dsuess/Pytorch ONNX.ipynb

## Pytorch ONNX.ipynb
{
  "cells": [
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "import os\nos.environ['CUDA_VISIBLE_DEVICES'] = '1,2'\nimport torch \nimport torchvision as tv\nimport onnx\nimport caffe2.python.onnx.backend as onnx_backend\n\n\nfrom PIL import Image\nfrom urllib import request\nfrom io import BytesIO\n\nprint(torch.__version__, onnx.__version__)",
      "execution_count": 5,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "1.0.0 1.3.0\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "device = 'cuda:0'\nmodel = tv.models.resnet101(pretrained=True)\nmodel = model.to(device).eval()",
      "execution_count": 6,
      "outputs": []
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "response = request.urlopen('https://3.bp.blogspot.com/-zBcdpq0NcLc/VrfAuIm_AzI/AAAAAAAAtCg/rymMidJo2-Y/s1600/individualImage.png').read()\n\nimg = Image.open(BytesIO(response)).resize((224, 224))\ntransform = tv.transforms.Compose([\n    tv.transforms.ToTensor(),\n    tv.transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])])\nimg",
      "execution_count": 7,
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": "<PIL.Image.Image image mode=RGB size=224x224 at 0x7F723009FF28>"
          },
          "execution_count": 7,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "x = transform(img)[None]\nx = x.repeat(32, 1, 1, 1).contiguous()\nprint(x.shape)\nx = x.to(device)\nimg",
      "execution_count": 8,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "torch.Size([32, 3, 224, 224])\n"
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOAAAADgCAIAAACVT/22AAEAAElEQVR4nGT9Z4xtWZYeiG1zvLvehH0R8WyaSlM2q9qQxWZ3s5tDDjmkBAggCZKQRvwxQ0ASQWgkCDKABhAEQRhBEgaYEdHDoZkZDmc4za7uZhsWu7psZ6XPfC9fPhM+4vp7vNlWP1bEZUGKH4mXYe45Z5+11/rWt761Nv7Bb/8XWmvTNDHGdV0jhDDGhmForZVS7XZ7uVxSSuu6xhhzzrvd7mq1QpQYhkHIzX/n83kQBAihuq73tnfiOLZtu65rQohpmlVVGYZRVZXjOHEcO44D//A8z7ZtpZRt20KIPM+73b5Sqq4YJjqKPErpxeWZbRiO7SKCOeeO40gplUJFWXphEAUuIWQ2mxuGgTTBGAdBhDHWWvOGcc5N07Rtm3POOccYB0EQx7HWOggCzrkQQkqJEPJ9nzFmWdZmBaqq6vV6mwcvioJz7vu+YRjwC5xzKSWlFCFkmuZsNsvzPAxD13U9z2OMmaaJECKECCGUUo7jwC/Xdc05RwRrrVutFiGEEFLXdV3XUkqllGs7WmspJdwVY8y2bdu2q6oqqrppmna7jRCqqkoIEYYhPK9BMCGkaRp4mwghzjljzHVdIYRlWZxzeEDbtimlWZZ5npdlGSxp0zSu6zLGEELwD9d1YXEMw7AsSynVNA08PkIEHj8IAq11URSUUnjSqqqCIFBKgWForRFCsGJCCMMwXNetqopz7rpuWZYIIaWU1Dc2gDGWUjqOU5ZlFEV1wfBPfu8fg2m6rrtcLj3PK4oiCIIgCLIsM00zTVPLsrTWlmUxxgghlNKaM9u2y7KEpzUMA16SbduRH6RpSghBCMErhBVUSnHOwVyqqkIItdvtIAgwxrPZrN0JirzSGiOEDMNCSCklMMZScd921+u14/mwHxBCjuNUdU0pFbKhlPZ6/bqu57Ol53mmaZum2TQNxYRzDmsEF4Unb5omDENKaVEU8NqEEFpreEOwTEVReJ4HNwm/o5TyPA8hxBgTQsAGM03T8zzHcVarFeccIUQpFULYto0xtiyrLEt4Q51OJ8syuASshu06y+XS9/0oiuDeGGNKKSmlY9m2bcN3KKVaa845pZRSGrbaQoimaZIk8X1faw1XdBxHCS6EqOva8zxYdtj8QggwMnATsFdhTzZNwxiLokhrDTsK7k1KyTkPw5AQAr/cbreFEGVZEkKUUggR+BHnHO4B9h5sMDBEMBiEUNM0lmXBhldKMcbCMIQ19zxvtVoJIUzbEkIwxsqy7HQ6UkrYt0Qb+Mf/6h/BAsFjuK47m83gp5vVNE0TdtVyuSSEYIypZVZVBXemtV4ul1pr3/c7nc71xWW324Ungbdrmia4kPl8DstaFMX29rbneefn54SQosjqJivL8pVXXqfEtG03jmNCECxK4PtS6CiKCCFlWVJK1+s1oYhS6vq2EIIxrpQyqOW6btNwwzDqug48H3an53mws8uydBzH933LssCRw4LCy4B96LouvHKwM6VUmqamaVqWRSmVUhqGMZ1OB4MBGBO8J9M0YR2apsEYY4zhO6Zp1nUN+9wwDLBp+Pf1dLKzs6O1dl0XIQReFtyBQSjsw6Zpsizb2tqCXUQIkRpVVWXbtmEYjuOANWity7J0LBMsDN70xkDruoaly/O8KIpOpwPbzDRNsEvYwxhjz/OapgGbi6IoTVN49YZhSCld1wWr4JxrjfM8tywL9gZ8IDgOCBFgo7D+hJCqqpqmCYJgs3nAQXieBy9LaiWlhBvAGDdNA56izCv8w9/5h2BMWmvDMMqyhEgqhNg4eXgTcAdaa4iS4Erh7mGLIIT6/X4cx/COCSGwUrCfwBPArYODsSxrOp26ro0xfvL0/W63u793j3MhuMqyDJyr7/uO48VxahkmY7Vpmr7vc1ZrjMqyrJoiDMN2u0MpPTu9aLfblJrwsiUXSikAHlJKKWVVVRD7XNeF/Q2OqizLuq5N0ySEOI4Dtw0vYOM4r6+v4zg+ODjI89wwDIgYG1uEbQz7B5wlxhjeJcQsuBP4PliD5diwGhD+TNOMogjeDcUE7Liu6zzP4adKKYQQItQ0TXCijuMYhpFlGfgwyZlhGIBJyO0XRA+MMbxfeEwAHlprx3Hg82GJLMuCv5JSAobZ4L2bsFnXWuuqqnq9QVVVsCERQhCRwP/B50PYTJLEcRx40WBCgCXA3mBxGGOc806vCxsYHgp2ThiGkgv88R//d8vl0rZtz/MwxpPJpN1uV1UFdwleEO5+85xCCEopQBzwPWATWuvt7W3TNOM4hr0LYQJCA2Os0+ncWI+UrVZrtVqdnZ3N53PGatfX/f6AEltrcu/ePcEVuBYpZcM1Uur68sp1bdu2B/3uBx98UJY5oThqte4cHriu9/TpU4KNo6MjpW5iihI3OxLiwGw2q6rqzp07SZIAcgKcBG83TVOwNoBx3W4XVhDsmFLqum6apowxiJLtdnuxWLRaLVhKpVRVVUmS7OzsbDanUkop5fu+lDLLMtu2wVOCZQh1A/vAvw4GgzRNb+xeo+VyaVlWp9MpigKsE9ZQagQ3CX4BVhL+yjYNACcYY7A8sNc8zweDASBp2CdJkoAPI4QkSRKGIdgZvBQArPDGwSKLomi327DTAJsmSQbmgTEGpAdWCK8edhe8aM/zAD0ahgGLvEFc8IE3EMVzZ7OZ7/vgYjHGENAEa/Dv/bP/N/wewGd4nbAvfd/fbCB4tuvr63a7TSnFGoHrBiSRZRnkCnEcW44Na0opBaCzXq+DICjLst/vA+IGdJzn+R/90R9RSofD/mR2fO/eg3t3HzUNw8jknDuOBdmJF7QZY+vlXCuxWszDMJxPJ5ZlUErv3L1nWGZZVvv7+1XZKKVM04YAIbmo6xr8N3yO67p5noO1QYKS5zn4P3gQyJMg9MNiAUhaLpdSStu2Ad5UVRWGYVVV4NXgGeGlQsTMsgwSCNM0OedgHOC3wLaUUmVdUUrDMFytVrCGhJBOpxNF0Xw6g6tAhpfnOaBnxphh3bjDTT4kpYTN5lgmIcS2bcYYfNPzPADBhJB2u71er+E9QtoAPm8DHuCLUgoeDtz8xs5gDeE3McZKIbgl8DuA1DfZEmMMYj0kymCCSinYb3A/SqnNhuGcCyURQoDfwNPDfvYcG/8P//D/dnBwUBRFXddFUUyn09dff325XAZBAJ4SLA9yTLiwlNIglHMOUbhpmg2kIIRQ0+Ccdzody7JOTk5M04TwCt6ec75JdNI0PT8/b5rGMIhl07t37za1oJRqTRFCGEtCyGI5izoDwzCWixmry7PjYy24bRFwWl/68jfa3Y5tO8fHx1vjHcdxmoZTSvv9frKOkyRpt9tJknieBwkNZJQQIqMoevny5XA4hOQPPMrGUsEKwVH9LCKH60Kyv0EFjLGrqyvHcSAPgOBomuZyueScHxwcmKYJWXMcx/1+HyEUpwkks0B0gJsBm2Z1s3l/eZ7btg17SQjBpYJkGWy3aRp4EK01b2pIyMCnAOgXQkCgg7cDCRmgWzBTiMiQQkFUhLgBMBT+dvNr8KI556Zpg2nC5wB6BmIBEqAN0QG2CAsIuBx2e7/f3yTNCKGqqQFPl2UZhiFEadM0keb4/e/+N1VVAZcBV4LbyrIMohukCPBxsGniOAYqBN43pFbwSNPpNC+L0WgEGxG2BWxEQHWwoJuEF8CrUiIMQ8e5AelpmjuOpbUMgkBIJjEpyzJPkyyJv3j6xLcNpBWYe8HU619684svnjm2hxDZ39/f3h7fXDevwMGEYXh1dTUcDsFD5HkO1laWZa/Xu76+hpBd13Wn0wFcxRiD0GyapuM48No2NwxPBMsF6BaiOSQN8E3wmlEUKaWSJIEkAyIdEDHT+QzCXxiGtm1/+OGHu7u7YFi2acEHgi/fUE5VVVHTArcEex6iKqRNtmmAR4e4DM8IXhbAH6UU7pAxBjgBtgFYEqSA4OANw9hAEYgGwJFBdo8xtm0XciywV4AZSqmiKBzHwRhDpAI6rCgKQBFZlimlXNdttVqAi5qmgRXIywL2/2KxgNseDAZlWRIs8X/6f/l7Sqm33noL8imtdbfbvbq6AtAZhiHkK/CEi8VCa23bpmEYdVNC6gebrCiK3d3d9XoN95Qkyfb2NnhccEtFUXRabXCogLVhXQAzQewAQAOrA/ivqioliVQcaKzj4+fw1htW2bYdRu27Rw8wNk3TbGresCoIHGpgzrnnRggReKimaVqtFuQcvW4EexpCIQSdKIoopVILx/Ng7xrYKPPKMk0DG1XZBL5dVGVd15AmKoxarVZRllJKgm7wXJqm8AKUUuv1WkqpMWm1WoeHd46PXzRNY1o0DH24blNLWFtITQBHVVXl+/5isbDtG7Ks3++fnp5ijCEiQSYK0B+MD+wGtuuGNIUVBu/Q7/fLsqyqynVdSJLArXa73cVi0e12wRECAbfJIKXinHPTpIBbIFZAOGZMUGICIQhADm5ASimEwFgbhgGMKURqAFGMMYSFYVBKLSWBFlVCKNhg1CSwhze0FMBux6b4+9/5DcMwGGOwYxBCQKzApjEMI8/zDZ0B36yqwvM8pcVmgQC1QJ5EKQUHAHkr3Dr4m3i1BjcG/gZgMvwt8Mngv8GRw3pxzvOsbnci2NYIqSiKyrI4OTnxPM/z/E6nV5WN4zicS4SU41paS601wZZlWXmeU2rCEgPqDyMXXjMkLhhR2Gy2bWNCfN9XSjV1Dd7CpAZYs4ENIFwawcFHmpYFzyU4h1UCywAMAP5DSDydXjuOw1jt+77SwjQp5AFa0Y1P2pgmgDAAA4DaIScD4Lher4HsA1wBxrFBkAihTWTc7PyiKAaDweZ9w1sGkAMsjRAiCAIwbiBKITfoD7pVVbmuDcwlMGsAflzX1wobhgUJHBiP7/uEkKIoEFKwDuC2weMAsrRswlhjWXZVcsAnluUgRDDGjnfD+8LuhTyhaRrPcQl4abhAVVXz+RySAyHEbDaL4xjQPbwVAEZRFEGGCBhig88gwAH+gK8oiiBqg0FAOAM+DJzNJpXboD3INiDQABt1eXl5cX4FmB3cgNbo4cOHR0dHnhdUVWWYyDARoYIaWkqZ52VVNaypqjIXvGFNkaWrskhevniqFauqAkoHg0EvDH34W61lFAVNXafrVZVnZZ6VecFYrbGiJkFEE4K44lxxzhvDILZrNbzOilRjxUSzWM2JgfMywxR5gdvtd1qdqOF1HK+63a5hEICDsL2XyyX4M+D8YcUgd4Q4vrOz4/s+5NrgfsA5QR0LMiFYK1g3MCnf913XBXjm+75t24BSyrK8uLiYz+eb7CdJkqZpwOgJIc+fP3/69OnmrcEiQ1lBCLFeryml3W4XjAywZhAEhJBWqwWm3+/34e1vMOWGxIBvwgMahsEYl1JicsM5SimLIhOCFUVW1yXnjRAMY+26dpYlSokiy/FP//C/gvzfMIxut9s0DVAMkEw1TQMQfmtrq2maqqoIIXmeOq4FYd33fYAjwAtIKW3TAkRimiasMiAe2E9QBbFte7VaIYTu3LkDWAQwB6SuwC2Dh1BKUWIrpaiBlVKcN4ZhMNYAipVcO46jNO90WkWRVRV3bF9rbdvubHJe1zVjtW3b63UC8LfT6SRFBRWjVqullEJIAchzHAfJmnPIHsyi4ggT07QxMYRQRIub1DjwLcsqyhIcm2FbSEitda/Xg8wJeGIAPwQb4JngvdrODaKVUmr1b9lHQEG+76dpCmQfJNHg7cArA/R0HAfc/ybF3nB/8E0oM0Lggp/Ce6SUwqXDMIQ8dWOvEM3A44CLVUpJxcsyBz8CXDgkuIQQpRAlppQa9skGu9u2nWUZQqquazAAyOdg4wkhuLgBxJxzrTBwJlCABXAMaXer1ZrNZhAcyjTDP/rd/xKIQMhdwM8DdL28vLQsC6p5YRgyxhaLhe/7lmXAcud57jjOer2GZTVNkzEmuYAiMjhmznlZlkqp/f391Wq1wTRAN65WK0ClCCFIGyG77/V6cBt5nlNiQ9ChlF5fX+7s7NV1gTGhlF5fzR3XMk0SJyvbtqOglcR5EERFUbA6I4T4vj+fz2GZgO6N89LzvO2t3X6/D/fmeR782s4oePHihWVZXCrBVW8wrqqm3ekhgl2LAp9Ss+bw8HC1Wk3nM+B6ArcFKBxsAsC0lLLb7WokgUW64eDW6WDY40xmWQYuEwx3Q1rBJaDsDhnxBk0C4IMHgc+HbwIq8zwPciCozIEtArKsqsrzPHCcUB+CkAWYMs9zMBrwfOCnqqpSGoxbwnM1TQMlQKUUpWaRV5Sam/uHGwMD0FqCD7IsKwiCoijgcQBPbwhNIdkm+QbLBpiXJAkhxLZcznlRZt0oxP/9b/xfx+MxmCZEWPAEEGeBMBNCdDodpdRisbAsK4qCoig0kpsdDP4fKldaKq21EtI0zevpxDCMoijG4zFsfVh0eLUbZhGKH3AtpVSapoPBACKR53lKkiAIEAaxAkEIYYwYY1VVc4EGgx4mGl55XdcEG0+fPrMsK3CdUX9AKV2tVtjAWmMpZRRFnt+yLOv4+LQostlsZhg35c1+v29odnp6bBgGl8I0zaoqhRCGbXW77dC74Vx6vV7DeRRFtusAyHFtD1AdhGN4zVATR0QbhgU4kjGGEb21QkIIAiXDpuSzKd6AX4FcG2g+YJrLsux2u1BZhb8C0LYpSJJbYQrEBKhTwKqCNQOJATYBd7KpBoEeBSoRCKGyysMwdJybSjXU/+BBKDWVRHleQtklyzJYB/DEoKBYrVbgHaCyCDB3vV7DZsjzPAg9Qohtm5Da27YLwYdSWhUl2GFVFSalBrDxsAmgLAbxDkoOwEo4jgNhYhP6pZRCMlhfKFpsmAvJbip4CKFWGDHGsEZKSJMaiGC4FmAggPObqvcmuwekDKSM4zhVyTHG+uYBCELYNEEpg2TJqqphUjAmRVElSTabzZqm8b1QNMy2XajiaK2BXRuNtj76+HG325VCZWl5794DJbVt29fX16cnZ7ujURD2GGNUs0F/UFYZQqqqC9c2O6328fFxEARZkqZ5ppTymBcEQbvVypKkqqqiKCB49ft9x3GEYJxTy7GFYMBIQPatta5rZhiGYZANew/+bOMmN+AS8iH4klJ2Oh34EYBOyBrB2iCkbgQx8MjtdntjcJA/AA7eMIAgKQKEBgkWUApaa+BtoKoOdg+XSJKk3e4qhTYFJEhdwJHD/buuC0EfeLSqqgBpUGpqLcGLYaLBuAnBlmWz+obXYw2Tkpumo5QwTdMyDPwH//w/jaIIamsbNIMQ6vf7YO/n5+f9fh8WESpGvu9KKaXigKaBKN4U2ZuiAoYIjK/b7W7UaI7nAtqAtBEqjRjjdrsNVVp4TkBIQDHWdW1bvpA3xYnFYtZqtYTgGiOltNC0afj19bRqmGhYw+o8Tw3DsG3bRIZpmv1e7+joCALobDpfrVZN03Auv/KVr3zwwQe9Xk9rvV6vt7e3GWOirr72ta+t16vLq/OtrVG8npumYVKd53lVcc9xoQK0ThKEEDEo3OTDB/darRYhCMIoRNKm4U3T+O3O9va2ZdkYY0IMpInWEAd1XZfwyP//lgSOBPwWIQSiHmMMfg39jB4SPDf8Idg0oCkw0I0saPNmgVuF4iogK/AL4FAgx4K6UZYnGywHdCl4ZUKI6/qz6aLT6X3yySdCiHfeeQfgI0BkSjG8KQgC4GIYY6vVyvOCKIqA1gDeME3jViuybbvMS3jwPM9D3weGwTAMJhD+6R/+V8vlUinVbreB6wdc73keoLfDw0NwzrAQlmWZJsUYA+YtyzIIAtjNoLEIHI8g9LOs26Z0FOcZvZWHQY14MpkIIYDCgJWCS4NoAAw0z2rXs2/ZKBtjWhSpYVlIk9PJpCzZRx9+Wha15M329naSroLAa5pmd+tgPB4PBoNut/vBe+9VVTWbzfb29k5eHt+5cyfLisPDQ63lgwcPhBBJkjx58mRr2MEYg3bE85zVbMoYe+uNL33wwQdVXdq23W63Xd+DpSA3tBdK4mkYhufn56ZpvPLKKxjjoijBOWWNKIrq7tG94XDYNLzVaedZDUJYwCq+789msyiKuGhg8cEpbgQcjDHLdGzbPjs7A/If0DnkfISQTYEafAR4Sj8MoMg0GAxM04Sq72QygawUNCtwCXUrxoXyBJiyYRhSAT+DgOrKsmx7exsUqIvFajzaFkJBwISUF+gawzAoxaZprtdrwLIQn4G9r2thGKQoMtd1lQb6Qtm23TQ1UQhIA9d11+s1sBBaYWoH+P3v/jdZlkEQgToVmAUhyDRN06KbujxAENijrTCCggFw7LclWgXE3mg0gtop7Fd4kna7vVEwwKvdUKee53FRa4WlVI7jIU201ggrpVRdl57rUGJUDZNK+b5fFI3QQkgtpZ7PVk+fPoUaRplXQND2+33LsgZb23VZbm9vP3v2jNWVSalgzSuvvAIcYbvdfvL4s/F4SLF2XXs2n7iOH3itqiour847UfjKq4+uL68cxxkPxxjj2Xze6XTSNHUcp9PpXF9fQxkzSZK9vTuXl5dZHmsth8M+bNeiqGzbRgSbplmxhjFmWFQrfHT/wWy6pJQ6vkct03VdXtVN07TCwLJNyRpQZjAmhFZR2GqaxrDs2XxZFMXWzp5tqDzPMaaUmFB6WK0XnU7LsixKbCGEEMrxbM65RPrs7Gw2m7Xb7XFv1Ol0AID+rD05jlOWOUBPyFw3wtmizILApZRqjQzDUApZlrVaxkA7QNYFiQdwFJCMQ0DAGANyAKZZ/oyODoxnQ0XB3gDH2Wq1QCgMoBnAQ1ky/MG/+WewFzdKUkixiyKzLKsoM2Au6M+IdMIwrMsK4gWwSwDSYcdcX1/v7OxAjR6AEdwrlFg2qMs0zaIoIGBRSi2bZmlhmpZtuwQbWmsubjyESYllWSDkyfLc8aK8SE/PLq6vp9NZfH5+rpQaD4dh0BKShX5w9+7dfr8fRO7l5eWnn3zy9ttvz64ngecPBgOorQVB0NT1y5cvfN+/f/fQ8zylhO85rGS9fqcscz9wsiwpshwKG48ePnx5fA4VF9M0W63Wpjz40UcfPf7s2Xg8JAQhrMPQH4/HjuNkWeH7vh+E19fXCuuyLDWSVc2+8fVvrZK03xsus6SqqmS9dB2LMdZtt0yLYKUdx3FBcoZQUTdlWW5t76zXSV4UaZoiLe/evUsIaRruuYGQzHVdzpuiyG2vA2VYfJuwe56n5U2BGyAsxng0Gv3mb/5m0zTf/va3TdMsyxzyijzPIapC8mRaVGtBKZVSAT7RWoPjADOAei+Q82BPcHVwVY7jtNvtyWRyWxq9Sb8A3UEoaLVaQBnBp0GiBtEJynuO4xDDxN//zm8AgtmgT6jaGQbR6CZTgW2hbhUMhmGwumGMAZW4SSehaIQQarVa0MywWCygzglK4fl8DtYJqBwURlBQ8XybM8m5QIjUFXMch4sG9DgAuWzbRpiWZX5xMZmv5kqiMGz96N338qLSUuVpClnkwcHBq48eOo7Dm/L09PTVV1+N4/gbX//6H//R91599VXTNIVUQoizs7NWqxWFoWVZrmV3u931epVn2dXVhdKs14nKMh+PekKI1Wplm9Zind6/f//4+LgoCuAu5vN5URSEGA/vvyEECyO/rssvf+Vt4Mybmp2envpei1rm+cVpt9v94Y9/8NqrX+ISj8fb15Ppcr0aDAZNUwehxxiLQtc0qUEJpdg1LUzIwcFRURR5nudl9fTp0zhNdnZ2ut3uK6+8sl4nju1dX18bhtHttSEPq6QkCCOJgsCj2LBMk1W167qXl5eKYBDKgNwMhKe+70OfSRAElFKQ6EPhkRCSZrFhYN/3DcPknNu22zQNJSbnPI7jbre7UYSARQJFCtoreF+bt9xqtQBtgx8EbwqBGl4rRFHbtpfLJXwm5CFA6hmgFrNtGygr+LNWq4WxruoCoKFhGODVQYy84dgAF65Wq1arBRkSEJl1Xa/Xa5AhAqXKOZ/P5wBwF4uFlHJrawtEkFBWmc+Wg8EgzwvbdglFSgvwuBhjpXEYtReLxdnZ2fHxMWMCuoWuLy6FxItlXJfV7s6O57mDbv/Bgwee6ywWi9n1y3v37pVF5pjG0yePe70eXPro/n3DsBoutFT9wej68qpATc3UfDItq9zz/KJUXzx/0QqDvb2d45dP/cANwlbDjcefPbu6uhqNRo7tTiYTg7qvv/bAoJbChCqTmvaTjz+6c3QYRaFpOw0Xj157fbnIzs7Oev0xpThNyt39o9ls4bgBxou33/7K8+fPq6rShGitaVl3u+3lanF0dDS5vuI1Y0yapnl4eEgXi4cPH56cnLC6OTo6iuNYSdw08uDgSAjBWC2Eev78i+F46DhOmZdlEfth6DhOkefU6Ue9sCr5RrAHVfs8zzcyOShFQnYPoTKO41Y7qqo8SRLHuZFyEkKcwAPyCMwRyl1QVYE0FNKjTX4M7Ph0Ot0UyaCABH4XqBUQsG6KDpCoQTp+fX19dPcA/+5//f+EjGRTCIGC5Gq1iKKI0Bv+DLzdhmsY9gcg4oKPBoAP1glwG/QNUGoDMQ54cth/QCQRQgBU1HXtOPZNjUFDHUy5DvTcmDVrPv7447OzM+hhWq/X6/U6yTIppSLWYr5SStmmGUXRg3v3t7a2TJNyzvN4enJy8qd+8dtRu3V2dvH1r30jyzKN0dnF1Xg85k2DEOp3e8+fPzepBUzK8fMXd+/e5aK5npxZJv7Sl16dXF0BgRdGQ875crkE2g80EGmaBkF0cOcwL/PPHn+yWMx++Zd/qWFVEARlUY/H21LgoihW68V8Pp/O52VZPnz0GkLG9tZu1ZRQfH/+4nMpJVdiOOxbtjEcDus0t217b28HLHgw6PW7XSFEXpaO53pegDSJ4xTE4FmWvHj57Pr6cn97DAWO04vzB48eQuHD9TytdeC1b7QXjrNRPIFEva4YiCrBgXW7XaBuMNFCNIQQ23ayLKPURAgZ9KbkA6IiiNcARjfl/jAMoSlDa93tdtfrNZgdUGabaOz7flEUoBeDdwqmstHFu65rWZZUHH//O79RFAUYDdRPwY9SiimljNdAMEE1DLIcy7JY3UAbFFDBAIG1viHMwcMDGw9YHiRRrVZruVyCRgFjDOEGPCiQo77vAUVVlpXrhEVRNA1/8uwF59yyTNGwuimLLK+qajKZNE2lDWexWAz6/W9+85tNWUVRdHFxYVJjMBjs39lZL1dRu7VaJ4eH945PT1999XXonYKaWV2Ui+VsPBwxxrQSlNLZdOV5TlnlUeC+9dbrz774IkkSLaRp2Kskx4iAEhQETZAMrZP0/sOHJ6cv58u569pSsEePHrRarfU6sUyn3x+evHg53hmXZRmnyWefPbFsJww6tu0aFCul6qZ5+6tvc86fPvui1Wqdn196nmdbRCk56Pa4aMbD0Xw+FZwfHBx88cXne3f2m6ZB2oiiKAxbZVn++Mc/zPK1lHx3NFBK2a61TpIkyyF9+fo3viGlHPS2fN+ntz1VsMdAhua5ATg/CJWA2aqqKsosinzHcSzLjuNYCGVZllbYNM35fD4YDCDjAeEc5O9RFK1WK9/30a1UGeiaxWIRhiE0agIZvGFyyG0TCFgqlMqB5QBbxwjhn/zePwaHxzkHdALi2aapMMaWbYB0Ej5xo1FgdQOcWV3Xg8EAOiSjKLq6ugLqC2hYKO7D65xMJgBHgHQA/vZWo4Udx6uqwjCoaVGtsOf5rFGUmu/+9P28ZkdHR59+8lG6Xl9dXfV7HSmllJwxtlyl8/n8r/21vzYej/Ms++Lzp1Bu+dKXvnR+eTYajR48fOXl6Wm3M7xzeLBcrMumef/ddx8+ePDixQvTxAZBrVZLCp4kiVJiZ2tXKSUl503hevbu9vZisUjTbL1YI8NeLBbDwZgx5vkOhIXhcNjUvOFqOp95ntPtdeo6f+eddyaTGef84ODo8aefMcYQUo7jtLsd1/XjJFUKnRyfGRjVrOkPB/cfvlKW5WQ6d4MAuLA4XhOCtsfD2fQ6DMPxYCCEYE3V7XbPL47DoBWGrel0XlcN4816vRwMekoLz7EDzzNNo+Ysz4okSeI4ns/nu7u7vu+3u51Hjx6BVgs80Y1uUN50IIFNgCUEQZDlieOYYRiWZVUUBUJkPB7nWem6bhzHQGBBbQKSdHBA4Izn8zlkP7u7u6BVBX3Fhh0HPAm5P1RQIdOA0trGQhhjWiL8g9/+LyA0g2DshmxvGlhWLpqiKEC+BPgAvLdBKJgmCErIbacV3CK4z+vr6/39fSg1gUChqqput2tZVlEUVVW12zehx7IsIZTWklICId5x3CTJXrw8e/7spRu2ri4uPcfCGDVNUxaZYZAsy4QQl9eTr3/1a++8887Z2dl6ufI878G9e51O5/z8/N6jh03D13GiENEab+1sL9epY3sUK9M0F/OpbZunxy8uLy/29/aiKFitFvfv3cvztMqL8dbQdZ2zk1PHcTAmlxfT/nDUNM3WeAdjPJtPgiAAysKg1pOnz8NWVBQZ4839+4e27dq2fXZ20e30FvNZEARlmWuM4jh+4403lqvYcXxKKcWEsVpq1O8Nnz5/dnh0/4vnLzudjmGZk8lVmsRf+crbL5499X1/vVr84s/9/Gefffraq6+6rvn06dPVKn7w4FFVVcvl8vT0GGNdVZVlOYQg27Jc10VS5XmupVoul0iqRbEAlrrd6bz9lS+/+uqrEJSbpjHJTRfrv5W6ghKS6Louut1u0zCov5dlaRr2piTearWgLg3qVWjaBOcFkhcoN3a7XchtoPgC6Tzoh/I8hxYMCMJg04Ax4ji+qSZQC/+b/+E/R7eTC6CXCIgDx7EoIVpL4FMANY5GIxAs6lvdJ7C7AEChHAoJI5SVQfEAPwV6ApzrfD53XRc0s5DKYYqlVJiaUkqM6XKxvp7OpFCr5XI+W5Z5blJCMLYs8+L6qqgLoaRhGK8+ePgLv/ALy/nivXd/2ul1t0ZjivHW1lacpkwJSkzH8eu67g1GjuMgpGzXOju5DD3//OLUtZ1PPvnoT/2pP/X82Rez2eznv/mtNE7LKgtDX2kWtfzT01OCTc+N5vM1l6zdjqIo6vU6ZVl2Oj3bdg1qnZ5ezBarwaBXVnnT1KZJ9/f35/Ol47iuE37+/PPAczcquKZpfv4Xfu7D9z8YDkevvPpGWZZPPv88aEWmaa6TOEkSoVRVVb5haYKX61XN2NbWFiHENa3RoKeEFHXjBu4XX3wRhuFyFXuex4XK8zxJ18vrK6VU3ZSU0l6vEwRBq9sq0oyLmshKcsGlTCtWCTadzYhpfPmrb//5f+fXRCk9JxBCVWUD6UcQBFm+5pyBvdq2naYZpWaelVD/fPLkyRtvvLER0kPsBp0Q5MebxBxmNGy6n0E8OZvNtra2iqJYrVaMsYcPH0IhatPptUnhkyRBiOAf/6t/BLADDAWoTYh0vuchpEzTBN8LCRqI1kBzAPARvCbE65+V7kI9nVIKeAImasB2ATlqFEVJktzQvI5tmY7CSCkVx+kf/MEfdKKO67q97uCjDz5Zpck6SfOyVEhjpKgQJiHdbvcv/KU/9+zZM8HV0dG9TqeTJNn19bXnB4SQJM8oMQ/v3rMtd72Og8A3LYqxbqoaKXV6evrg3l1oJl4uFr1edzmbM8b39rfn8+nde/tQ8Fivk7LgBFuL5eSVVx6GYUgo8jyPEhMh0usN3vvph4PRju+7CEvfd6HeOJstB4PBp5987nj2aDS4vr6GZFlKWVXVnf29q6vre/dfuWkPYswLg4YzyAwWq1U6W/hB4PjeZ08/Pzw87Ha78TJuinw0HFomtmx7vV4/e/ZivLWDMS4bFq+TNE3Xs6uqyJSSxCSe5xkm6XbbjuMo0ZC6YIwphSQ20rIo60prnZTp3bt3f+mXvt3tDDEmUmBIHizLKqucsRp8m+M4tm0jRCzLWq1WkCVTYm7kRPK2HQUoJ/DE8P0NCQVAk/5MkzqU8iFpCcMQxoUAbIXeGACTQij8nX/yn3Q6HegN+NmuPCHYaDjkvFkul47jgHIZlGZQAgZ2AKoFUM/ktwMwgOgCQRfoEqBA3G63oQQKKruNgskwjLyoPC+ASPTRRx95nrNaLkE1c3E5m8XJ5Xxes0Zr5BDcsa0//0t/5p1vfO346vjF8fHW9n5WlEzo0Xi74XK5jo+OjpaLFcY4zQvH8cqy3N/d4ayuiyIMHN91l8t5p9U+Pz8XkksuCEHz6XR7a5camPNmMOgxXgkhWSOKoonXaafrh5EfBEGSxIPBwPP8xXyFMaXELura9/39/b00TTlvut3uycnZYDD8yU9+0u12LduGbtif//mfPz09vby8fO211/r9YRi1z87OpJS258Lq5XlOTaMsS1PjNMsm81kURQ1nTVV/61vfmlxeMV7bFiWUOo4bx2mSVU3TaETjdXpxceGaCjq3EFW8YYSgvYPte0d366ZARVWWZVU1XEitbwYcMcaYaHbu7v17f/WvWKbTNCIMQ84kMEEgflASIQziFUwIoZTUdW0Y5ga5AikOCS7Q7PS2nwLEQPBrlmVdXV1B1woI2KE3WEoJfc/gyMDSwOmC2VBq3qTnAGOhqAPCHyiiyNuRLLCTwMY3ygNQG0D+vvkCfAN8GzALP1u8Aghr2/amEQ/2HEj/oVmvrmvTMAjCmOgsS2ohudBKE6FwXddlXhzc2R8MO0WRXF5eBkHQ7ffysnYdP05zJuTzFy9fvDxdrmPT9pKsuprM0rzMinK1TuIsk1IuFgtICj3fdV3X8xzoak+S5PT0FCFSluVyESuJbvorXItznud5VZXQeNk0da/Xs21bKp4k8WIxxxgvFovt7d3FYmVZFkLadR3OeZ6lnussV4uzs7M0TYMgqBp2cnb643f/xPG9yXwG9ArGeLlcQhdDp9fe2h5tjQbb2+N7R4dS8jyNNZKOY2FCTs+OLctEhJRlKaRstVqO7wVBVDe8KEsmuFbI83zb9aqyWS7XSmqpkUaUUMMwLNf1HccjxLAsB2PaNA0MMbAsA2RlMPUII2pQK01TaKexLIsQDHRpWRYbXkkIURRFc9tWz39m3BWATqgMwfehkxtKNps+fficjT+GChOQQpzzyWSCP/yj/3aT6pPbWTyEENe1HduO41Wn0+GcLxaL1Wq1tbUF7arktiF6ozYA0j5NU6A8YQ+1220QQJHbMRtQZgDCFSK+ZVmGYbXa3aKoXr58+ezZs52dncvLc6hOJVl1fBnPl/FqtTQIcm3r3uH2/+R/9O8oVluGeXJx0R+O0qJx3CCI+n/8wx96UScKO/fv3//pT3+6WCzmqzVsSt+2sOSBb4+7gdZ6e3s7DAPf8548+ezO3v5yNT958RIRy3XtXq+3Wi/29nak1FmWKYmSJFkuLh88eEAIvn//vpR8PN7+4P2PR6OtOE5Pz8/39vYoNYGMOz4+ZoyB0hEpdffu3eU6kVIKqfr9fhC1JrNZFLUP7x5ZlmWb1uXlJdRgoe+HMVbmqed5UJKJ16nnO0VR9Pt93tQS8a2trfls+fCV18uK53nxne/8zutvfPmDDz64vjzlTc0F8zyn24uw0mmyOjo66PU7kesjrTmTomFu4EupkyQps9yyjFW1Oj07+/KXv/zrv/7nOedKIdfxoXYN2wZERo5rUErTNAHC26AuOCMwKeCGwCvdCGF/Zr6aZVmAA6EIdxMrKAWmE9IeqCF7nrdYLJzb+Vmr1UopZNR1DV7whhqVEsjLxWLmex6kq47jjMfjXq8HcmPQ9IMTBpiSJAnYPmwdyIRgSwHHCWJhYKbAiQJDe2vuxmqRcCHef+9Dx/fmyzgrmsVqGadZlhXTROXr1DUNi2jVZP/jv/wX2lHw5PPTJElef/2rs8Xy8Ojh7/7eHxDT3do5/PDTx2FQ/PDHP9WSI2LMlwvOeasVLqb5Vr9rYlW5tN1u13UVBL7WstVqhZF/fnGqlHJde3t713XdlyfH3V6dprnv+1JxP4zGwy7jNWNisVgcHBycnZ0XRXF2dua6ftQKGGNR5GiNLy8vtdZFkZUlHo76omGEEMHqVqcHmoSXL18e3r3PmDg/P79z585qsWy325AFH905WK1WrGkopbZtTq/Tuq4ty+FNvbu7i7FerRZFmVqG6Qfe6fGLwWhHK26aVLDSdc2yZkVR8aawHUcpRTRqmma+mIaRP2fCsWyDYqcb1BWzLdeLWp4fSikL3QwGw8ePnzx48KDf749GYyFEWeWOdiz7RtYDLA3Qk1rrpmGG54KTA7lnr9cDntU0zW63G8cxGBUwnZAjAhsF/XdgbE3TbLAmGOVsNoPvgDiJc865xN/7l/8A9gEE4tVq9eGHH37zm990HKssCtCNwo9AyAhBH7wmpOfQuQIOH2wdqlUgfYIrgfVDMQnqGYATYOs4jlfk/KMPP86byvW80/Pr49OTqmEN50VVrdZN23VEGe+OB3//f/Uf1HU5mc/SIkcIXV7FhmUT00HEjLMybPWePP1CI7JerxFjaZEjSvI8dx0DCb7d6/Yjf29/cHBw4DiOH7gUE8exri4vGaufP/1isLUPwQG20/X1NSR/19fXW4PucDh89uzpl7/yFmOs0+6dnl4ihEbDraxax3GaxNmjR49evjzhnM9mE9ezlRKuaQ36ozjNoiiqmaCWc+fw7nS2oJQGrYBSOr2+fuONN4QQF2fniotBv6+0vri+2NkaCc5kzaElYxGvB6OhF/qKNZZhLlbLN95463oyNQzj7OJqMpltbe/+q3/9/aZpBG+6nRDLRiouWLW/u7O9M87SQkqpJN+7s7u9vZsmOcY0iXPHcahBKNVS8efPn/7Nv/U3lETr9frVV181Tev4+Bg6SRBCnN+04QO5ZpkOQujq6koIcXBwwDmfTqeEkF6vt1qtNkOmIE5uJHywmJvwvfFZ4K2N22Fjxm37pGEYnEsC5gyg8+LiwrKsr371q+q2dxaaTmCuAcgBAXeCm4RGZiheQeyGR2q3261Wy/M8y7Lee++9jz/++OTkBHhdyNnhojAND5z0arVarVa+FxZFk2S5pmYpVJyXq6QILGIT8b//3/39/8P/8T/KsuTy8no6Wfd7e61we/fOA4nMk4tpI3SallLqMq9ePn/ZVA0QpVVZZnlS13WepavV4ujocGdnp9VqQSP8aDQC7QwkDYSgyWRyeXmZ53mWZZ7rh0GU5/mf/tO/uFis8rxstTqfffrEMp3nz19SYrqO/8knnwZB8Pbbb1NKy7IGoVOe53meW9SwbRvqLsSw2p3unf3DH//4x3GWMinLunIcp9vvv3jxAqYyQWltsVjcv3vXc9z5dFbXFeCob77zDmPs9Vde3d3e0VqPh6OyzC3TGA76d/a2HYvE6/n29rYfRUxppbXERCrEhJrOZ8cvT68uL6+vrgzDWM4Xn378UVOXGCnLNJqmsS0PY7OuxKNHr/zn/9k/qOpiNO5dT84fP/lEKck501rBKCvbtqXQGFGkSZZlwG7atg1tF9DhCbPToMYGuQ00PzLGyrLM8xyQJLhYiBvQdgZEKaRZAIWBInVdl4CoEXrcBoMB6E1+VnENpVIwVrBLYLag7mfeTvGDtEndzjmBElGv1/vqV7/66NEjy7J6vR6htmXB+DLlOnaaxJKLPM0QQu999P7hw/tZUeZZaRAjXxfLy0UdVy6mnq7+F//hv9/rdo9PLs9nacrJ0WtvHz56g/pRJdFw62Bn905VNdS0333vp3mZaa1ms+kiWQuN8qLwHFsxbts249VyNZ1NZovZwqCUVeXk+rxI4mGv2/ZbjuVenZ1SLYbd1mo5jddzyaq6SGVTv3j6xeGdI6RwUwmsrZfPzwe9sWkYo+FQCC4F/eSTp0f3X+EKKYMqTH0/qkp+Z+9eXQnTdBG2CbY1siaz9atf+vLR3XsI40F/O83qJMkdL7yazBGmu3cOsqqWCF9fTqQQb3/pdcfBg4FvUbGcXj86uocVWq0X1MBh2FovY5MSy8RVsX7l0f10teZ1oVhpGCSpmnXTxIxnJZfayAtWMFVLvEjyOMvLsq7LLE/nBik7LWM6uQwDZ2e7H7nhG6+89S/++W/98//2NxWSYdtUSjHG0zRbLpecN1JyQlWv31rHc/BEruv2+31wQ8Ac9ft9rXEcp2VZa41N03769FldM4QIY4xQVdUZIaTTbUFJKQxbq1Xcbneh+7ppuG27CJGyrKuqMQyrqEqCEIKeJtBHMcaghQ0MGVryxO2oD3U7Zre5/YI/wbcjaqFDBRJ/8Oqe543H44ODgzRNBeObYmlZFyCkgtKAlHI2m52cXZycn19fTVerFaHKtkgrcP/O3/k7vV7vj7//o08ff55k+d6dw+cvjt/74MOsKB3Xffz5E9OxryazvMw6nY5pmkJJjREmZKNzdV2XaNTpdJbL5SZ5XC6Xq9UKprWkebZarbqd4dHhA61ou9XvdUd5WX/la9/ww9bl9VRTqggRCDlB8Ppbb338+HFe1ycXF3cfPPjs8ZNHj159/vy5UuqTTz4JQ39ra/j6669y3nzla1/VCK3X68VqWZZVVuRXV1c/+cm7aV78m+/90Ww2y4vq7Pzc8/3LyaRh7Ctf+cqXv/KVsiwn17Nnz54RQrI4gxeBEHr6+EkQBFEUrddxnue2YyZJopQKPP/evaM7ezu72zvACSRJUZUCUVsImuY1NRxqOOs4Yw1SSD99/uzk5CyKot2draO7+3kSp+sUaxIFrTdeex0p/Sc/fjdPK8/zQXMEwRqUeFC+Bm0HBOWyLCFqQ6eUaZphGMLcZ6XU/v5+u912HCsIAvhmURSM1ZRSkF/u7+9DIiWEgAwHAixC6KZb+Lf/6f8D+jFASAJ+NU1ThFRVljAgaUO3gmeFdngI7lAwANEUVO3h16CMBF148CWEKAsWhJ5hwNypmhASRR3G2Hd+63ei/jCvxJPPXyRJNrmeNXXue3ZVZP/b/+h/3QrD9957Lysa23EQpddX0+5waLv+Kl6zhmPDPL+4sG0XRvysV8lisUjTFBNNsEEoaqpaNvV41B+0/bdef9WzTNu29+/s5Gly995hHMfn5+cnL14iRPywhRB69OjR559/3ul0ID70er3pdHp2eun7/nQ67XQ6BwcHsESmaY5Go08fPwnDsNWJgsB7+vwLy0AU6cFgwKt6tLP78sVZlldlxauGmW5gmHZaFZZljYf9sizPz89fefjAtu0oCtIk2d3doZSGlmOZtKkzx8BlWZ5fnHpuVDd8NBoL1Pi+b1t+HMfreN7rdWzbrGtR5OzZi/PL68np9fXLiwuJtGEYjkE7oY81whQJIShBrcC3TNIKfdsyHMsMguDw8FAyyWruWrbjOAIxz/M++eyTv/RX/j0uJNCWELshrkJsZI2ybfv8/NzzPDA7ejstBqIuYErTNJuGI4TKMm+aptdvua6rFBKCOXZQ1zUhBuBa0MiDi4SFhTHfUTs0IL23buf+bJyoYRDTNAkxN5U62DQgrkvTFCCBvB1vBEjUuh2tuBFAbZjbLMs8LzAMQ2sphNAYY2I2TKxWqRd16orP56uyZnndGLbFGemE3l/69V8+3N/5/X/9A2y4w60BFJm2dnbTrLBdFPrBisWz2eTVV199+fKlaRvlopyv5lxKoVXL8eq6bipumqZjUIxxWTWfPf78L/76r3388cfj7ZHG5JNPH2ONyqIOw9Zysc6r+XA4nM7mw9F4sVhAbnd8cuo4jmHZqzixXQ9TYxUnB+1OzZaGZT978XK1jDvd7tnZCecsjFzfs/u9rpSyHfWm0+sg8huh4rxyPP/04lpIREzDtK0XL15wzt/80msNF4jQP/npe/eO7p6fXXU6HR0oJYWB+CJPmqbqtHvL5TovC0JRuxdlWYICoyiyXq9jWZbnuUFgEZweHe55njNZzna2x5Pp3DItwyBZzaqy1ELatj0a9BWiEtNlXGCMfc/mmLbWy2F3ZBBzNZtbNrZtGwn26qNXv/v73/vyN74ShmEURScnJ5sOSrBC1tRN02xvb/9s4weYJvSNQClnvV7D79uOCdNEqqoixLCsm0616XS+UfLDtCnLsobDIXBBSqnlfGGAk0vTFFQnUIF1Xde2TcG5lHzTs7dBpTBS2rg9TgDqsHCj0KUEHhSSMih4lmU5GAw8x1+tVoZxM6USEevlycX5xaXSZDpfnpxeLNdJnueubVOK/+7f/burxey3f/u3/c7O48ePw7BV13VRlfv7+0jpLE6U0kxyhPT52cl0Or26ukIEN1XFOcda1lXBmbRt2wvCps7zPDcC9+HXvrJYxxXjWlGC6fPnL+/dO2qa5tPHn//Kr/zqn7z3ftTunFxc7u/uVg27e/fu2dnZ9fUky7LhYJtQs9Pp7Ozs7O3tPX78mHFpWk4QktfeGp2fHFMD3T06SNJlO2pZlnV6ehocHfWGWxfnVxpTIeV8MmuYXKdFVdeO4zSsME368eMnJjUcx7l//+4qTq6uLtrt9je+8jXDJNQIPIxrIYXSxKD7+/vHp6eGTTqdjm2b1MDX19cIqcFg0OsNRqPBOk5G48Ff/PU/fzVf/s4ffrcoirxkjNdSSiJx1ZRVc92OAsZqJYUQwrDoqN9xDdRrdxmTruvOZ5d7ezuC8SSejvvjH37/B7/0y3/Gsqx2uw1DX6IoCsMwjmNKbDAbmLkA3CcAPKXFYrl2XVdIsX9ndzOtzjAMjKFb1QrDMImzIAhMM97a2gLbAOoe+oTX67VpmqZJtdb4T37/nxRFAfQQKKIppbPZrNUKPdcFscgGOAKtCviS345mhQYAqHBCPZffTo9WSq3Xa8uyWq1WEAQE0bIsuWRlXXGpuVC/+Zu/3ep0o7D7/gefrtfr2WyCMeas+Zt/828URRGvFkmSNQxdXV2Zpp1l2a/+6q9+8cUXQAaNt7em63Wapqfn5/3+IE3TLC9h3DOlVApmGBalVCokeOPZ1qjfiQI/tG3f93lTfuvn3lkvF5eXl9fX12+/+eXnz186UfTg4cPj42OM8Ww2A7acEDIejxeLZa/XS9P0537u54B+evny5Z07dyaTyZ39vW63naer1XJmWvj1V1/hTVMzxrkOW73JbPny+MKwvM+fPp8u47JqmBSUUiFrx3Fs2+y22oZh2I5FEB6Px3fvHi5X8247/MZbX2nKtCpyg5KTl8+pRXd3t/M8Xa1WW+M79+4fzWZXV1dXhknDoD2brnqDrmHZs3ny7oefTpfpZDZ9cXwsta7rmiBMCLEMirFWSmGkpFK2bffa4et3t+8f3fUd18K0aTKlhGVZ/cG4rIU09Gq1+Pa3vw2ZMYg5blzjKoN2zSAIQCEESNR1XUw0xOEsy0CxnqZpp9PxfX8+nyKEhFC2bTc1BwkbUK1RFG0+AXJx0NPVVUU2nZZKqfl8DtJPqHamaQo8JfxBlmUg/Qf6k1IaRREMYQMGFISkaZpCRRSoK6gNCiGWy+VyuUzTOI5jIYRGZLWKg3a73Rk8PzlZx/n0emJgxYrka19+raqz4/Pz6So9n8Wz2cxzHIOgQa/z5NPPKNb9dmtr1GNllq6W69liZzTO1uteq2URbFBqmaZvO+0w8mwLY8xZbVkWNkyhECL0/PLq8npCTOviarJYrb3A39/fX8XrsBX1ej1QI4DSYDqdjkYjKeXz58/DMNzd3R2NRlmWgep5d3e3ruskSZRGP/jBj5J1rBTqtnsGNoq8wcgmxCGG6wft2SKmhm3ZvlCy5rVQsuGMSSUVaphiUlWMzxZrhMnl1fXTL140UklNzq6uPv708ePPnzYNC1utpqmOj19ANkkN/JOf/IhzPh6PMcaUYs93Gs7WSYIxfnD/flMVvKwtakqBbcvHxNLIYBJxRSWxJLE0tWuJl1m1Sovjs+OL6eUqX2iMpvPFdL4QQiAsHMeq6/p3fud3TNOEAJUkyXQ6hWI4COfyPAf+EZpdQSwCkxM6nVbTVJw3YegjpNbrpbqd1Azp6c+OAtkMhoCkHKyRc25Zt6Zp3M75BuPDGDdNzRmrKgX1IZjNBMQVtBOAS59Op9B1BcAAPgHyrfV6HUURtMIAP5qnRRRFXIqScdNy3vvwD0fjnSQt4nWSZbllGFpkd4923/n6l3/3X39vkTSE2nFa7Pe7BClADu125LveZDLp9XrnZ2dMIYIwkkoLaWBqGoahse14GGNMtFRKIgnBSDT1ZJJryUe9YRzH7U735fEpJagu8+Fw6DnOeGvv7OpKEZKV5Xg8Ho3GT58/86PITlOJ0HKd6OPT3d1dqfHu/kGWZYZpm7b78JXXnj596vv+arWSgh/t7zQNz4pqHHZsYkuF4zhL0nyPYCZZnueIaEqx1pg3ikuBMZ5M5xSTMPK1xqPRyHXdNMkD17m8vB72Bk1dnp2f7+3umBYtigwhvL+/HwSBEN3j4+P+oAdDDBhjRZFwoQxiW5b31htvPv3iedTuvTw5nS6mpuUohbRGhCiBtEEQxlQhxPPq+fGVY+8hSgyqqqJCGhd5+fnzJ2+++aVlkg0Gg6qqvv/977/11ltRFIG+E3rooBYI4iNI84FRbrcjKDSC5NK4PT/jtpcDwfx/z8UgeYOuJmhzgFk6oB6G1Krf6ximgfI873a7SZIHvssYQxphTMLAywolpdQY2baNCSEGdagLk9NgBhXcJUhT8zzfyPDiOPZ9v9VqbWawQ6wkmJZ1Qw3DsqzFcl3kNauUZBhJM06WkUvf+bk/fXCw//xifTYrudCCNVEQ5EWdp1laJZoSTfDB7s7uYGwb/tfe+ubJ5dWzly9s0x0OtjAympr7gduJWnmel6wxTVvKmhjUcW9mE6yzdHu8xbT8+PHTYX+gNe5vHWmMe9v7NWO7R0etMOr2BkVWCsYX81hIfH493buzl81n42GfYp3GqzAMm6qQhmHb9osXLyLPtw3z+mrS7kQ1UyItesORF4aN4JgQapGt7YFpodV6iqhCGtd1rRCJXIdzTkyDEmyaRplkaDC6urgOPT8u4y+/8Wav03r65HGapq5J792P6kaMRodpnF1frRb2vBW6jkVdy7yzvff48eNur8WWpWMZo63RfBH/wjtvjQfRv/yt3/MMMghaldDawExwQghWAiONkSZYIUKzUn7w5GJvb5tSX2bpVjdSXCSrhHMe2Samtmc4J6en3rc806ZFXSVl1mn3KaLPnz+HERJN00BzXJqmlNI8q6MoWq8yxpjv+1ppUOKBIgQhVJWN6/igpa+qihCr1QqFEE1TQdJiWYbnOaZphqEvhDIAem6mSgDldDt45EZljG4HS+PbU1fU7RxQSHU33wc3DBNGAX9sqHtKKUGUGFQj1LBGKQQ1pDTNq6ryfbepsna7zaSK41hqRamBTSKEmCxXrG4k0UgqhdX5xVWTVnIs0yAdbm2fX106jmNaVlmWjmlRywQ0w7iQEmLHv218cV2bM0mJaRiGUNJxHNf3EEKI4CRLaU1Y3aRp2ml1n3z62RtvvMEEf3l2wRjbKLVhLDIIv6FJzTUtk1JIeKuqGnfaaZqajn15dd3ujqG18OrqCqA845IJCV0GkFwCgjIIhaZb0zSHw+EHH3ywNRpEgb+zs9MU2bNnz5q6bLVa3XaPEOLZjlKq1Q7ruv7iiy8MwyiLutVqQUms3+8n6RopaZpmFEWrVawJwRobBCOkFVIIwXhHaBqmdcOn83nomJYQ3dCxLKuRLI7jreGoZjB8niyXi96ga5qmYVhlmZvIQreHE/V6vcViAcYHY2Wh6AM2A3VyoJ9ghAdgv1tRGAgPbKA/Hceht8c7KaXyPA+iFoFFV7djVDd9TAA3gQmr65rdDsyGNwEHDAAHDqsMU1Zu8zUMSiWMMaj9oKmA/tvTrujx8bGBCdC2QjKC9PbuLrXMx0+efvL4SdMwjKlhGPPlMmWswkhhRwmT1bhmep6kT46f/ei9P2nqctTr9ttRlcZFFlOCMMZp2TSahGFo3h5lBOUA27R4w5Ik0Vq3W91W1GGN6HUHlJgGtR7cf3R0dC9OE1iKe/cenJ5dvHxxsrW11Wl3W61WfziGmVjQxrC1tQV9V8QwqGlsbe0M+iNoMas5Mwwz6nTfe+/96WyuMeJczmYLIFMcx8GIYkKElICdTMdudzuO5wLvvVque70+3PlkcrVKYk30aGu4XC/yPEUIffjBJ3BMTysKjg72toYj2/Z81+13u3mazSbXy9ms245+7pvf8Fzrzv7OVq/dbwfjQWfQjWzLoAQhrLTWiGAhsSLGOq1zrvpbO6VQ1HR73eHHH3y6XC4dy/I8bzgc/t7v/h5SuMoLxmrXczDGAH8RQpPJxHVdIHbOz88ZY0mStFotOCYBWuE2rwDIVFA2QWoOBXrXdaGkAmAUUppWq9U0FciobuL1zwo+HMexMILGERBKQTcTHDcR+oG+nc4PaGOjEIUpKJBqgKYOtOKMsSxLXD/AGDuO//izJ48evX55PavLarlcNnX5zjt/7t2f/jQrSst2+16kBWmaxjAJIiZphKxKg1quHUmE0qpsGCMY/fGPfrizNXZc2tR5Kwg934/z0rRdSF1t0zIIrerCpAbWiFLqeQ7UMKBMfP/+fc45OEVCyNX00rFcx3Eur6/uHt3/crf96aefconKoiDScrOsqqqDg4N1HHOp/bC1f3D08BX3xbNnV5Pp3s644WL/4O7jzz8zLOezx59fTheUOIiYJ8cXEmnHcQmiSmiCqGlgjYVhGEmeDUcj0zTdwC/Lcry91W63OauLonh5dXXv/oESnGA17G87offw1QcGotPraRD1y4JvbTsXFxdbo7Fpe30naFh1eX2hEBptbYdB5/jkbDjodttB6LudTksIbppWVlTH5+dZUdUNy+tGa92IBiOsMHp5fkUQfri3I02HYm1g6+nTZ1/9er+uK8uybNv9/vd/8Gd/5c/WdV1mOVIkDEMuGo1kf9AtioKLhlDUaoewnuCnLMsCmRsk0xsmP4qiuq4hPYLBgFCLhxANvrbVaq3X6yDyDXB4oBcBVAsJGudcagXGhzG+e/cujLIARwicKrh0oOWhx6MoitFoBJ108vZUULD+6+vrrdGYMVbXzAtot9snhCRJUpYlxtq1Tdd1V+tUE0oNbVlWLZqqyLTgUlU+wocHO64faMNFnhcNtp48/SKN44RLNV8UTW26nmEbs+t5xWUriLqtdllkSgnQyzHWgJzKts121AbRFtw/QOeqqs7PzwlFg8FACLG3f9BwNpvNDdsiEjes2t3d1VIMBoOrq6soii4vL4FpgjLHncODpq7S9eqdb33j8g+m+3cPP/nssRsExy9fnp9f/Olf+jMffvAxZ1JjTU1bS6Uk50Rqgh3f6w8HCCGKidKaSfHs5YtvfvXrP/mTH9mGcXx8bNum4DXjJVL813/91wOYMmSa0+k5NhVSYrQ1fv70eHvnyLKNMA24FOvlCmmSF2m7M9rZ2V4uV6LJfd9vt6OiyHeH/fk6TYqq4ZpLQYmtsVZarZP83F4Mu11Kzbwqup3h2eTlixcv2p0epeZ4MH7+/Pnla1dRFEJwFzAISik4QpffDtJnjQA3V9c1CMOHwyFkUfBf0KdjjIHbB7/med5sNoMDeqBqFQTBYrHwwwD/6Hf+AXSCbnRQIBBhjCmk6c8MqtwMTDQMg2KyqS2BvgT+F7YCiMTA6DcNqXmeb41HaZa7XvDDH78/nS0Nw/ri+fF6vZ7NZr/+q3/6w0+ezOPSctyacc925vN5XpWirluW/PZXX3/97gHC9DqXTyfrjFv/9L/+Zx9++OHf//t/r8oTzzKj0EVSVVXleIFhWGVZm0TDGRpxsoKirOu6hknu7B7Ecfzmm2+Cg+90OrPZ7PDw8Ec/+tFbb731k3f/ZHt7WxPsOK5l24wJII0xZ51uC5YCWhDH4/Hnn39+9+5dpGS8Wjx6+ODs7MyyjMl0vn94+MWzFy9OzwgxOedZniuFYDqV1jrNS601w4JS6piWaZoYU5MSg1JAjYN2N16uRuN+3ZS+7/uu1elGn3/x9GBvu+16d+/ebYWd68lFlk3H4/HWcJCmRbs1TOKZbdtpnjIhXT9omoYQ9+mzl8fHpwZRlmGCmH+dZJ88flbWTVrxWmjHMqVWQkmJJNZiu9/ttcPdbmu7P8C0KYpiOByOt/fn8/lqtaIUf/2dr21vbzmOC/3HECoh92CMDQajLC1gDPd8Pt/e3obMCZJ0OIdoM9dJa31znoTW0GsAxsZvz46yLOv07IqARwRZHoifYR1BtaRvT+qGYj98xOaSUNSCziR2e74YTDiBK0EqAC1KQRCsVgu4xJMnn0dRu+FSSKm0rprqwb07q9WKKeT4UafbH4/H9+7fPTq4E/rueDjc7rXWs4t4dZ0nk3w9LeLFP/sv/+Ef/avf/eVf/Lpn4eVyPl2u8rrJy9rAxLXMYbcFQhbgceHYBnw7WWUwGAAnspklDb2zdV1vb28HQZDnBTWsrKySPDMsJ8uK7d2d+WI1mU1vZgxhY3I9A5q6kVwRnFWlZdvLVfz0+fPPn738/PlxlldZnq/W6zzPm6YyTdo0ldaSYkmQwEJhoQTjkgslOMF4aziyqGESahjGweH+nTsHgiutcFWr6Sxx3M719frdd9+/uroWivf7Xd9rn51enl9OgiBotV0tpGDN4cG+QTTnDUIqL9KqSMPIbbjM64ZSKjlrqgIpgZRWShmE2iYedNodPzQkIdgsGpYzViFdIemHLan1Ko6TZN3vjVtRJ0mSn/zox512JKVcr9e9Xm8jf4NBN8vlnHO+t7fX6XRGo9GzZ888zwNN+vn5OcRSUC6D6hK0xeL2RFrQx23yclDxGeBvoZdZ3J5pAnZd1jfnkoBqFfwHFJOM27PJIEkHzwqHJIEpNLdnhsIXbDLLcrTWWVYAfCaESiGSeP21r7710YfvCyEqwebLVegHvmlXRcl5E0QRq5Iyz+7vj7lG62nquu724M4Xn/7UNsxOqF7d735/elllskgLionrukWedjrtVhQBkywFklxUVdXv97dG46PDe57nwbBMGL0LM2TCMDw9PZNaXVxf7+7tnV6cG6YphNoaDA2Er68ncFSwkBpp7XgW9OAiDM0I/kcffXzv3r28qrq9wdXlpGmYUghjXZZlEPoAdSI3xBppxbkUvJKUEKQ1UdqkxLOdqigNQiPXb0cBQujly5e+H0zmc5vahm1xqYbdNmnLDz/8kGD02muvDEc7/d7QMEmcpthAlm06juM4zuHh4eVkmqap7/mPXnnAGvHuR0+KLHl+cho4tpQyDEPTlaVMayYwIoNWC0fIwUbOy7LMq7qIQjsIPLvAMEZzsVj4XsfARifqLBazJ0+e3L33gFIyn89M0zQtWtUyDAPOOUKa3R4pizF+5ZVXVqtVVVV7e3tAOG5knD97roN1e4okqExgDASUx13fuznfFyEEXcwwUQ0sFSpUcGYFpF3QEQ8jRoBAAZ3yJhMCxe7mm87tWROg02u1WvP5XEoJcFZKKRrmu/av/tKfURJxzl3HoRqzuqmb0rZvekIqZP7wk2cfv5hME94b7I2G291Wa3/Yubs3QHWap8ujwzuO40QRHDPFlsvlyfGxcXuKTRAEUisv8B++8mh3f09pEYRep9sqyqw/6F5dXQHuYYyB9mUwGKzTpNPtdrtdz7WLouB1A2os4Edc3yvzoshyJaTvBLPJdLlcUsN696fvb23vFkUFpTXbtoXknu9CGgujN5um4pxLLm5iEcIUE4PibqdFNMIKJ0kWxzFj7PXX3phNF/fuPhBCXFxctaLO+dllwYS2jD/8o3/9j//pP0EIRe3O9dU8jlMYChlF0XQyz7Kirms4/vXgYB8TzQQP2h1EqGHZnuOGfhB5vk2JTUleFrJpfNveGQ5GnQ5RWmt5OZ28PDtN01xqBYYCCiPJVSuK/uW/+JcgLYIqTFUySizORZKknU4XDvRI0/Tq6goMaXt7O45j6N4EBQnUosEAQEkMTZTQ/QIzHMEt1nVt5HkeRRF8brvd5ren2oAyHITGMA4K3jc0XnqOS2/n+wDIhR9B6xxQpKCZJ7cnd0kps1oM+nsY604rWC7ns+tZUxQtz2iyy0WcRVGbSWEYhMsqSxobjnwNQ980keDPFgwtrht2aVmO6ay++ALxulEGUbhtBfaQeOv1GlN0Prmq4Qi2xcqyDIlRzRrb8+7dv29HAcNyOZ/OFteUmt1+JyvS/cO9pmnW1yvDMhFCBwcHZ2dnjx49vLy6KsoiCkObGlVRrperbqvtO263FV1enve63Xa7e309TeOUSNKUsmbccaP3P/hUYlrUtWS8YrHneYxVw60e6BZK1XCFDNMzLU0JazmRIahjaUw0UczzApNa3SgkluRFfvzF56FjrS8v+91WFIbHL58TgzYVY9Qzg57lBpcXV1vb426/dXJy4riB5qLmvN3tIk3WWSEV7na7V+dneRK/enQwmy688d5scq2Vpkg2RdJpOVfzFaZWVma2oR4c3SvqdhIvS1Y3OW9caTkeNawg9JIiK+ts0Bt7XhDHseLk/fffP7q737Cqrli/P9LKVFL0+8OiqBgTwI+2220YJgweCqAUmBacVAYQC6aEAiMEZBH4NWhpKsvawBiDVgXszHEc+Mtut7tcr4AEBUYAKGVoM6KYIIRAAwW2CxJASOqBU0jTFI4jAfUAIcS23HW6LsoMIdQ0vK5rxstX7zzM06yoylarVTcCIRJ5Vrxc3TSzMtmKAi1VEER107R6rpK6Zg1jTBHiBG2NsUFQVTVA+zdNc+fOncvLyziO/ch3PfuVV14JQy/P82Q1K4osdLwoioQQrG6m15MwbJmmaWCzKArLsBarZZIkP/zhD+/dv1/XdV1VhutB8yHkRk+ePNnaGl1dXRVFVdcszwrXtre2tj59/FlZV03TWK5z00GrdVVVgeencYKho9w0tcJCSYyxF1rDVtcWhCJh22bQ7mnDObs4tTgiQgsh6vXKNE2OBBJ4b2vHpGSxWJzN5slydXiwL4jx5MWLUjDTwKOt7ZPT81G3Y1lWmuerVUwwtgwzS1JCyCuvPJxOkziO16us3W5fX5yGvkcIOp/MA89JqptG+LzIEDZeffholSWnV2dZWlxdz6RqirqKwnaWFraZ7e3fsR13Op985zu//X/+j/9P8/k8DMM8z7WiSqmmoet42Yo6sFYwLg/fnvYEwX0jbea3Z58CJICoC2Ec6hfQ9bFYrAzo1oNJjTC7DAI0/AOkzpD0AIMNWRtcFYxvc7GN+wRNA/w5FOiFEFVVxUkxHI4RVoZhhKEfBq7sd7rdtlTItu0qaYSQtu1qIR3LzqqaSxF4YVFknU6nZhXGNMniqmxs22GCSylpU1PTSLJcSQHUAXw5jlOryrTovcPDw8NDxsvTF19wwe7ePaxLNp/NHNsLgsD3w6pqyrzyPa/My6ZpAHus1+unT5/ev38/iePri8thf+C6bl1XG1Td6XSyrKhr5jiO4Pzly5fw+FLK+Xyuta7rxnddAPQKacc0oUKtFVZ1ZZqmZ3oGIUJJ6OQWiHzy+bPZcnY5u2aNxEgx1pimSQjyHX82Xe7tbFOMjrb2VqvVyfOzh68+RAT/8U/ee+3VB3fvvfLyxTPeiPFoQAgeDXrgk8q87vZ6s9kiiVPfdWSkLy8uBGPapGHgGYYhysqi1A8iy6KWbSNshJbR6XfCVvDhh++v4tz33boS7ciry1p39HKdIWJ4QVTW2b/473/zl3/ll4q8ahrmOq1Op6W0sOwRawQ0DSOEYP4rbADgPkH9CZYGx31A3AZuBBgoCOZN01xcXAghjK2trU0r5snJyWw2e+utt6BU2vCbc6VARAflrJvB41yAR9kM4tnMD4OJEep2ajjAA3hVg0G3aZqGc42pQXGexOPxMAxDqVi73b57b/fzp8/rmiGiDcOIWgHjUgmNtOE4HmcSI5RluWW5rJFF1Zgm5VLM4xW0ZYmGQa1MCNHtdr1tt9fv7Gxt/fgnP3Rty6DUtVzNRZZlpmFPJhNCjHa7e1ORM6z5fN4b9C3LAl4sy7LJZOLYdq/XS5JkZ2v75PR4PB6DNwVBlhCi1w0oxpfX18Awb8SvnudZpglacSY4FISLunYd37wZimUKrjrdjuH717PZbJGcXl4xyXQhkbak5JZliLrBGGc5z9Jytpi7nmVRh1JTarLO8qJhrd5olZR//MN3d3fGTMiirDESgjUYA8diCMl7vU5VNZiS8Xibs5pqVRQZMQ3Xda26aUpu27brOuskUUoZjlvFsdSKIHo9md+9ezjotTzTvMouWSOQ5lprjAkhxve+971v/dw7nu8xxrSW09m141iQpTVNA1Ui0GNsJtrB/C/btre2ts7PzwE6EkLAhAAMTKfTNE1brRawSYPBwIAwDT06juO8+eabABlN06yaG6kf9PVBLwdk6HVZAX0AHmujC4REDGqs4HE35JTruoQgP3Iff/Hk8O7RfLbkvHhw/83xeKiQ/Pzl1Xw+f/PNN58/f+n6zrvvvusFUVmWpmm32mFR1pRSjAkTXEgNQiwgOyBq2JZlGyYIl6Cl6+DOnlLq8ePHhJA4TSyDttu9rKiisEsIsiwnSTJCyMX5JWwtwECz+YxzTiwTKDrG2CpZYI2m0+nu7i5CaLVaWJaRJAkM6tne3j47OXn99dc/ffyZUopaJjSvAi0A46673S54Zc/zpNKO42ipJDOJbZ9dXz27OEuSTCllWjRwRhhT2+ln6YqLlRKKUoqUKKVuGtZQMg5vTh24vLyyLGu1cjqdTuvhQBOD2mbB6o7vtlsuQTjPc99zFstlVlamYREm54srSlAU+cvlrCxLrZRJDaSbpMwNm0qFJGeGlMSg09kcaaOuxHqdYC13ht3BYFA3JZciTVPTchBBQpbf/e53f+mXfqlhlZSqrtlNdR5zSMNBPzQcDsFxAkSEPWwYxp07d2CVgDXnnENTfLfbhSrPTSEqXRuDwWA2m2GMW63WaDTCt/NBweBA/AeVAygVWLdnPcHIBtM04Tx0kOJDs+9mKA+wr0D3GIbh+35eZLPVpGZICLFczm2bXlyc9YeD7d295Sp/+fIlzAwaj8fUtC3Lcl3/YG8/SRKtUVEUlJhKIYwpNkyMiawZVGXt28OsRqPRcDjM83wyua6qCmPdGwwkbwObhhnLb5PEwXiU50V32LdMBwgKxphj2bZtU8vUCMVx7Hue7/tKyCzLHNeezWZRFMRxXBSFlHo02vriiy8k56s4hghe1DdnjFPTgvwUUgRoxSamObmeEUIQIXbghN32aHf4x9/7I4xxN/IODw/vPfpmkvBK+pPzL0y8zOPzLI5nVYkxFsJYJmmTVdu7O4TTmjOERF3HlmW//8GHcbo37rmDbitwti/PJ+N+l5W5xpgauNttE42oQQzDilcJZ3UrirgQquJZ1biue3l1XdXN6w/vp8tpmWcY4zLPEUIGDDzjhWsoJKTt+sPBmJpGUVUy5r4f/vjHP46i4J1v/HwQhFXVSMlhJh7GuNfrAXW/4XnAisBZAp7cHKATxzFookF/DE1OGGPf94VkZCNTgoNwoFwEqBEKVlC0BC/4s7opYF6g0gUkFNQGoC16I0uBY4fAjGDAnVI3ZzUbBolaITVInpfAkkApFUZJgYrPtu2irhvBFUaIEkIItUxNbuIpMBRA/MIVIZ+TUlZV4bp2v9+PoshyPNN2y5prbHR6A8cL8rIu6koTTAyLSQEfXlUVge6lsoTR+t1er9vtBkEAAnLQPoLcBCIJoKtNIw5QcjC0ApI8fHuKH7kdogY5pZDy9Pz8hz/4Qeg6kWMdjnuG5n/7b//N/+n/7H9e1cK23O3R8Oe//uW3XnvQsg3cNIZSFqGVqIUW2KAIIcWF4gJj3AielZVpWUB9+74ftYJOp4OQhjMCPN9ph4Fp0m474pwbBjFuT6rWWiuMsixLsrTdbpvmjVjb9wOhVc0apRShCBormqbhihdVDqk6Quj58+e2Yy6XC+gzc10PqE0YW2TcDtoGBweD8TcN8vi2vdb3fTBfMFOQc0BxByFkwNk/kOvAj+HjpJRe4FdV9bOCOsuyYGYQVJsIIc3tAGXf92HmDucc2NdNHz70TcOsRi60bXtFVsSrVb/XSVZLrfXO7v5slURh9+py2um04ABFs2pizieTyXK5arfbTcOQJtigpnFz3AmMTIFLw7u51e+llmXd2dujpgmH0FFi5DXLs9KyrOOTE4wxNUiSZqA6qKoG/L1rGkopoZXv+34QwE9lw8q8aKqaGqTT6UjJsyxZr9cg2g3D8N/2D2oVZzeyG8a5ZZlwwjt8p9PpnF1ertdrRLBhW5PVIk+zyCRbrfHR/t7DR1vT9frs+eernMznU8zXGU6+tHPnzmv3x67zr//4g3kjaoSIY+dVGYW+FlxyhTCuqsp2nels9tU37zVVsV6vEa+QrDHG8Sp2w3C1zlq+s1rHhuFaFnFdO025YZrwlinWlmVxIT7//POHB3v7OzuUUkzMi9n62dlpWVT9dt806YOjhxdXs6oqoyhESJ/M14apXdd9+fLk7OwkDFvQILler7mrYf48ZCagQYajMqC+Y9xOZgRScrlcQlMGhCCIz2maAmBodyLDMIx2ux3HMWQ8AOwYYwcHB2VZaqls02qqejQapWm6zAuEEJMNUAMgWYWDV+DsDhAKbd4K3ATAZIxxU8sg7HT8QZXwO3d2ry5OsO0MBlFe5aKsBS9cIoRSrCCXZ6f7ewdOf2i5nkDIcr26rouiQJwajmkyC5xlXddhu2XbNmUMEsDpdOr4nlLq+HIKSMPClDUVRJOqqqhlN02jJBIasbJSEtVMAFmW1hzaqanlaWwqZEgpvVarVtgyyGK9jrO03W5LRMfbO2VepGm8Xs/vHtxN8sxwrDjPsG0USWERatk2kxJbpjZoxRhCKD47u7y8lFIOh8OyqmzbtFvdnus8ePDAtc0mLx8eHj7/4I9++sGnHWoxnobUFHnCDNQPzHe++up/990f1dQiipZ54Vh21OmyqtZaS6mzVRpF0YfvP+11W7zkoatboVOUmWeZHjFp2M6LrGmqpkmDMPRafi3kYrU2KHakqimyscYYN1yczpatdi/yHCQKzxCsqUnQygqOsFkxZRlhXQlWxut45jgW59o0bUb0J5989Gd/+dur9dQ0vFY0mC9nYSuyXQe4xSRLEUKtTnu1WmGEwHlBnIRedughgzq51homG0PO026365oZMFcN3I99e0Io8KgbI4PKG/Cgg8FgtVpBH+qGkAe8BanS5swu8LibiTxaa8ex4ziF6Y9Zlu3v7yOEGGO27SolpJL379+dz5dRu98ZDK6vpm+//eY6zU4vL6FpBPw6ELEQRIbD4WZfEUImkwkIlNI0DfyoyCvLspr65vzJuq5hSLtlOkC2waMBqNc/MygQEkxocoAF9Z3Itm3FBZB/MHKC88Y0zevJZZJl27t7geO2A182DLIreUs8gfOAEPTX//pff/78eRzHmhqL67nrOo7rNU1NnZ6QmBJ9sDfs5FlT0Y6vHVuWWRoGfSsXUbslGlSzqirLzDCUlLCXDMOoleacZ0VTlte8E+y+eW+dJlhLw0Sz2ayoeH+3BxVzSgxoGYfUAhXYFVhhg3CpEU6y/Hq+KDzbNnRelgBIlFJF1SRJ5jlRVWVZVliWU9WFUqqpuZLo/fff/4v/7p+3TFyVAmSyoB7GGMdxDGAUVJ62acGMWFDcJUnCOYdfAC8GlBMcZwAp+GKxMmC+EjA1UBZqtVo/SxvFcRyGIRx0aZomNDvDj8CskySBnB0U3QAvoLl+Iz2BpZRSGphEYZvz52DQtm17vq81Ho/Hk8lsOruCTfby5cvXX3tjNps1QrquG2d5HMdwkN5wOMyyDG4GEDfU/fM8r6oqz3NofIPjIgFdQOERoDBM7QfEHEURjLfczLuCKdTwaKvVCj4fIdTUnHHmuu52tFflOVAfYRhy3liGKZRazmftdvfOaFsULE4TSggxTc/zkiRpt9ughOh0Ot/73vfgzD+XWq+/8oho9Pj4JPR8QdV//B/+3d/4//xGrz1ueUaeYcHy2aowDLtIii9OzxqpTMuMgl6v12t1O3meg44dxnmGYWhZhhZiOls9f3kcupIg5Zgk9AemQ4CxPjk5i8LunTt3PvrwMz+MKlYbhPqeK6sGY8y4wAa9ni9XJvEdA8Yu5GlhEqwUchyHEhOOImpYDqS6adq2YzR8dX5+2Yr6BJtJshYKGtpujuSD6g9AcJBoitvDawChwsASIDGDIJhOp5CLg2iz3+8a4JygA39z/g0kH5tmN8/zDg8PgT7d4NE0TWHpwV/CH4IcGHQrYOgAasFTuo5tWbbjaJhxB6661+sVRWVRZ/XkKTVtqtBkOu10u47nxmk2Xy29KBoOh+C6AJrAiAsoeSOE4IogmgZ1jBBiPNqGMSzAvQEGh30IKf8mjYMW1rquIb2DIa5KKTiuBQ63pZZp2pYUzLIsfzRaLxYI04uLi93d3aouHc/dGm3XRVkWxd5oq66qLC8kuRnKegPoPQ84wtVq1QrCXtjijOV5HnZ7jufN8/Tf/w/+3nq+kpxFvqG0kJLHWUooLRpRNNz2w8gPeq1OEIZaa0aJZ1tKKUJQIxqRCs92yiJHdd5rGb2jkWkRVsSMl9u7e9PlynacbrdrWz4mpN3tYEwDP0zT3BRSKoWQ8lyrkbiqa6VNqTlWQkjZCf1Wq6W1dByHVZxz7vsu4xjObMiyzDC14zo/+P6Pf/3X/iKhxLRuBnhBbx1khwB2TdNkUgE1BMfBQUCD/Fve9rfBcsEQe8MwijIztra25vM5lN1h3wB3CjQe57zdboNpbth/oOthH4Cv3Ux2gJNy4PSZLMvgsJLNbH2EkCY3x4KJur6zv91pteuyVAjB58fZSiusMNoeDheLxXQ+c21PcTFLZnEcA5r2PG8wGCilxuNxkiRVVc3n89lsBvcD2uowDJeLNTwkHJ8KfCSYL+iYoF8KmIrBYADHSsBGgiwS2hggFHiOU1UlQsioKqVU1On02m18587J8YtOr+057tXVhebCs71Op+PNPEwol4I3DDpyQPkFktlut6uFTFaJRLKRTUDbtdbIsLrd9u7uncDzJ9P5fLrI0kQ7LlfKDpG0itB1onakFSrTVEo5nc+k0pDCg+bXMS0lJZViNlt5jopCz8CN48r5fG5bjlZqvlyOB3YaZ4PBIE1zrfDe3t6TZ19gosMgKOu1pbWihLGamJ5WkgvBpSiq4ujo7TzPMXKqqiqKzLSJbbtSSkqxlAJj64+/96Nf+3N/gTEWhW1EcFVVUPeG/Q8WqZQa9Prq9nT7pml2d3ehHg5updPpMMbu3r374sULkMsopVzPNs7Pz1ut1uYMJJi7lOe5vB1CDk0k8AoBa0Irfl3XH3/8cbfb3d7ehvZThNBwOASmGu4jSRLwVSDV812nYkxKRSi1bdtxnNlsZjm27QfUMLKykEoJoQ6ODpeL1RfPX3T7wyTPDNNSSgNwtiwLmvyDIICIA2fmBUEAZieEOD8/Xy6XGFFCyNnZGSgLLcs6OzuLosh1XSg4gWIQBqQBgwEulhASxzGEJEARhJDValXz2vM82lBANaJpTNO0HW+6mLeC8NXXX1svV8+fPt/e3u12u4v5UimFLUthBL5hI4q9qSqtCq0lNvFsNkMItdvdi8trz7a4YIEfRh3f9eysSIUQEqluq2VipAWuJS+rSiipuEAGEVLUFcOUaKnqoqSU2gRphOqKMVa3AjuSejZfRp02obTX6UABpSzL1Wp1eHCvLMs7ezsvT844F712uFpnncinpg08EZxTEHj++fn5zmhY5UwI0eu16qbUGgFqUtqoq6bT6X/3u//m137t1zCmUt3MTQ+CAIyquT2hBmhOYFog29lUyzeRdjabgS+DPjmttQHpLeCGLMvAF8LHQXwvigIcDwjwIOH43ve+p7X+0pe+1Ol0oJqvlIJEG0ZvAgMKzXvglYUQWsvxYLhYJVqpTrtNCBkMBjVrTk9PKbIsyyqr2vG9999/v9XuENMoikJq1DBu2A5sEnDDQRBcXV0BzQRyQxhUOZ1OYZCJ4ziW6QD3BJsYIDUQ8uAXYX9vqrjwv71eD+aggGB0A10My/IMQ0mFDZqkuVKqEaJsGovSVtRZx8vPvvicYAN7dl5Xe3t7GJH1ep2WuWHQTR4JpLTjOOv12sQ2QsrChmZSKbUoZp5r2kS5jlnp0tDUoMigSiuJNa3yrKiF1ohjzTlnSlqWnVcFItjzPEwJ57ypaoKEYVjrNG1HViv0et0hodadO1tZmVu2c3U1aUU9qfXR0ZEQarVatdttxup24BdVjRs+7EZFXfuu4fRaeV5eTxfD/s0R2ZBKNrdHwpUlyIoRpYRzKSX/wQ9+VFXVX/2rf6VpmsFgAAQL0N6O48D5g57jQmWHUgoLAnYFzQ6AA0HfBEcBwgFiN+fcAIwDQSS4PVBJwgVgZeGAZahTf+lLXwIh/SZQbiTrYA2QzUBaA8iDEGJZdlEUT58+tagBZ+teXl5S0xiPx0VaV9XFbDajlg0dtJPJ5OGDVxCheVbA7UEGVlUVPAO0XMF4oMvLS5DHbzhdy4TzUxBIW6qqgn28UZOAxGYzBAViCqDGMAwh4gB+r+vaMCyEFGNsuVgjJXzf79g29BljjB3X9VvRahUjkzKsMCGEENs0GWOBG4EjgVgPlC1CSCLtOLYUrN8dSC6KvKEIF3XBFG0Nt4s4bZioqoaxuuCcUkqIllIrKVqtcBmvfc8zLYIIEUqWTW0axAhsopFF6fZo6HkW6MGVkMBUpHECridq9+Bo8SwtLcsKAq/dblOaKZXVdd2NPNt1bNct88I0TSVkKbltjzCmRVkzxsqyHm8Pyoo3DFFKTZMaZotSXFbpxx9//Jf+0r+LMZ7P5+ACQVoEPTObDkqobgBUg5YmiG8bPTH0fLZarTzPF4s52RSNoHEJcBvwAsPhsNVqlWUJXSaGYaS3p15DoIfcAvgOGBENh8eJ25NAOp0O1J9s2y6qchEnabL8/h/+jqzz+/f2w1bQane7nX6d5nmcYIx9P6zKBmPcbncpsU9OLy6uri9nkxfHz03bMC0rLwohdV5UUqGGCct0Xr44uby4JtgwqGUatuv4ju0d3DkCKSqQbf1+HxSK0IsMpKxSClS3lNJ+vx+GoevZjNdZnlxPLl8eP8dEmxaVimskq6qAc31c1+30BgqRohbLOJ8tk6rmUdgr0opVLFtl86up1sgPA8Oxo15bKNGUJZEqsGzftB4d3aVK87omVDWsQpRUTRnnccWLWrOyERrbi9lcKTUYjZUmO7sHW/1hYFnt0DEJf/DggPHi7Tdf9W2KFdOs3O61O7Z5NB4OnKBnu9u9Vug5vKoMStM4ZnXNsnIUDttumxVNkqTHJ2erNEnrchavLNuFMbOeF/R67XbHbXddy1Sha7f9VuB6TdM4ns1Fk5fZcpWlpdKWm3ImJcdYU2paloc01YrYls+Y+Pjjj7WutrcGjuXPJnOCtW0R1zECzyfo5vgbaAKBrABsyTCMPC99P0SIcC7X66Qsa0pNhMibb76J3/vDfwQafcjHCSHgmYFQ3NRP/3+USmDB0D8A+p1N34iUEtICsFEQizRNozHqD4f/r//7f7KYLu7s73/rF38hzTPP96uqMjBxvej9Dz5axOnFxZXERCpsukFZVHlVvvr6a7PZzHGcLC1c17UsB4xMCDG5uoZ5/WCI0HnSbreBYwOntRGqQnsx3DZMYYXPgdFChmEgfHOwDpRz4Rmh84tgA8YGAVwBfXi/36+qSssGIeR7HrBpnu14jts0TVbks/WCc24i4toOwRg+EBOSFXl8e+ouhB0gSiGNa7WidB0rocq8IBSXWXp4sG9SVNflIlmPx+MwDJkQrzx8+Pu///tHR0dlWS6XSwNbnu/KJh8OozxZtFph5DkEa9uww6DNObfD8HI6l8g4vZxcXl5JoQfdwd2jAyGY7/unF8dVlfVH4zffeHvQ3/3ud3/0k/fep4bsdPzXHh061J5cV2cX10f397nI89UKYwpijzSNTZNSg2jNpeL/0f/mf1nkDGnL930uKtOkhmGw5uYUjQ09B3J4YMc9z+NcwumJhJBWqwUToKSUDSsMaJejlC6XS3iXEAevrq6ABYAoCbASBM++70PJFbjP5XK5vb0NWHM2mwH7FYZhu91++fIleHXTNLkUFxeTz5++GHZ7W+Od1SoOwlAIhTFdLFd4lRZF8eTJEyHUcHsnTvLr2TlCyHKdi4sLwNFwFi2Ey9lslmWZ5ALiNeD6ra0toFHBqkAYcHJyspFcua5r27YQAnSfm8kOQMgLyeB/AZBAngSo0TIdOO+e3p7Um+c5bMt2FCCk8qKghIDcG8Dr1tZWUmZNVdu+a9t2nmWQOFZ1bVnWzs4OtHDBFur1eqAhb7VaGmvLtXgttnd3nj9/1usPRqPRJ5++b9umYRjT6fTy8nJnZ+fs7GxnZwdAzi/8wi88/vQzLbnhGMN+Z9B1LcvK4uXW9pZomB84GnvTxbqoqv07d2fztZTScbyGVc9fPIU0pawTjeRXvvL23t6dPJVCMillFHlSyihsI4mX8ZXGGmPNGCNUY6yRRlrfUH6UEELN5XIex/Ggv3V1OacUM15j7CiFlss1lH76/f56vYaBoyAz6vV6y+VyMpl9+ctfBiH9yckJjAthjHV7kQEoCpTMG6VSp9NJkmS5XIKAA9ADpTQIgvl8DiAVWCuoua9WK845MMaQrjLG5vM5ZMqQgWFK4qxyXbfV7SCCHdflUiBKECXj8VgIdXJ6vrOzw5g4v7oKow4gNttz4dJYoyLP67o2DXtellCuhcBtGAacDoMQarVaSZLAUJokSXzf397ePjs7e/To0cXFBXhHyFegHAXoFkIEwje83QYybtoNWlGn0+lAbxaMoAIJFWMMG3S1SgLXCYJgHaedVrupeZIkVdkEni+5qOu6LMvQD4SUhBAQPpq27bruyckJJKmQ7UE7TdSOeoN+npYUkwePHl5dnmd1rpC2HDuybPAmcADVsN9br9f7e7tZEg9HHdM0PRvXeTwed5USSPpYy6gVTGYzx3O54hjj5y9fVFVFNOJNYxDU6w0Mw5hMZrPZ7Nd+/VfefvvNpuGdVgi7kRCMpFitYsN0Fstlv9/3XS/0zKvLEiGilUYIpI8WxpQ1TbvV+85v/au/9bf+VtQKpBRwQHen0+l0OpAPQEzf2tpijIGdQIje29sDVwjoE3asYRhxvDL07fA76C8DLQgonGEt4N8wvwkIVcg2oA8OPGt9e3AY/Bd0mVA5gHwNMPJyudza2TYMozccMMbypoIIlcVJluStVgsZ1vX1VGPkuu46rSzXgU+GaqRtu6ZpVmUFklOoJkDRMkkS6PQghGxvb5dlubOzo5Rar9eQXcGkfdu2AQCAmwQKgtxOPd8I8jctXQAeIMjom6E9N+dPbjoLwAPleU7ITdMBEPJN0/itmyIKlDbgA28OFRHCdG74E7giQgjOpMyyrCgy23Qn8wlCem939/LycrFYfPWrX+YNA+IW2JL5fA4Nk/P5fDBsx3Hsdbomluv1kjHWDnyE0Hq9RkQD1lqsc0jAwVBardaf+7Vf6bR7k8nkw48/eOutt66vrwkxLENvyEHLIGADNWcAe0yiEEKEIIWQVghuHtYHY312dmaaZq/XA+UQpMJSMH17lsHmFK+6rnu9HpgWIQbUWSA3gsS93+8jRAy4PFQL4S1at+eJ5HkO/At0JmGMp9PpeDyGUXVg8pD2QuUUslqg7oUQjx49At0JmH5ZNT/4wY/qvLCIgRG9mlwORiOlsZIoy4p2u71cxYvFYp0krGLz+bJpWLqO3cAHdUxd10IoqIxt5iYcHh7CTNOmad59910AN3C5Tz75ZDgcbiRXQKPCbgQ1MQwL3t3dhT5DrfV4PL6+vobi+0ZVCMhyvV7DCQ2c89VqBSMpoR/w+voakLdEGiG0TuKm4VIKhFG6TpqygpcdF5nvuK7jenDsWl3DwWhpmoJ4YIOaHjx6tL+7e/LyxDGt9XoF4esXfuEXPv300/FwlGXZN77xjdVq5TjOvXv3ZrPZZDLZ2h5RiiyD2LbRaw8Xc0V07tqOkurO4d35clHVrKgrPwzSgs3nc6RVK3L/9t/6G4apTNO4c7C3sz+2LBNjUjVyPo3TJCcUtUI38B3TctdZSk3DNBDFUMqmjHGoYbZaoWFYdc0Y465rU2ocHx8Ph0PTNJfLpWm4SZxtiB2lFDDWq9UqCAIQzwshhsM+FJY32rT1er1areq6JK7rwimfkKICboN3CQkElIgAvYEbgGml7XZ70ycKNdbFYgEbBRDhZDLBGK9Wq+VyeXZ2RildxmtEiGlb55cXYRj2Bv2yLNdJvLu/f3l5CUQVgFdy22tvGIZnO5Jxx7QC1zu6c4CUcixLS/mrv/zLUKUAoSHcQBzHaZrOZjMgikej0Wq1evjwIRQnjo+POeej0QgopDfeeKMsS6htDofDDSe3oT83HOpqtQIC1XVdOLkZQI5lWf3BSCpUVk2Wl+1u37KcvCykRoQYURRtbW1prcF2f+ZAJgJ4C6arCiFGoxHUQre2trSUT58+vbO7d3R0dPfwqMyLe4f3siR/7ZXX+932L/78tz54790yTy/OTo5fPHMsYzTo+Y7TsKrda3ueV5Z5FEWtbi/NC4nw9fXUMG3HD5KswJjmaUawxli+8aVXAt80TbsVdRzbcx1fSYyRSbAVr7M8r6SUCCnQLSRJoZAOolBrSbHR1Mw0bgRltmPleRqGvm2bdV3btvv+++87rrVczW3b6nQ60PEL7DXQ6k3TQDcI6MuklBcXF7BLZ7MZvHQYUDcY9gzwcFAkhIInBBHLskBGChBt0ygHWUKWZZv4DqwNDCqH1wbFfcjrd3Z24POvrqeUmpblOI5X12w48k9Ozvr9vlIoTfMoij57/HnTNIzBwaOR6SCtdbfbXUymZVF4nue73mqxTOOk3+9TTBzLTpIEToYE7AFiWzjvEYLIcrnM8/zFixfQX3r//v2dnZ3z83PIoH/rt36raRpovB4MBgjfzJQE2ADk3KYuAmwulCqg6AqsdafXhTq+EOLq6mo0GGxaE6UUSiliGoQQjRHGlHMu6sb3fYU0Qqjb7Xa73X6/D3WBFy9eUErfeONLjkEF40WaDvr9Os86kbcz6o3Gg8vL8/V6vb+/D5oe6L4Kw/DlyxeO5/X6HV5lrVanLPNW1A5sN16tTcMsKuY4HqXmepUYBuGiurO/8843v6J0U1XKsT3LsrRASvG6Eqs4y7LKcwPf1lEYWAYxDGu5iqN2ezTul3HMS+F5ERwgjDGmFA+GXa1wr9cpigITcXJyIgTrdttZWpmmuVisPM8hhGyGwcPgemCdgNK3LOdGNeb70C4CWgjHNQghBDzfxnmcnJxAvxKcAQxLDybY7/chbwAoBiHPsqx+vw8RELge0KuCQnmjqP/Od74TRlGn00WY2JaDMaaYLJfLpmmuLy5PTk6ANp9Op0DIl2UJx3f7vj8ejxFCkMBCW+Abb7wxGo2+/e1vg0Zra2trOByCpnCjXG61Woyx/f19qLOD8v+9995brVaTyeSjjz6aTCaTyQRS42fPns3nc6CQRqMR8BUQ4jfMxmq1ur6+hmFSgDogyBi381dM01zFMQyPqOsSRCqQbIHTBYeqlDJNo2nqIPCTJD4/PwM51de+9rVvfetb8XJ1dHB4eX7hO+71xfmX335TShkGweePPzs6OgK5bb/fH4/Hr7322mKxALjW7fal0FqThmvX9rKsaBoWhm2lEGNiMplNrqecy/V6/Y1vfO0v/5W/vLe3c+dg7+DgIM/z6+vp+fk5TJTd3t7+6KOPCCGEIGDUOZPZ/5erPw+SPcvOw7C7/9bcK7Oyql7Vq7f1+rqnZ4DBDDEDYGZAEgBhmqAcNhkmFQqJEhkGaFGmZVFh0yHbDIcphyJs2Q6H6SUcpkiIGjAoLiEQBIYDcohl9hlMb+91v7X23DN/+939x6lKNFnR0VH9XnVm1v3de+453/m+71SlCPlyOS+KoswLjKi1Ns/zzWaV5xvGWKfb7nTbVVV5j+u6Xq/XURTFSQg9cLiiIdgtl0tgsiuldnd3YTWggw1VCjhUdjodxli3s0OAKQjCCXgq4/GYEAI1OPxvEFbhMTRNA+8HBT7QkSDKwsZyHmOMtdaMUKO01jbL6+Wm/t4ffNBJh4y1Ge/0d4aBELs7o4gJZzQRbry/G7fS4e5eq9OlnHniw0DIpjZNI4Kg3el5RDAlymhlZJSEq81SGZkXGxEwwNKdN0kahZHQRk6ml0C8As72m2++CTjDarWC1CXLMuBoI4RAU6W1rjZ5scms0pvlCgpMbUwQhlIpxgkmHmEnVS1VjYkf7PTiJLROM4rbrSQKxc6gh7CjDGOGPfEiCtdF3h+NwjBqtdr9bj8Ooslslte1chYxh7CZL64w9sfHR1qrbqf//NnLydWs3+s8/fhJXixPL5/VuvhnX/vvju8f/PNv/FZvNHjx/Nl6tSQYaSWTOGKUEIxkU8dRtNlkQRCHSS9K2trRIEy7vUGxWceB8I6KIE7TrnEybIs//gtfGe/tTOeL09NZnRW3j24N+u00iQIeUkybqrp7fFDXi4i5ME0tDy8Xa4z93rAtmzyMg7Tb9d4bZxEmUZym7Y5U1lhUlpqL2HlinGuUrGWVtjphHIiQO2QxRVAwAZWkKIrVagXeop1OB2NvjBqPR+PxCL5fLucIuVo25OXLl0DwUUptqfbQfQHEH7C9rQbIew88NKjpKKVQl0FXKY7j5XIJ7uvWu6KoirzyHr88OT84PEaYrLPCY/TkyRNCyMXZ+Wq1Wq1WVVU8evTRfLF8cfJShIGyTioTpwm7cSADdBN6OZTSPM8/+OCDb37zmy9fvoQDBn8L6GO32wVicp7nX/rSl5IkOT8/h7P0hS98ATIZqKu2NAWgqyKEYLZfGIbbzAGwCNBUgTIJCsGLiwsIzMCgg6Xf1lWQYIG/WrvTgZQDQgimBHjc0+kUMvgPP/zwM5/5TBQHx3eO6qa8OL+Cxl6v11utFnfu3P7a137rzp3blOIgCI6Pj/f29nq93nw+Pz09feeddx48eAAwuPcexgZFcSiE2Gw2RV0tl0sRsLIshWDayD/7Z//soD+klEdh0jRKWVdWDWVC8HC9ydN2p6zlx0+f7eyMjDEY00F/uFyvVVMlUWyVZYRCHukdttZKqRgVQRCURY0xRohgjLud/je+8Q1CWFFkl1fnSjXgpwRYClSWGGNIlsA37+rqSmsNRg/gvgFpOkKIQBq7Rf6uEUGEgGYP+Sg8MJDPwgMA5iiUU4B9KqU45/P5PG1d+39LKaM0YWEURMk/+Ef/mDKBEImihBLeKIMwPj0/swhxHkhtoiRxzt06uj1bLJW2kAGHcQQwELRugV4A/QaoOSDrgIj4SeMdsIHOsuxv/+2/vbu7+/Tp08PDwxcvXrx8+XIr6YLzBoB8kiTgEB1F0Z07d+ByIDezoIBtCCkQwK7Q4IY8BGrEraEVdNdgFmqj1WazOTg42D84iOK41e3cOjjq9XpMiNFw/Pbbb8Nbv/baa9/97nfTNFmtVqPRaDweh2F4cnLy6NEjKNSOj4/TNMUYI+Tef//d09OXe3u7rVYyGu08e/bkvfd+tF4vVV0T4teb5XQ6XW82l9OJNHbvYL/V7bTb6WDQ1VqGYdjv9zebfLnaRFHa1Eobt1it11nuENbOVY3K8tJ65JAfj8dlUa+XG2/17aMD7Py9e/cRIp1e94b0XSNEsiynlEMOs78/NsY0jXr86Om3v/Vd5xxjFEpYSFgHgwHc+N1uF862EAKSQyAQA3YJxlJ5ns/ncwJbE1BA6F/bG0snqNNhL8N+B6Lxzs4OeLtBXjWfz6Ge5ZyPRiMoreAJUREMBgNEcFlXPAgd8mEYOozAXEqEsfM4bbe89x6jq9l0NpshhBAl4/1bkN5BCrjNqQFkLcsS/KiWy+WjR49OT08BbSiK4uLiAkIaCNaOjo6ePHkCasx2u/3o0SO42aGwg6MMsBRk6Mvl8id+4ie2wJmUcktqhEoRbgz4ZYGDApEYyh1AMJbL5cuXL/M8dw7xMHrvvQ9evjzlPMizsqirVtqpK3lxcXF6ch6FSVFU0+nce//+++/HcYSQf/T4g8lkMhgM0jRtGjUcDlut9suXLy8vLzHGe3t78AvC5wFETAjR7qRRwHvtlggYIYQL0e12s01RVdXl5XlZ5XByvv71r9+6deuVV17RWg8GQ21cVVUeY0w5Juzw9lEURR5j6zEhTFZysViGnO30Bzv9wcHBwXi875y7/+Du66+/PhqOh8PhzmC8Wm76/X4YhmWVE8LiqI0Q/u3f/pfGKmMMZRg4flAbHB4eAg7qnAOMDyzDoyjabDbtdvvhw4fQYYGnwOI4Bl4c1D3Ad3r33Xdff/312WwGhTmEaKDvL5fLfr8P7wfPCXBEuENhOLNzrirKJGkJLjwJfufbvxsmnaqRvS4vq1rJpqplfzi8mi6Kojg5eZF0wjwvGROz2cIj3O32KWW1a4zzDnnnECBW7Xb7yZMn4JJ33esn2N0MvgFsqGmaW7du9Xq98e5+0zSXl5cg5ZtOp91uFxqMMAsVupow72HLBH3rrbeklA8/9fZ0MQceOOw/gPFA4A/9C2ijQ/IKrQroCDDGqqoCgPPWrVuU0vHBvjEGKFqU0ko2QRxJVbXb7c0mB/KOc25vb9dafXExtdb2er1vfvP3bh3utdvp1dUVJajT6WX5Ol8tj4+P9/fH7fa14+nFxRnGuNfrXE3O2+12XmZBINI0DQRDzsdp0lR1FNFG2m63O18vldU/+tGPut3u0a3jJx+/6PS6Iuhxzq+mcyGCx48fd7vdFy9eHBwcMkbSJJG1bCcJ8dIYk22qsqhn08n9V1/RWr722hugtEGIbPmWjIr1Oo+j9mw2qUqVtkJKRRjGvd7AGds0DWCf6/Xae19VFVSWgF3mef7s2TMIlEAuU0YT4OTSm0GG4Izz2muvbfUiQIkH8TJCaHd3t9/vQ1iFWuzg4IDdDPtumqYocoxRmqZlVeVlvdxkv/t73wzChIUR5bxqauscQmi92Xz89Nl6nUmjp/NZXlbL1cYinLZaaZqusw0PBGOMiQBIpSAN/cIXvnD37t29vb3hcLizs9PtdrfhHIRgMIYHWiwwUsx7D0p2mLEHcnvoqsO228oSnHNFlp+dnX33u9+FQcLwV4Awc85hOhTgEttbHjY67HuAuuBNrbXPnr+sG1U3Smk7nS+VMpwH7Xb3+PjuaLSrlN7dHUN/azK5nM6uvvPdb52dnzx8+PDFixc/+7M/m2XZK6+8Mh5fjwq5fft2p9PJ87ys8k22+t73vvf+++8fHx9fd7nqyiiJvaeYBCFvddqEkzCJq6ZuGgNOUpzzMAxfvHhGCJKynkwuQxE0VSllPdrpU0oPD/bns0koWLedhmGMEUGIdNppEolWkoQiwpgIHqxWi6qqVqu19+jFi5fO+clkNplMgD6CEQ2CaDze/+Y3v6m1Lssyy7InT56cnJxAkzLPcyC4QY9tvV7DSsIuAlIO8J4I9PGALwdGodtVBhOSLeUEpIygxry6ugJoMwzDVqsF7XholsZxDNt/uVxCtvDd73/v6mpaluVgMMSYOosQwtKYsqyjKCqqarXa5HmptVZGtzpp2mmLMEySxDtcNwrkAVCTQe8BqjdALkejEZj+9Ho9EOg55+7du3d0dPT973/f3BhKXhvlWfvqq69CD2mrbQVMdzKZ5HnunDs7O/vwww/nkynAH3Blw4aGhirQEaHcARKMMWaxWACaobXeTqYCSA4GVAAYJLXe5NmHjx/98Ed/YLS7OL8aDAYY44uLC5hrI6VcLBaXl5d1Xb/77rtpmn7jG984Pz8/OXkJD7Lf7//4Zz8DpdLnPv/ZV1978PTZx/fu33n2/EkShYzTQAgRMDhv7Xb3ajoZjnedRcZYjxFkZevN8tHjD07PXhzc2r04f8kZCTnLN+skEkY1+WZFMMLeLVbLLCu89xT7JIqN0ovFsmkah9xqtUKINE2zmK8CEU0mk6Io1utMCKGU9h5v1kVdy29961uDQS8MBWMsihLgbEAQBY95SulwOOz3++v1+vLyEi46QJeBTFeW5fUwWQgSWZaBTwYhZDKZQK0APAZ4KlvNGrpxUtZaP3/+HHDje/fugfNWWZacC4+ItOhrX/saUKciEVRFDUg4VB6ttFPJqlGSC1Q12lrfbvVFFCqtq7K5bsBin9c5QJLQIwDCfKfTkVIqa7Y0AJCsAI316uqqLGqoCu2NoVeSJB9++CE0f7ezdd947bXBYPD8+fOzs7OVkkKIppbg3wLoL0DuW77t1glja2bpnAPrtm0jqq7rs7Oz/f39vb298/NzqKLW6/WdO3fyzeaVV17JsqyqmiRJsiyD6Q6tJFksZt5YWdUffPBBv9vLsmxvf9TpdF68eHH76DCKIqnKKA4wxnfu3Pmt3/otiAVvvvlmnuf379+/ujwriuLW4eGzZ8/Ge7RpmqU23U7/4uJKSessCqN4vdm0emlZlovF7PbhEedstz/aFLnWOo4Cq431fqc/iINQy4bHwlrrtIvDIE3TOE6drdN2q5alrPV0OkWe+hQXRdVKO1Wde+/Pzs44DylljXSmkdoqSune3t5iXoxGI4I81JRRFAEjDHYLHGZjzPn5eb/fB1oJpExxHBNGwyhstVv9pjaCx84SggXyzHtcVc2WJwoEBWAlQ1YArUXn3Kuvvgq3D2hGry4vMca1bGptDMIY0X63E3NqqooTSqjPynV3p5fV5XQ9Obs8FUI4x7QXSXd3uVnPpgujtHdOKxUIoZSRsiYEKdXk+UbKOs83VVU4ZyjF3aTz4Ph+J24zzA73DxkVO4PR0yfPCWZxmniMpFbL9apq6qSV7t86SNut1XqBiW+nMXF23Otdvnx5/vTZIIz/zC/+SYb8fHpVrFdlkZXrTDdSXRu1IeN0VmzyMnPIjscjhNyw3xt0O7s7g51OD1lnpPLGQkTHhDjvV+v1Zr3sdduckSgUcRK+PHlukV3n66zMrpZT6e3l1VW+3uwPd0MRxGEUBeL27cN+v7vZrIaj/uX5y6Nb+3vj4WJ26U3VjUMpa6WassyPj4/v3rm/MxidnV6sV8VwZ+/28Rud7j5lCfJ8tcw4C1qtFhPJYLC3KZuikcbpxXpVVRVGNM/qVqvlnamNCsJ0tHuQtHsO+U6vI0JeK1k12lnkkWLcWo+yTb1YLqU1yrjB7sFw74AGgSUuq/N2L82rXFuXtjuHt+9qixtlorRLRavTP/jVv/f3ETYMo+VsWhRVEERRlHiPp9P5YrHCmDImnLuWbe3tHRa5zLIKGC39QRchRKAZCqRdqFgBeoT0Hy4FKP79zUAFyN4AwIcCDYpoUJ3v7u4CbQ9j/N577wVBEEXJ9ifB7IlSOp/P4WoAEmocp4QQ7zCI7wB/Be4mVDCQCwKeBXELsuy6rg8ODnZ2do6Ojrz3EMxOTk7gHhkOh1t1x7Nnz3Z2dmA2EsBYSilvHbCYCaXg1NA0TVnUgN5ba60x1uosyyDXBE0LdC4AWWu1Wp1OB5J4wGuhAkMIgSAOIQQWBrCGEIOhJgM/x/l8nkRxEsVwvUCKBknL06dPj46Out0upbjX6332x37cKO2t01IFQlBCvDNVmU+vJteJnYg45+PxGDoxQRBwHmBEocmMEMKIGI+aprm6mi4WC8iACWFlUUOeHYYh8jiOUymlMQ7S9KZpnEOcc49RWddbbbf3XmqFCNbWKKPLuuKcG2cRwoQzJfW7776bpm1MrylRgMxAagRXPJSVMJQW4itCCEZ8Q5FKtoJDuNSu2e834A5QgSAJC28GwQNgDjsSwIvgxsARBKZArGJMfPWrX4UkMo4S51CR5ZSgOAxCERRlCc5KzuOmUbD5IC+GXbJdBYyQVgp5TzBmlDprq7J01hKMW3HSTlLALOFsrNdrsAwA1hbAnHCbhGH47NkzmLatlLoW9AUCUPff+Z3fYYy1W91WpwevZoxB1hljrDGc8yRJoKc6n88pwpSQJI6VlLB6sBtWqxUAdsCCBU7ucDiMoujk5KSqqrOzMxjR1O52ACMbj8evvfbaydkp9GYJwpRihN3z58+jKOp223/wB38QhmHTqKIovv+d72errJN2jo9uNVUmGLp9ePDwjVcEQ2F07QcD2BnnXIiQEDadzpNWu9vtzpdrKXWjNMGUEDafzw9uHcHhB3JgFEVXV1dF1cTX6gBoZwvZaBFEiBLCuENEiJAx0W53ESJhGLc6PWWc9dghkpe1R6jb7Vrk67pxmEhlHz9+fHZ2EkQCCLtJkoA1OADVMCgLRvhZq49uHwI9FGrZsiwZbAWgDrmbacftdluqWimltYR8FggpwGzaykLAvgGiCOQWgJ9zERhl4oh7D87kBGBF5xzx12RKra0IotVqxZiI004jdRiGHllKw6IogyACExVQQW21b5AHwzcYY2u1EMkr9+9dXp5n69XTjx9/5Stf+c53vhOGYa/TberGOUcx4ZSlcQLUhKP9g+l02uq2jNK93qDJS865BoqKs0HihWAOeeyRQx6wcSGEwwiAWKtNGsXwGYCXiLxpdzuA0sMBK8rSGLNarV595ZVnz54BdwQsQCCNLsvy4VtvzSdThNB0MSeE3Ll3t9dp/7Pf+PX79+9bZ1QjKcVVVR0eHiRxlG1W3XaLsXDYH/T7/SzLTs9eJsn1FLK9vb1WO16XMoyixWJBGZZSStm0W61Wmtw6OPqXv/u7cRpXVcV4QDCz1nPOL6+mr75SF8Wcs8haL8LAmOsRxb3egBJukaGUuRtWV1PLspJ5Wff7fcJEFESJ1M77upbD4e5sNtvZGTVNk9d5VmZcUMZ5u9Uuivrs4uru4V1C0GQ2XSwXULbGaUI56/S6AFeDB0eStFardafTSpIkiqJAROO9QwJxUQgBEdh73+v14A692QT2estKCexPkDsBZQE6igDfwC43xjRKXkyuAOWCFgK0ztqtllGaU5bn+WaTY8pFmCBCl4tMGt00jTUepM/AcJtOp5EIrLUME6eNVZphQhEWlGWrdZnlztqmrp8+ffr6K69OJpNXH7zywx/+EKjvW/AySZIwjsI44oGYLeaz2WybsXzmU+/0RjtSq97OoNvvvfL6G6PdvXa3t7Oz0+210zRtpfFouGO0KstyOynVGMMZg7uFYbKlG8MX1Itg+ANTn9frNRxm4JUul8uqqr79rW8RQhxGeVloa05PT4UQr7zyynK5xN41TSWlfO21V1arldY6DGLnEEa0qJqT87OT8zMRho3Sw91xd9A/OT95eXYaxyHnOAgCo918PtvZ2QlE2Ol0Vln2/OWJ1q4oG+18JZU2rlE6K8paSpBnQUeGUrpYryHoIkK8w9Loqqow46ssHwwGjTJhGMZJyyPy8uQsL6qirJerzXS28IiUVYMJYyLw3pdVZYxbrjLkxcXFlfX29PzUGNXv9+EG/9a3vhVF0XK5tNaenp5SyvO8LMt8MOhQitfr5Wg0DiB+QWyDUwLXKyRSzluMsfcIei2QTkHZDgkrwI3byhoyMIxxK20Zj/r9nfl8CSyvTrtb15JSmoRBwAYOO+O04EFV1Y3UCBHvXRpFziIwfdjZGQoRAlCV53lACSbk2pjEe8bYarUC2DXDa6WaRtXOG0x8Esce9e7dv1/XdRimxpj33ntvsVjQOM5X6zAMP//jn/3ut783Ho+vrq4+85nPjA9u/e43vxXEycHRbc75K6+++p3vf2+z2cwml84ZGE0LlDlvrKobTih2ngsOWxNZZ631HkM3GCHEAlEUBWCojLHLy8v9/f2LiwuYbQf1/tXV1cXFBTBzp9Mp9o4xNux1v/71r9+/d4cTop0eDod37x0XmzWwxbutdq/bXq1WZV31+13GWNIKvNbK2PUmC8Kk1++/ePnMWZRlxXCnv7d3sF6vO63OycnJ5GoWiGidFcY4wJJrpUPOsHePH3384P6dRlZKmjhNPnr85PXXHxIWhkEciGiRzRBCYZIwztM4rZoaIRRGyXK9xh4lSQqIJGO8aWSSJFobax3GGBEsaOgxkbKmFH3nO9/5xZ//Be89omQym+3s7FDOH779trY2bbd7vd7B4eHZi5evvfZalq3LKocGmHOIEII9YvP5HHjHxhjw3VRKtVqtRlbQNYG7FdxsqqoCB3zQaoKtNVzuAOw755bLdZikYRB/9NELzgLAtJRSnAUAuxpjLLKLxQJRFkYJQgRyVkrJaDSCBBQABKhCkLPIeeQ8QohR6oxNotgYk603DBEhhJaqzAtBWVNWQoiTly+Hw+GzZ8+gOhmNRoPB4Itf/OKjR48+/vjjO3fuOOf6/X5V19/81rf+0l/6S0rr3/zN34zj+MOPPn72/KVSKttkAadaVdaagFPvPWUBpJUEY865N9fDKikmrU6nrCsIog4jaLBBEkIIWa/X4/G4qqonT55AMbdYLBhjupE+CKWUnU474HyT51opTuje3t4qX2itLy7ODsZ7R0dHz54+5ZwXRaW13r91MJ1O2+32s+envV4vy+vbt2+HYdjpDikH+74WZ0Rrs1qtQxHXtcyywhhX17LWWgSEhsJ7jBAJgnCV5Zj4NI3DnXi1Lh48eNDr9ayne3v7y8VaSt1qJd1Oy3tUy0Y5TyjfbDb7t25dXkzKqgmjBEgRjdStdhdoSt7oPM+jJBYidJYYZy6vJk+eP+n1esDFYYzdvn0bik7Ybxjj8Xj/7OwsTgTIvFrt5Oz0ajAYEpDRAJ5njAFXQWNMXddKN1A8wkIDCQPd2MtARQ8aD6DkQEpAKYVD3+r0irqCQCKl7HTSuqoZ8gghKuhsOk3TtNHGOWScIYxDCQ8VGCGUUg6gbhJGwY3dPeQbUHhCXQKYgPUOY9zpdUEY2dsZPHr06M7xgyzLmqq2bcMI/erf+2+UUu12++XLl9APe+XBA2vtV3/t1xhjo73xcrm8nM3X67XRWkpZ5g2jPmAcrJ1CKiBhALoNw+AQzSDGZ0UOVZRurLVW1vXWsAQggmtdq9ZAOyKEFKsNJEXgv06xPxqPN5vNnTt3smp9cvKi3W4bqTabnHPunI/CoNNqTxazOE33jw7t6Xna7p6dnf/O730LEq3Pf+6tMAw363JnMCqKLAzDdrs9uVq3Wp2006aMaW88RlobFCHtvNKaSgmXIb52q+NPnjzRykNTp9PpUIycRXm+jqKEchcnrTCMLs6vgCMBJDJgwDHGQNBmnbdWl2VpjAvDyNlqPN7/wQ++94u/+IsX55e3Do6gaNlsNr1ebzQavXjxAmPstbt1a//y8rzTbVlrz88ud3ZGUAKxIKQIG4R8EFLvrfNqdzzcbDb7vdF0OoXkCYYFwh4F8aHRaLMuISuAln2appQwrWWaiIAGhNLuYEc5X1Z1IEKn6kiQLHc8CLVpHrzx2m//9tc5F1IZSnncbqVpv2kaxkPONaVUVrU3lhOKrdEeWaMxxghj5B2jhBLcbqVVVRGGAb9QSk2uzoMgeOvhF4qiuH/7eD6ZBUHQb3XaURLzYH+4u842Sqqf/ImfhObNcro03nVaXcbY7Gq2Xq9X64luaoSQUzUjxDsktXWeIxzopqGCM0qtNUrJJG0p55VuBGV5VfZ3Bp+UHHLGjDHXunHGgEoCaicQ61BKrcXr5ao36LK0TQgriqIoZVUVlaxbSTTs71JM3nz9IaHo6vw8CqkxzcVy2h/sJkn6/OnL27cO5/P5/Tt3YerNcDjMVhJ5vFmWAc/m82m7kz599uL84iptJ51ee7leU0GkrAmmnLQZ8dga6t1yU/3wvfcH3UEogudPP9wdH51Mz62uojihlgohOAt4f1RVDSesrhutda/Xy6sSE7zKNpThWlfW2vPJGeecBqgqq0C0KUOcY9kUsQgERvlqSZHdGw3XyznndDGbNFWVHux98MF7r7322vvvv99t94wxh4e3T05ODg8Pw+P09PQ0juO6lgyIS3AIoL0O52Oz2XDOIdVbLpfQvIYIB0JNSGiiKLp//z40ovI8RwgRRrkIPGJp2m4aGQeJ1lYphbzziBZ5djWdfGrwVhAEjPGQBpDDQba+Xq9BPgGv7JyTZcHIdQkC5T98DNBAZ3mOPMGIjoZj4Ci9+6P3Hz58mOf57VuH5+fnzlrBOYxwgE59nq0Xi8XtO8dFmWmti7pCCJVVtVqttLrmmwJSAeUdiLGwN6Argj9pmqY72IG2+9HREdSUl5eXUAkBaxb6+LCexpjj4+MsyyaTCUz8ntm5c66pa8Y5IQQ5XFXVeNQdjUaj/uD05clOv3v68nkYhmHAq6paruYIoebcOI/29vaeP38OCpN2u3337l3IkTAmu7u77U5SljljTEljjDk9fZZXhdJNlIR17SgVWmvvSRAGzrnlajIY9LSWSRghhMJQ3L9/12GEEGp3u1JK4x3BRCkVJwkiJIhCAAehGsHEr9fXow2VUkDCj6K4bHKHvFQ1tibkAmD1siyHw+F8PoW2Z1VVu7u7m83m+Ph4Pl1Ya2ezGaRDm83mwYMH0+k0jhHp9XogwtwaFYF/xlaGK6UcDAYwtxOkm9CegqofPii4OAHbL9vUs6uFbAzDAlkia3X68qwuG90oo/IP3v/B7k7r5//YV7C3cRBqaZpachYsFovpdApmnNC1YjdekFCPw+aAvj9ArQihXq93cHCQpum9e/e+9KUvHRwcSCm1Nq1WG2h1g8EAEDH4/LPZ7OnTjzH23/32t87OThaL2eXl+enpy8nVWSMLkF8i6xgmnFBOKPEIvoEDDBhc0zSYUeBAQCnKGAMSN/T0Ic+hN9N8oKcAQCzA6QcHB51OhzFhrNXalmUZp4kxViq1Xq+vri7u3j7mnDtr4iiMoogSlMYJsEIJIaenpyDfA735D3/4wzzPZ7MZ0Amqqnrx4gU4UxRFcXh4qLUEoU8cR5RS54w1xntPCLJWb7J1lq/Lqto/uJ1tao+E4PFqnc8Wc8IoIng6n2lnRRiIKBRRWGkJQDr4ZPV6PfhN4ZBoa4qqcNgpa+I08gQVdSG1yYtyKwCGGsAYAyBxnufXTRPv4XNyzl+8eDGfz41RDFreW7Y8UEDgd+v3+xC04IgATxQ+TRiGwPoGdBpqAgi3SRRbRIs8I9jtDNr5Jo9icTk9bSWpM/Vf/0//49cfvvny9AR5WzeV94wQhpzjPARXCK0U4IWQboZh6IwlGEE7HjiBkEpvlfsIoZOTk8ePH+/v7x8dHT179vTHfuzH66yyVhPCdnaG+7cOpJTaiu9//yNOcJGv4jRZLafW2jhNWMSCkBJCsPPwy8LxY4wxzAhFjawIwdhh730QBJ1Op6obwrn3fjQaCSFWq1W/3wfvc+DpwJIChxoAKYCKhRAgTyCUxmkyn09FFAKeGoYBY2y5Xsky2x+PPRLz2bTdbq9WK2uafr//5PFzHsbd/k5VVWcXV8PhMEkSbS0PIo9pnCSM8clsKlVlvdtk2dVkFoaty6srpZQQzBiNMQ4EoZQib503QohBd4Csv3V8LKVeF3mvtyOltN5UddFN21VTe4/jNFHGlU1tnd9C0YDte2S9d/Dom6bRWnqC86rE1FGGtWkizvJq+d//3B+T2vfbLch5xuP96XRqjBmPxycnJ7u7uxdnl5D4QoEOLbxut9vrtwjwmLrdLrhiQyOOcw78edjOMC/LGNPr9bb9HkoplPAYY0D+YI5Rp5sW+YJivTNI/tyf+bewbxaz82yz6PaSX/7lv8AZujx/ubOzoxtjjUMEBwHXWiKEoLLbTiSxN1YTgB7D0pRlCRwCaDxa6+bzBedCSgVN0KIonfM/+tGPhBBlU6dpevv27d3d3aIoHj16FEXRnbvHh0e34ihMkzgMgyAQ7SRJwpAQjL0nCKVxTDEmCBGEAs4JQth7pZutvgA4r3BgptMp1F5g/A5JOTgPApdqa64GmoeqqrY5VZ7nPOQeoThJmqbJsg20wSazaZym/eFOq9d79ORJ3ciy1pusIlRkWQHNFFA/TiYTIQR4os/ni9lsCs2koiim0+mDBw86nU6rnWh97eGBsKOUOm+CgBNC9vbGR7fuakm/9lv/8m/87//mdLqom2aZL+48OMLMiTAABnAYx4PBIE4TsEfY9nhgUCcQ5sGdAGhcmF4nOdbaTTblgr3++ptRlMRxWhRFEERwA0C1BAkSJGDgUwvOIv1+H0pw5q7HylZAioYVNDfDuOD2B2NOwJjA3hZjYowBW3jAm6CP4r2fzubj8RgRvNrkt2/v/vX/9V9zzgnOGSNO6e5eH1NGCLPeSaODONHI7Yx3AZ0BvQvkeYCKU0rruiYY0WsDRyGlXK1W0BCfTBedTocQ1ut1Wi3jHAqCaLPJMaaz+SQUbL1efvzx4/as8957747Gu3Ud3EBaFDB8kKYghJw2oQhgCwJvH9Ax0CVyzxr1h3NyldZJdD16IUkS8CHbAsmQfsCfQCYNvwUwkWGQRRRFxjljlUckyzLrfRAEjZLz5SwQ4dV8cXF2MhgMwihZbUrOKSY8z/MgSghhtSyt85PpTAjR3xkWVXn//v12KyEEf//7P9DGVLIZjXaUNUm7NVusjDEOWY8sIcwjC0y5JIpu3TrsdwfTafYb//Rrf+Snv/jy/OLOqw/qRZN2oqrOvR1yypyx2mhvPHTCJldXrVarriRc0NBchDXEGAcBx956ZyhjhKCIRyi07TTqdfsORIhaQxsWCrvFYtHv96MomuoZ5I3QJILJCowx6ygBR0LYiGVZQtkOgJG/8dGklEJIgLwH2HcwriEIAqD0QauprmuYEbherkwjVd0gb/nWWVxEm7w4Pb+spJLWiThxBPUHAx4wIKloreH6hm20TeC2NQc0saBmUkqNRiPIxubzOfSg4YCtVitAGxBCy+Xygw8+CEUAmr7j23d2d0ZJGLfSdsCDdrsjmEDOE0KctZQQjBBnDCMEVRckG/AuAM1uG2ZQF85mM2CFgtoJqP5QzMEnB70UMHKgnbFYLIx3o9GIMAp5ERPceGe809ree/XVtNu9/9qbQdKOWt1aW2lxJd3VbK2MfnHy8tGjR2dnZ6CL32w2l5eTr33t6//qX33jO9/5DnzaN9988/j4GNxelVIi4De0BEQIwdjHcdxqpXt7e5TSO7fvfulLX/rg0cdvvPX2//m//L++98H7X/jCF7JsfXF23lR1keWqkUkUN2WVhlHIhdNmPp/neQ76HzjD8Ozgoeim5hhRh7xTy8X0L/77/+5qsVgvN8AlgpIRqCG9Xu/y8rIoisFgAC6qW9r7H7JBtkxKCCRgHwKlKMykhy4orDg8CXhOW2bndv4GiM6MMVLqplEYY8F5UzZFUWzW+Wq14Zx3O/2dneFytbHWARTAAjFbLvyNhSwsMUAEEPwhjYNNCdsC6hV7478PlFZQVkGSACkRJNZwbLZOfWkcg700YwyCtNbaGAu0AchxAfL0zjlr4ajASwFmvAXhwfQLf+ILoE3YytvPD81kuNbhU8EUAaCkaGusd/BoAaV49NHHF1cT59HJ6ZlHOE5aRVkbhygPLq6u4K2FEDBVotfrgXYUErPJZAJlLpwiyMTsjWMm4LjkxiuqaWrnjLX67bffvrq6arU6v/Vb/+If/cN/Mpsv15sCiGNbG80yv74tkfNQLIJ6eMuOACQHOQ9ICCXYe39rf9xKolBE69W1qhFIHdsUDsQz4A6ulLq8vIQKDNKnPCvJlgIMxO+mafI8B/v30Wi0dRcHEooQIs9zSE+BbB/H8e7uLiwB/IBW1llECU+TNvLk8PDw1t6tKIq8w6tN1mjVHwweP36ctlvWIcLYbDaDbAOuXXDBhUPCbgbQw0IzxsbjMbghgDsAcFVPT0/39/dhG21HKFVN4zFWSjl07aMbBMF6vZ5czQQPV8sN8YQQVhW1tZ4xQRD13hKCMPaEIO+tECyKAsEpRg5IXs45KODg9gf4HRYd6N5JkoCNBbhEbQ0dQI4CyhCIDYvFYrFawt6SUjpnBGXe+6aWRVXOFvMPP/o4iJNVlmdFnrTajz76OE5bg8FgOBwOBgOE0MOHDznnQJIihEyn08ViAQpJyDoQQrPZTEpp9LV+FWDEOI4JxYeHB5h454w2zVtvv54k8XQywxgZg/6P//n/aXd0CJ8TcpX1arVYLC7Ozp2xW1u/sizBewtiHCgXKKWMkjgKo0DIpv6rf/U/4pxFQfipt96BOa7A0Gi32/v7+0opgFmg2AUWPYCbzjlokTBKSZ5nnIs0TYPgesTM7uiAUiqVW60LzrkIkqKUGOP1Jmu1WnnRLNcrIUQYR9qa+XIhwiCKY2CmhHFU17XHKC+L4XAICj2rZcBpmgSeoO//4HtJKy3zojMYBY5wEtRZRQgBzoq1tpESWweRO03aVl2Pd/YCbdYZwVRw4R3iIpgtpt1Bn4bscn7V6/U6O93laiWdEklgla5lnfKUIGyMqZoaOd9vd4M4WufZYDSsZOO9pQxh7ClFyFtMA2u9c955l7Y6UitrrSe0cc57RIMgCAKLUFZUURR10g60YYxebdZ5GIatlC6XyyRJ60paawlmWl+LQr33AJzBJdA0DaUYa1MVZVMWgjFKUJK21us1IczVTumy1WqtLqb9ft84JLUO0nCxmne66fOTcymlt+jFybnWutOKwUL2zvHtoipVUw563d3+CCG0mm3KorbWCyGUNdgTZkkUBBSTgLMkSUQQbNYVIWS9XgWUfPfb3+nt7B0/eOikWy7X2jfT9dpaW2sJdSpBGGNU19Vqvgig98sooxQuVW2VQxYjlERhXUwo8//2n/vF4U5flhqzoKibdtrhlCWxMEadn70Yj/fLvGgaBekH1I7D4ZAQkqYp+DcpYxlEHUqBFHw9v2u5WMdxHMURRAiIlOCoVtf13t4ezMOs6/r+/fvQzHDOAYBqjIFhGuDflKYpBJWmaZqm+tEHP7h9fJ9FznuPrKtriZxvjBUhn8/naZoyIKRqA649QZx6jEQYgFctEJ2klEEUEkY9RnAKoWKDiwxUK6vVChlirZXGhCLAGL/yyivOWO/9YrEAaL2u6063jTFeLBacU8boNfmVMXdjJWm844HAGJubmbCQgEKvFRJ38BYwxoAgGIakA/4AKwP8setutffWWu+vZwZA4gQgxta3rGkaKRUozpCzaZrGONaNnE3nnDJKKXDb4jje5KXVylp8cnKSpnE7jrrd7tOnTzudznK1Ae45xphhQq6pjw57D30vuPSMso2SP/8nfm5Z2k995nO1VOt8HabJapk7Z7BHlFKCkLXWWEdphTFG2BmjrNXe4zASyBLnLEGYYFxV1aAfMEHefvvh22+9s1wUEY8XixlynlJelnnaivf2hovFAiHX6XQoraSUR0dHi8UCdhRwlsGsGGPLwGXUewQYHmMMqjNCyHYAaBAEMPRzNBo9fvx4Op3CnQskechIQGcHxojgMwPp47YuBhbIvXv3VuuNwbzf7zdSKZlb66NPOHYYreMk8QhB2FeNMlqHYVg755zT1lhlO50OIYQwWtQVUhIeKuSgYB6BEOIgRCaYBUIEQRiGUM00lQSuHeRkYGLBOQfvYCjFnPeMMY88IYRhgkCJJQJI7MBVFDYrmEfAWkH3C+aNwPr4m2mRAOxBkIAij3MGGxo+A6T1kNNDumWMhrIvCMOmadqd9GK54pxKrXRls6zglLU7qTGaUBdyMdwZAGycpqnWNooiv9xAH8tab63VUrlYBJxyzpI4bLdacRRZzSSpmaBvvfVmaaLe7vS/+ru/enx0lBdLxlgQcIA8YU9LqeBEYeIQxkkaam2w88R7QhAiSGBMQjpfXP3lX/mLb7/9sNhUYZAs5xt6PYiaxXHMOHny5AnGOAiiTrsH2QJ0GVar1d7eHmBBEERmixWRN2Pf4bjneZ7ErdFoBFk/Qgg4Ad77oiiurq5gH4AurNvtnp6eQhYLTxoIHFCyLJfLKIomkwk4yzVN89FHH718+TIMw+fPnxdFgeg15QKScYiOsKeB1gppDWBvW3Ep1BZSK3UzMw7G5UC1BK2mnZ0dpRSQbeFQVlXFOeecL5dLqKzBhQrANWjebuseOFRwriIRBFyA7AHKMkjvYLnABmPbVplMJnBUADTFGENIBkLuFt+F4gyqUhDNwfdQ4cKBh3wRdKTT6RRwTSp4VlZwJ2pr1pu8aRqtrMPXBWsUReBDfXFxAWsOsYZg7L3lHHvv4iiAK45R0Wp18nzTyOrg6PDn/9gf/6mf+unPfvaPrDYbaSSIT+ry2p4SeELwXKyVCJsg5Bgj7x1GjiDPKKIMWdf8R//hX7lz515VNd3OYDQacx7cunVkjGuaBqCVNG23292dnZ2iKCaTSyAQjsfjV199FepjqP/A4YeAQxi0MYHzAckf0OPh2W8dR0DLvL+/3+l0QK89GAzAehP+ChqAsC0IIZPJBLYFZNNvvvnmnTt3er3BW299SimFrIMbcytFhc8QCMFuLGSp4FESG2cd8h4j650Ig9VmDe6vAD3CvDkpJdiDaa1nsxkgRB4hxhgLRBBHRVUaZ1977TVQLUNKDg4lQKXbAhEQ+aBCh8gH64MxBmMVMFqCXQVnXd2MjYS9CBU9YGcANtEbR+ateAE6t3CeoUUCf7VYLABt2QI3lOF7x3fee++9IIiSJJHaSm3a3R4PhAjCvYN97DyArIPhThhFWxkZiMMoJoSQgFMtmzCgUpaMoTiOMULZcmWM4yywSv/v/sZ/9sEP/+BwdzjY6abtJI7DIssBc8iyTN84BRmnMaNCMKUaQh2jXnDMiW9Foi5Wv/RLP3///l2nfFNoo+VyMev321rqVtJqt9txEhrtVstNntWnJ+dS1YOdHsKuqgqQ1YPDVKfTOTs7M8bEcZuBjUySpFVVYUzB491op7VuZANBAtyzoAiF2hl2Oixur9cDGjz8GDSpIPsEp2oovD7++OO6KnbGe5vN5hu/+y14YBbIfsaIkMNtuAURAS0Pk3Ryed7tdgHoXq/XUMWfnZ0JITC5JpqgGyvkLZ4StVJCSFVVDBNz0+FVSjVBA5RWY0xZFphcp4mc061lgBBii4Y454w19aZpd3sQ25xDjDFGBQT4W7du1XUNXgFgMrBlf0MJj24GgsGvtoW38M0QaLjm2u02KLoGgwFcKVBgCc7qun5+8nIwGsKzaGrFRDibLihBGHsWsGGvD2H15MVLOHsY49U6v0npPHYeYwK5b7c7TOJYNTXGXDWKU6GUsajqtuJv/PNff/jWp2yTLa6ubh/sP3j1lScffTwcDj0YW3C32WyMcZhdrzP2LoxC5JRWcr2e/+k//ad++gs/OZ8tD/aO6rp2DnHGy6IwOqOUGo/LyjEqKOUAJsBYlcViEUcpY9eajiAIxuNxWZZlWc4WcwKRBh4/qBPhVgJUIgxD2AGwI1ut1nA4BKAR1hegWnDSgf8FrkutNWSKUIDneR7H8cHBAfTNHz58CIcbIeS9hWsOoEFoEyiloOUVx7GIwqKuaiW1s4QzREl30PcEK2ugDQYHA0LONjjBJrjGAhmDGKa1BmRYSgmOyXAzQKoNpxFAX0BSITsEMzDIxqBZ32q1tNYQwsHFBVTqYH8HmAsUQ/pmGg5gq9DthNMCnFHY1hD+4ziGOglATfDkgYUFNZzS2ljPRADpNSKYcgbib6hf79271+v1VqvV5eWlu3Hf5ZRSSgEvIxgN+r0oimRdI+sw9lHYSpLWzmAgvKNIPnv67vTs+dH+7vHx8f379z//k38ErkRnTFVVqoFoRaXUWmtMkNKVNs3to91/+8/9j778Mz+52eTbWVlWES0Nxrg/SHvdSGvZ6XQ455RyWM+btCoAZH7bzbbWwoy4o6NbBOp5SKomkwmQO2FXAXQHMBss02azgRkaUMbCXFeIl6Cg/WQ6BfNAIJc6Pz8Hgs+LFy+Gw+G2GQObA240oGw1TXNwcABG0QDIf7J1BFna1gkf0Dg4QhB9IWgFQQDsYOIR/MJFWZZlWdQVD4QIg1o2RVW2Ou26abQxSZp6fK1WhZeFi4Jzvr2yIbjCxD2oji8uLsBfF9J0pRSQcaDWhMMMxxg+OezR7Z5TN05jZVl++tOf3iqb4d1BHwIrA3XDer2WUoMtMGT5PAgYYyBons1m3toXL16cnJzYm9HOUK4BL4xzTim+bokRqqUi2EN5l0Qp5/zW4bjfid989d7R0fjo8ABjTyn98pe/3O6kzhk4ipBKeYcxxsgB/q9ee+3+f/Kf/C8+//mfYJQKHkWx4AL3+11KGSEsTdNOK0bI3jgC+W63C4pdQghCDoYsbpMc2Abtdrvf7xOCiHN+Pl/A5CvGCKVYyjqKebsTr5ZTJcteN5VN0e0kjHrkdRzxnUGnlYZRyNIkWC4mSSyikOXZ0lnJ2fW04C3HGUjXDiyrPb19+/Z4vLNaTjutlCBsvEbcSVNuux1pGp+fn3hkPDI7w97V5FTcjBcnlIZRJJUqytIgXylZN6qR2mhnjffaE0eIxdggVUonra4Up4IzZrSiBGutELK1aWrTWOL2bx84ilkU8jhqD/oGEam0sc46jwhBhChrKyk9pdpYiviwu9MK007cbvIyW6xiwRlBjKCqKSknPGBxGnV6bRHy3qCbthMmqDaGca6NKcpSGyOCwDrnvNc3kgRo7mOMHz9+DNJbY4xg1BktGAO1nTIWU+YRKytljCOELdcbi3GUxFESM0ZsJYnXR/t7x0eHX/ziF6MoaXd6UZyeXZ47b0RAMKIYU6Md50HTyKaslGy8sbpqkHVJwFSVV1l2+/adf/pbX9tUxfHxYZIyjDUm9v0PfvjX/uNfGfUima/jOGahKFW+Xk4ZVtZlvR7/U3/qZ3/lV/6DrCgb5RES3qJW3KOEG2OYIB67RsrzycLTYDQaCyGcM+v1MgwF+KdCMSqEKIqsrkspa4TcZHIZx6H3drMqGFDxQAQHCwQY4WKxgAIQICSIKGAMCQZ32wbAZDK5JtolCfRbAdBijMGftNtt8HqdX06ooM9fvDg+OuacO6cZYw75uq6Rp3At1lURhkESRZBUee/rugkYJ5zFcbzJMihXdSOttRQzKLG11hAmgExkrN0UOQSPdqdd1xW5ntKJtuI7oE4CaAB0LaiCMcbKaMaYd1hr3Y4ipZTUypcFHLY4jlut1mKxuG59ReFyuRwMBtPp9Pbt2/P5HDKHuq5v377NbhxVgV8LGcI22MM3sOxg4Q7VFcAFkOvbG+tWGK4SBIFDmNwM5404Yx5BX9s5d3FxVdZVVV5fKRC/BbkWdXjvd3o9wXkURWEkiipP28OACURRlhfD4TiO49/9xu/+iV/8Y5PFnCdJVqp8k+0d7P+H/9O/jDz/B//tP57Mp/oy2z86/uyPf2o46IzHo/v37nIsNqqKRCvkYbKbApsHYgoEJoyDKExAD900zXC4izG+urro9TuQiUGqA+sQRREI4MB1ggGDHTjzW/4OFLBBEGzLVbiewKgRMiS4WYBkBeEadiGlDJh70JzdklHyPN/b23v8+MMi32CCer3efJEHoUAI1VURBK1thQtvXRTFcrkMOGOMSaOdkhjjaz4HxnEYFUXBMEbWUUahAMrzvGpqqJYE4wihzWYThBw+PKXUGF3Xa8h3YaPAJgAtQCAE5KnwdEUgAHLy3iNKgzjiLpivltzZoijSNEWUQP4AvXUohmB7wYucnZ09fPjwO9/5DtRboAGEX9Poa/h2i1gB9ZtSipwNgmCbIjPG8M1EAMisKBf+epAuZ4gM+oOdnRY4kM0W8+lkzgKhlLLOO++9Q8irgFNKqbd6sViMd4dhKCYz1mp1OoNuVpRRFPVEVJb5wzfe+vijJ71O7+XZqcChUsZj/uv/3W98/tNvE2T/B7/0i5hhGvJWGJZVTpEf7e68ePECOzYY7FZlXakaM73tCI73Ri9fnML+m06nURyAKxbcDBjj5XIZBJxSmmcVjFaC0dpxHM9mszRNF4sVgV4w1EOXl5dKqX6/f629tBbmKKCbCQFQwSwWi4uLi48//vjp06dwVcGWhdYIYJOwmlt9Y7vdPj09ffHsaRQF4/GIE8wIbiURss7I5nrO0A0hA44gxhhYqtqYOI4551Cj6EbKqh6Px+ClDzXZxcXF5eSqkDUVnHDmKfE31NrpdAoyZXNNIydAdIC0En5mu1khCYbYBn4+Silo5QPPX2vd7XYJZ7PloiiK/f19aKRBj2A8HltrgfoJhcv7778P/Bt4C4BIYVMCBQLQDPqJ0WHwSWBfwiEHkHgwGEAEAcB4S47e2hcul8tnz55pZ+u6qaraGEcpD8PQY2SRt9ZS7AnG8C6NrDd5JkJhrc3yvCzLMi8G/Z1ut//Bux9wQoUIlHbWoH/2W7/dG3RbrXR/d3i0v5+GnGOy0+k54+u8Ho/27t+/v1rMm7pqd1KAoqHam00XMNwHYQeynbOzM38zfPvs7AyKJOhTQit4MBiAv9Xh4eHR0dEbb7xBhBDPnz8HWh3GGHzGANQEypIQApYGIg3wPqGoh5CwFSfBFQno3TY3h0EwgA/s7e1dXV0hi7zHrSQFcjTUNPB4ABoEtBz+CiKf8x7iDUckFEESxUWWW224EPPVMlutldFwo0FQAsHaNumGhiRIqAF/AP0qfEj4gl4lIUSEAcRvuHcA4uCcA9wL9QqUU0VRfPzxxyClhbLU35iyAvcFfMGBTgGHHMIJwK7bwAlnZkubgpsKlhHWH7Yv/Eb0ZjoAwdejcqngk8mkkRphSpjQ2oBrm5QSU141CmI0wo4xhp0lHnmrg4Bbqwj2xpiqajD21lvs/WAwWG3yOEq0VtbIJIm11kEQeSSVroyuI06Kolovc4pZWdaCcdlU7W6r22/1Bh0gFkIXLYqi7aQ/6Ky2223QPkgpf/qnfxpgHABt9vb2EEL7+/sAhYIH5XQ6Jdbaz33uc1DLj8djiGFwQKHUB/QbCFEQrtI0hUEfoAHaospSyouLC38jGoFhQt1uN8sy0NF3uq17d49//LOfiaIgiqJ+t0ew7/V6FF0bgWzbV4DeoxsdEkDoAO8lSYI9sto0VT1bLubLRaOViEJ4nGDkj/71L35jAw0gGkRKuE+vqeY3s42990Zp5xywXeGXBWwImDjWOUKpNqasKmNt3TQwuVRrXdf1hx9+eM06Qwj2HFzK9hNGzPpmug20lODHPvlpt20kOJxwkOA2I4QAyxNjbN0f8gEopZ1+L2m3jDFBGBZFVVWNiLZz6pkQcFqZEEKqGpyK41baNJUxhlO2Xi+TNEriThx1JtO5dSpfz+Mgvji9zLIc6LZCiEbKzWaz0x+04iRJkkBEZVkrbTElVVM/e/YMIEVQv4xGowcPHty6dQuagpCCD4fDT3/600EQnJ6eKqVevnwJYoTVagXX7GKxAOACzBkZ7D9KKai94JhChNjb2yOEAOfvujTRGrbRFpGRN3Zz7GaEF/y8c+78/BwQ07qur0f9SZXnmZRys15+4Yt/5Fd/9e8zxpwzwJygNwxfpVSYpvBQ4e1k0zBMvLUe4aaqGSFlnq/X67WsEaPo5inCBice+X99g0JWvdlswlD0+0N3o/CCzjIkgpADQNkHD4OHATQbsyzrDHYGOzsArALp8+joCFYZciRYPYDoYQtCRrG9uGGJYBmdc0EQAfa5he4/eZwgO4INvU2TIIlCyA2Hw6ZpnLOM0CAQ/bR1sD+q66auG49wURTGO4SQ8MRaGyUxxhZTQjDhlMRx3O12ESFFVQZRUmRrU8swSY3WhKBOv7fOG2lMUVfU25OT0xcn5/1O9+/92le//FM/+dZrD6212ppsPa/rujPoG22dc8YrynmctgVjyF1nOEDuhHXGxMOeM8ZgRKfTaa/Xm0wuO51Ov98NgiAKW2VZgugvCILpdAp43GQyubbuhoKLUgrUB8jxYehqp9OBMw0sQ4AAoX1fVRX8LWQV1984AhkeGMlCaJzP551OZzqd3rp1aLUJw3A6zRACqhslhDCG/XVhhYUI4JmBxNkjRCnFQCgmlDPmrb2aTp1zhBLtrofIeO8pwtRBHP3XgijsEiiAQIQJbRvYBHDRNE1jGmmcxRhT5+q6puI6e4HgAyTw3d1dGBoB0Ack68D5gEQCypdPJpRw6uCNYI/ChWM/IVTcVkvoxiYNHq2/pj5dR1OEkBCsrmVd11EgGGOybhrGX7x40e6kCCGPkHOI0etqjzCqtU4Sjr0jGMVh2O20EEJJFDRNJcIwjZPlrJKVtNZa5Ps7vacnZzvD3RcvT995440f/vAJccYou1wuf/3Xf305WTx49ZU4EatsWhSVRqZuVBy1i7rK8iKO4163GzKEMb7OQWezNE03m02v34EVhvtQSg3zEpw3AB/N53O42WCPQffx2sqlamqpFWEUU1LLhnKmrekN+nGagFULmHwXRQGHAKAQggXMdaxKVeRNXelsUy0XmZS6LDLB8WY9d1aOdwfOSka90XUSiyCOnj59fnFxFXK2Nx4Md9tC8G6/F4ahVWXEBUXUG8wwh/Aj66bd7tbFCpmGIicoFgQxwTLd5M40HGHsBMUEWexsQAnHiHpEPWLumjyBMSaYIU8ZDQjmGAmodeaT6SDtUGU5IgJT4pGxNmqniBIqeKNVlCZlVSGM4zg+Pj5ORRiLoN/uTCcTYNoD1RVSK2hXfhI8gkscNhnc47Bj/vDAOOQRQZha77Q1hBGHnEMOEWectx4Z543zDuFaqqqRxvlaKquN1Y4RH3CEsDK2icKg22mFQoggoIHwhBbVdecPY2xNg2zDrA6IjwPRShJifciElubenVc4EY2xtSylLby3GPHTq5dxi/Ekenq22DT63t1+v2Uioo/3dq3B3/zuD/7uf/3ffPObv28IijsdaR0PIh4Gu7u7nXYiOPFIYUqKqpzOZxdXl4TRxWrpkN9kxXK1cc5UVaF01e6EO8O2dbLdTtvtthDiwYMHMFsVDiowig4ODkQYMDjTcJVAdOScg4lXqa+zHITQ7du3YfAU9Hii8FpOADGD3gxehmDOOYMkFYAnoPBlWYYJ6/V6/X6fUPry9LLVatnLqVJytLsDpnPQQkQINbVstVNCyGy+oIJzRGxjsMeIEilVkVeEUu8QQg5j7DHBBFuMLCZ/eFEivM1o8c2X955xboAMKptOr7der611LOAY0bppAF8jgitnAR0DxWaZV42SUIMD1ratBujNxA+YdAH7D2oawOm2YBY0kAHzIuQPPyq6qZYwxoQg7/61mIo/8cUJNcZ47MMw9M4Qf41MR1HEGDNaWWuh/+duhqVTShBC2DsoqtJW7L0nlFiroyg6P7/Uzsoyj9Peer3Mi43xJIij0Wj89Onzt1+/s5zNH9x7Zf/WblZWzhJr9UcffbTKVl/6mZ+1FldVUVUVCE4QQkHA8zyHvi7AC1EUZVkGUzeKojg+Pm6a5sWLFxDpFosFlDoECyCnwq8MVhSLxWI0GhHofgLbY3d3t9PpgAAUOsvQ4aWUvvvuu3Vdv3z5EtYCcjWAVbMsOz8/n0wmgJgA9UkIsR0xGEXR4eEhzCKCXQIZhhCcMQZIqhACIUcpdcaWRWad2fYYlTR1LZtGEYSTKHHSWqUFodQj4uGhEkyYRdgibxGGfzxGHiOY2Qrfwz/aWee9J3iTZ1VTA/GMOMQ85ow5a2XTeG103YRBAP9JMJZSVsX1KkESCSNEoFsBzBKopeBAbq9pIBb5TzBFIFXebsHt5oMD/8mN+2+cLoyx8e761ay9f+fu/fv34zj2GFVNA1NPoS8Fxb7zZptUMMaiKGAUU0rhV55O55vNZrFYOGdglktZVzfIhgiDeLFeRXHw6qsPxuMRF/Tw8BDyb2vt6cnFb//2byPk7t493hl2tZGQtRt93XSI4/j8/Bx9YrSctZYxcXZ2URTVG288TJJWnpe7u3uAheV5DjJa4JdxzvM832w2CDvWNA0U2kKIZ8+eQVDc39+vqioKQiiDtuB+p9Px3o9Go2xTwS3snDs+PgaaGcZ4sVhG0bXsDjI8xth6vd5CgHt7e9DUzrJsMNgh5FlZlta6IAjKQuWbtffeI409a5qKB5FSGjmMLDLGSqKF0Zsip4I75DB22BGMsPMYeYc9PFqPEEEIOXINKHqMscPb3E4bzTlHGCPv19lm2BtEUaSutxHKqxIYpbCNIA5BNgkCt6ZpRCBAmQmlD4BTgK9tRYWf3FgAdcGeg/X03nPGMCEeqL4IYYwQBkKWJ+S6WNxuVnwj5aOMEkIoRrf2x3EYWWuN9xgh6x0hmHMOKhQoYb33zlkaBoR4xmm73bbebdZ5ECXK2B6Pl6ssbXfBLLKWDpFQKcWjOOCcMWaMohRbpxljeZ5Zq5MkCUOhVUZwMJtMvv61f3b79uG9ew86rcioBjERRdF6OW+322VZjkaj4XAI1mIAU7RardVqBQzMbrcLom2YRUEwS9N0tVoB64hznqZpEASqtgSGGUO5ClO8d3d3EUJFUQCE2W63KaVpmsJcFULIJ42uAVSH4inLstu3b2NMAdkG6QWIg8F2BrBDYFD3d3qcUWSdcw4RIuKEC4aJp9i1kmS5mv7Rn/vjBwcHcRyvVismeLvbZUKsswwR74l1yDrkEaeeYIaJoIwyzCnmmHCMOL6+N+EIQeC5BmgZxRhjRi1Gxruzi/O6rqMgVEoRhOMgJB5h50MuNsvV7GrSlJWgLE1TTAlgvf7GGx/wI9j3gFxuOSvb7wFd2mZK/ma6zZbctMWhtlAJbM3tz0NYhRQWGjDWu02Rg54RiNIY4yBOoigimFVVA7mEVlaI0GgbRiJNY85pHMd5Xmpt67peLBbOXqN4m03earWAJmyNRwhRhu/cuV1WxXA4NFbBj7VaCfQFESJhEC8Wix/96EeTyaVSMk3jMGR1XQKTBmiZP/zhDxFCEJXjOF4tN86iVtpZLTdlUQ93dvOsXK8y5AkkTjD2xFoLg4yNMQ5ZAtwz2IUwvxoqVihdgS6klCqKAhhc3ntogUD/CYIHu3GJB1/+Fy9erNfr8/Pz2WwG9Nuqqt58880tgj0YDJpGRWEMKSmMizXGOK0x8di7v/7X/1d/82/+H37t134tjKNbh4fGGIJ8GobEuYSLbhi1BW9xxhmmDBHqCfUcI0oQJ5gTzAkVmApMGaGMUOEw0pY6hLQViDCEqUcEY+895Wy2mBtnQSp5bZvv3HK5hHEzIgp5GMCBBJaQuRErw+b7N762LTHIBLabchtfgdUFkDOEWEop9p5izCkVlMGOJDeiGngRoDJSzqXW2CMj1XKxyDYbpw0hhAUiTWO41iFTgvqMEBTFYRgGURQEoajrEpQtgofWeueQlLIsS/gwH3744RYuBNjxK1/5CiA8jBFCEOecMmyMCgRx1iZRksatb3/zOx89elwWm6vLsyJfoRs2GUIITNwBwwbuIgCxQoirq6sXL17Im4nUnHMw/r1379524hSlVNma7e/vg58MgPDX0mGtd3d3y7yAVwcmGDwh8MAJRAJXCfwbalVjjPPGWjsajaIoevfdd2E8OrAcgCMHqBZCaH9/n4nVYDCgQbhYrcGkWNeVty7Lqp/5qZ9CyFvv/h9/62/9/M/98dt7B3K1tnW9E0cMeeecDyOE0KIqDfLGO4yoJlA0Xe8NjwiCyx4hyokEQA4hYj12znuEMDLWMkoRwi9PT+8eHxNCnLGIeIJwGIYUkyAUUspitbmmV1sLF24cx+qGe4r+9S84w9s2OuxCSNHgpoP93e90z/yJVZoQxChxxOPryUweYYZurPyg5IJoLYTw2HPOVVMPe12njTfWWssCFqfXswPhGcGjpOy6/YGQgw+mtb7xKsftdtcjCk9NCDFfrqfT6d3+AI4ZxPJPfepTsqg/fP8DeBFMfBDwXq9X5jUhFFSKYSjee+89Qt1wOBzs7JVltVpdX8UQkqGLYYyJokRrTYjjnPf7O1BK5nle13Wn0xJCdDodoOCkaQq48tuf/hSbz+dbucxms+l2u9CZgFQX6s1er7fZbODN8jxPkoSzCGpY+GWg80kIqcrm6OiIc1oUxe3bt8GaAa4hALegkuj2eqWsjdXrdRamKedBEHLqSSYlYygQ0e7u0FrLaHh0dPzG22/NTk9DgqNI9GkYU4a19lZ7goexkNZKby3y0nvtnXTOWOs91p9EQj1hIVVGU4c4J957RAkmhHivnSOMGWeLqkpbaV3XWiqPEdQ6EGCUUohxZzSsNeEM3Rjgf5Kuus0XP1n9kJvp04DhQ8UthHj0wYfYI06ZRxYjxMi1IIQQ5Mg1cRugq1deeeXJkyfXdFLk4GWFEJVsjDWEX1ddjdRKNcgzcLaBnU0Zwfh6eFBZliGhgeBlUROmiYi4CC8mF61Wa7PZbFu7yt7wV5Cz1rZaaavVyop1GIbGNBjjNE1lrbW28LJxHJdldnV1Za0uy3y4c9TvdZpaEYT39naVUlJqkD4TTGBBwHUBUlJKKXQZtdYnJyeMMWhKQydytawYZ1ErDeu6ruvaWaKkE4JZa/MsB5sQQEDBJsVa6zzN8rLX5Y0s66bgnAchJ4S3Wq08zwdhz1q92ayAAQCEZRh6dD28RjeMEe+MIJRjFgcp9SFz4W6v550LrS/qQptmnm/2kh5CWEr1V37ll//af/o/R8gMu6Muirqepol3XhKKlPNRq+08WWeF9mi2XFXWKeIopdr66ydntWJ42uQI2Z6IwsB6jwimFiFpqcTIIkSiYLpZHHJOjKvzQoSBJsgSr5CRxGvknJKEM2eNsiYSHDIcSDGVc0DegZsdtiPc7Mh7xjm9GURLCKnKctDvP3z4xm/+5m9SShhmFGPjTBT8oUMJJB5pmoDk4+TFc+Qs9s5bRzH13iFCS9U4rTDGUSzCJDDGWO8QFRgxJbVgDBFPnAoxsk2+0xoxxmSZV0pZ7gedHhV8sZq2Wh2jmlZ6WFR1VukGOe0N5QEhTCnViVvttLW/d/jk0WUSs0qXcRJY44whrTZeLedAyajLAnm/mK2LVf65n/iJZj1DYdhqd5TK9nePammuJqvBcLTZbGRTEoQxJe1+T6r68dMncZxWVWGtjfgN5Zegbrfvr62S5Yunz1jTNLPZbHd3tyxLEEMCMRFiANBtIBkvigKoEltGCKwgUCsQQkmSbKl30EYCsimM6pJSWocG/b73riwrysJeb7BYrPqDUStOZNPs7Yx0kbUG/fl6ef7ibG987D0OguBnfubLURDePRxVs9XDT3/mp978FKrqNBFpmmbGCiEcQlVVYY9WeTFZztd5hhBaz9dVU5d5kRV56Q1f19Kb42Hv1uF4d3f33r0H3/3eD7713e80ljjkHaFVVWVXZ51O78HxrWVZzsuilUQEoZAjYrHHCBQm0KfYRlAo0mFxtxS+P4SHEEII2ZtxtMBJ/fKXv3x6+pJzDiibt3ZLJYFMYGdnBxFsbqyy4PXRtTiTOOcYoUqpSHCEr9+XUpoXlXMIEeS9JxQj67z37Xa732vv7IwCRteUZNmmqmVQFDwIrfVF1cRh5L2vK3l2eRHHMSIUjpagLM9ySthstmCMUUepo0HAFNJ5XjPGGOdWO0xwHMd1XYdh0JTVy5OT+0dH2WblMDLOZkXeavco5pvVuq7rXisGfKMqsrKWaZru7+9/+OGHYRjqRm6BZH/jftA0DfeOgVfMtUnGDWICNQHsV1BsbmW4WmvOKeEcWnzwVEBnAw9gNpuBLAZjvF6v0zSFS229XkfR9YS1IAiWqyKIKcbYKJ0kSVVsuKDeOcHCYa+fr5bEI+s0xtxq89Nf/MIH3/52J428t72dAa6bNA445xRR54yUkiDkvQsZGXRaw1GXc06J0NZoqaqqOp1cfvfDdw9u3fqx1x+OR7uUckyIfLWySn/w+KOqqrS2QgTW1CH1DDll1WC8c3F+FQShIFw7TwOunfU3w/Ugs9xW3FtI6JOdJNign+xkwrG/urr60Y9+BMW+McZb673nMUc38/UopSK8tvf5ZFJ7veGxF1zEcYysIZRAQQOrSgjBlFl/TUwJOAfITGsN+xWevVIaEWoQ9VqLIFyvs6qp8zwPg3ibqFxXaUEQhiF0ZK21GLGbQ4cwxlIrjJyjxHrnnEOU5HlZlTmmxFvtnTPOYe+TKMCYEhTADRyEfDVZZGV1586dRlaMk7QV1/6a8fjJ5a2qKvDmmrAMhDFoFOV5vru7C+gALKv3Hpjk0HMnhGTZGg5up9MJgmB/f58Q8vjx406n0+/vVFU1Ho8huF5cXEipMabj8f5yubzh5vid0fiDD58AgaBpmlKWi8ViOOhzjOOwdfboI/ILsFQOEfKXf/lX/p1/8bV8Of3Jtz+1qrNRu1cpTZVhjHCPBRdB0mp0k8YirrnUyhqrhWdxmKRp2mmPD/aHu6OjV+87qUPHrLXW6LtHt0f9wZc/9/nVanV2dvb02TOPm6eL6e0Hb//dv/UbCuFf/KP/vdV0jrxllOqb4wdF3rYiQTccF3hsn6R7bmum7a7VWvd6vffffx9Yi7DdW0lCKY2CUGsN03ubpqll470H/gMghddMGqudtcpZa0NOMWMs4AJ50tSN1QZTRrw1RnGGKYFK0QrOq7LBGDtlEMJBFFEeiCAhHnEhKBPGUodYUTbjWwNKKaYCMsUkSaos29vb+559n7OA2RojxplgdLzJVkm7Y623VnPO4zSx2oTtUEq52qzTNKnLUgiRRIJYxZGhjDDMwjD0DmOCOu02pfTi9OTLX/mZqsomk6u7hw+A1QoF/p07dwBApeB4Ax0nxthyuTw4ONjd3QWTNGhCQFje2dmBEh4weRAYQM317NkzkFZ2u92iKBgTzrmtJRhjDNLw+XyeJAnkBlKZF6cfycYih4UQRuskilXdtMMQOxfFYnl25mWNRIBIYLVO4/B/+zf+N1/9O39n/+5hstNbrIrD0YGS0jcVxYQ67CxKeWicRcyHOJBSFrWniIiAC0+VMseDW6jyFHNnFUaAENh2EpEw2ul0Hj54UP/UT753+uyXf+4rr3z+86U2IWuZWlNHolA0TaW9J5xZa4Evt2UqCSHszXg+SED9J2gf/obGiq+H62mIoEHAMcZhGMCyY4zB22K5XELaQBjdZgXD4dDd8PQwMkEYeO+xs1GcILD2tQ45D20R44zgRDYVpTSOI06ZlJIi0mm1VkUmlUEIGeOstbXSlHGpTdPo1XqjjRMipIRTzpHHCCHwfG2aSmvpjfUOW2e8Q8455K+5tvbadDvigXDGIoSqWlproiD0zjCEraxL78Mwruu6adpJ3FpMFu1OLKnttMRqmZdZ/WPvfDbfFNBkhjgIdhWcc8YEAW+MbeIIN0tRFEEQZFkGlFuQYwP2dEOuQUDjhXACUAU8htl0YY1varVZ54DNNrVKkzaYeUPGw0T8yqtvfvzs2YsXL9770bsUk7vHd+IoQASnrfh4/0BVNVK1kRVCjglBKX/ttdf+yv/srxoWxKMdE6bx/Xu018UHexsRrBFyrcQnsYkjEkVMCEwpw1ZXuazzstqUdZbGIqSIOk2VYdrGmHWDuBvErTBMRRgKEXTaX/of/plXPvclhyJBAlk3dVZwRuoypyEn/JoRtw2H/mZ07CcRUHNj3rRtBeFPkFNBMr/1wtiSxBlj8D28DvD3/M3oXggKCCFCABnw7TSKwoBzKgQj6Fpcb7XzHnNKnDdCsCQMPLrmeSHkq6YO4ohzXslquVqdX1065xCmHrHZYr3ZlE2jKBOEEBiu0jTNzd14K45TjHEUCjiQjDEeCBFETIRUcISp1poQRjkL01R7gpjggQiDeDmfLqeT9XQi61w1xWiwM59OtbZXl0trfRimm1XR1ObyYgac2vV6DRSOzWYDHr9FXl2PtwH8HO4va22n04GEACB6QIhgYEAURWEolIThYrrT6cC2BmOdVqtVFhLQtVar1Wq1tLbeY4Di9/d2pVaFr9JWG7Po29/67mi0P5vOs/Xy0XubvcEgDERdFroubu/tnj19cvjGm95ohximzOFwZ+/4T/2Zt7/7vR/Uxnzzh9/jnP/Yz/9cR/qX3/u+RpIxFqWxXM/r2bzVHYi8lEVTayWxDvut0tS+Nl6bXtgihFDGA8CdEUGcr7HcOTrovP6Wwcw5a2X9g9//VrclvMcbK2tZOUS3d/pWH3K9X29ITP9GGrrFTWF7QdIJmIb3FtA+aNYTQrx1gNMBKRvYnxB6oVSCaBoLnqQxQNIEYYwwpsRbq7UWVBDOlGqQs+1WipwNOE+imBPWOC1lTQlqdNPqtAgLlFJJO0TMV4WulM3KJk5boBJ2zhNCOGXYosvLy/29o/5OD69soTYI4Zu6UCDkwHp8sZxZ66y1YRhKpcIo8ohY64qiGHS62loaCkYI5+zR43d7vb5xaDgaaWWN85yJfrdPMWVBUNd1q9XKsmxLMJJS7u7uMuC/7O/vb5nhcE+B4ysAKADI2xvDWEin8nyTJAlIOs2NoxBMV7916wgkbIwJIJVijNM0ztebWumsrOLO8B/9g/9Wa5tnxWhnx1tHnanyohsGlGGtNcfo5MmT/miUDPcs8WHUahqlLXI8+vGf+spyOul3uihg1iNMyfPZ7Es/92XPECZkcRKyMK21NsZzT7QNgoiJIECVbBHXaoXeO0Swp4RSigjlQVQjzwad/sNXDabYWoaMc/Xf/+p/1e+nk+lMW0UxNdZBBN3WmNCK/CQYDP+53akQROmN4w3suZtOhwJ7f2ttKATnHDkP7imwFyGmwt4F3xHOeRBwiiQhiBEGbAF4C0aFs4gx5pCrm5JzqrUk2Pf7g06n462lgjJCrdVcMGOVsxZhPJvNkrRPeHeT5atNNhzvMhFSSp3T2BOMKUKo1WkjRu/du/fDdzNqqUZOCGGNF4ToRiKKKGfIE2u1R0QbZ4yplHaWYCK4QFlRaS1xSRyCsW+BkhkXsWBea6Wk2qxdUdUg2AcVxra7AfloUWZssVjs7u4+f/4cyu1WqwV9C+iNBkHw7Nkz8AgByF1KCS8XhgLQY6icttbgUURBawbXE2BV/sbtyCKMymq9zr797W+Px+PFfJ2t13EYjbsdR1C72wpJqyir9Xp9l7PJ+flB0vJpV2tNCa+axlFONCahUJtc9NseW1ypiGHEGPbWIzor9GvvfNZ6R5vy6e//7v7hYfvubW00l/r093+gPVUqQwQzQghhmPDSWR+Fo4evl5gQqyKLyzz7+IM/YEYnrZhnQi03GDNKsXNAQUKEIO/9VgTnb9R5CCFAPPANY4tgvN2szjnQGhBCtJZVVQHMVFNKCOl3e1tbZLhDt20kYxQhSAgmBOMEBwGPwtCb64JMKeWNV0pZ5BFy3nvGabfbxsj1B11Kad00ImB1XTdVgZF33vEg8B4FUTQe73/7+4+vphMqgl6/zzlnjEAv3nuPMAqCaLVY3L9/f76avTxVGOmmlhgRhJyyxirpnAN3RCllVRWcc0QJZezs8mJvNGzF0SDZkbIuihJTEsbcWt/IWsl6Z7zLOL26XAxHoyAImkahm4EZQRAsl0sAKAfDPhsMBnAQId2EYFBV1cHBnpSyKNfDUc8Ys8kWYRgaa4KQIexmswlcRsByAl601jqOY86uU37vrXM+SSJg5W02G49sFLd7nnb7w4DxJBANp0l3ONoZLiZXcRSsl5uIs4SyUrrf+71v7h8f3X3jNSk153wxnwRh5JtSE0xX6/nF5WDZYbsDvypHiKCycO3QlZWbL5F3Cjkh4g9/9EG720eO8iD1sgqH+y7Ae/d/DGGSPTv92j/8J60k/exXvtA72veCEu8dsavVxjf6D7757mce/sT3/+CHuWhWNC/rCjNPwFHHOuo9xth5RzDx3lrkHbreoDxgHjuHPAFxNKYNuOB6D0UnYdR5T7hAHklppDRJEnFC87paF3mv1xMBU6oJgqCRFWfImooQEggSBsR7g5FFnsApqRoJhbpDiAihjAZX6E4a9XutOI6dMY3TjrokSrFFXjut67psgii1Buuar9fm9PLKM4+ZbHUExY7gWAjeNI4JzDAVSRvhyGj35psPEXIfP30GeZ1sDGNMUEIp1aoJBWuaCmMcBAFyylPSWDPL1lz0iVZJFJVlSZDvxT3jHaGcUvbsw4/3Dg73dgd1WWaLeZqmQRRlTZ50uk9PnydxXClplNps8mtDpTRNwXQKSEm9Xg/Ip2Bh1e/3gV4KWxAa9CB4gOoSgnOr1ZJSchYBetzpdCAGAMuEUso5nc+nYRgzRtIkHo9HcRAWed7U+XCnRynutBKGiSoKZfhBd/cnP/f56clZd/+uXm/Ksysx2CFhTKO4nK3I2Tw/myWTeTNf71jiPn7mOyk72OsLhlabIOSuVA+O7raTjnFOS8UIGrxy10W80hp71759pLrpj/+Jn+seHlnspXGUE5evZbY8f/by6YfvGWMmV5Miy4Hsxqztpu1KNo2VBCNMiHbWWossYoI7Z9CNMtsYgzzCzmPntVFQwWCMCcYEYewRRhhh7JyFZqYQjNzMZCmKgisqME3TtG6usyylGkqvEa5IEO+t995aRQiHyGKstsYxjgnFrTQNI4EJaiXxajkPuAi4gDqMc+69jaMEeWadC6P4g0cfWesopd1uOwgC5CnGWAiWbTZByDkNEKWY2my9Idi/8847nAcXl1er1SqNYm+sMZB4BFpIYxRAb1YbxyxnRGsbp61+t1vnBbCLlGrSTpsxscmKtBXmmwVhIAVgjHhCEWdBu9V9662d+XweiuDo6KjKC7YVfG0pZMAbbRrCOAHfJfBG5JyDdrHb7cq6AZ8woJrCdQNi573x4bWW6GZIyM7OTqfTmUwmzul+v4sxffzRhz/5k5//+te/7rTxziXxDqFGREGpyoDwtJWusk1rlX31b/+9z/zE5x7SVj2dXX30JLzj+92ByVbyaumySmplyoJoJ2u1ybMNsjxJ8otZm7Bg0KvOpm1HUN4w5aiTar3MVNM+vkudCCJ28uz9h59/p3U0rrVHhDGM1WZls/WTH/5gen6ZBvTDp0+Xm422NhYs4K2dNPniF7+YFfmvf+03EeM0CDVyZV0hhAimThtoKGdZBvQUjBAC8QnC6KaB6Z3zN+JVfjNKZTAYCCHWa1HXdVUUlOLDgyMYBill42/odnDtNrUKQi6lZFRocy2EQoiIgBtdE08Qct66/d29IBDOWGdtXdcB46Db1sZpj5y2cdS+nM3PJ5cO4267vbu7hzyBpJYxXpZ5EIgoSMrVkrEwieJQsEaWP/ZjP/ZqUXzwwQfnlxeEEKtVVVUojFTdYBSnSWKMIWAbHUXSyOl8wRjjBMetVAhRlnl2cRHGUb+/E8dxUZXW27paIYSCsFdV3jlfVU1VK2OcVuXv/v7vCcwYoBgAEhljXr58OR6PkyTRWnpkoVFBCOl2u2dnZzBUGWbXMcaKogAlHqRiZVnevXu3KlW/3we1FDhkQPk1Ho8bmSvt5vPlrVsH3/z9b9+9ezyfThnBHlnKGKHII98YyTAxxqrGUMo+evfR1cUspQRp9ULb9SqvlW5pHFrjPSpVQzELogBTIuu6G6TWuPzjk/JsatZLToU+nQVRuy5W0/OXgwd3sG6wEeViTZomJRiXVRR2tVFaayqbxWz6o+9+1zbq8fvvVZUKkBWUdOKovzPi2pw/fWocGrQ6mtFSKqUVdNGUkkkcM0otMNhuDGMxIY1pgKUGfVHAAYCmQggL4mRnZyeNY4xxGsdpHEsph8PB7nA8m83AZVIbiTHzHiHkjbEMIdnoKIrkzTgeQohzSClFCaYUc0JhzvF6vSYwC8JcA7EIY4u8sWSnv9fp9H/vO//UMUYZa7U6rbSjtQ0CDgmelLIV0t3RTiQC7KnBdj6fDXY61hlK6b17d3ZGw9/7nd+lQkDfst1uX15k16Jt7Zw3rTj2CE+uZgd7+4h4RnkYRyKkdV1jTEFx2aiaOGQpIYQs18vd0S3jvaC8cdJbRzlp97pHoz02Ho+BbQRQyOHhIdRJgNsBh29vbw8E72DF0Wq1AFIG+An8F2DudJZlcdQGSxxCyMnJCXhFNE3z3nvvvf2pN4qiAvOJLF8XRXEwGnPBCEHr+cJayzwVkZhcTnb7I0yY0XbTrDkVKKDOm/km29FI14p5XFeZUsoKsl6sUhFqayprUyo6Igo36zRND/pJpJW07vJfXlJuIsHscoZ3djwOvKqFUsvTl7rXJR3iHHayrvL5B3/w/ne+/T1GuEOozZmIo53h8LOf/ew6z67Op2eTySbPumlymS2TJJGZrKUKo7iShVW60cZ7HzBurYUr3lkr+HWjHCHkPdk2LeM4rivVabUF4/LaPYvXdX3nzh3OKbuetQAWctQaj4kXQmDsgcMKqnN4WTCINMYLQiglBHuCnKwb5LyWmlMaiaiqqrwqIxEIHmGKJ9PVx8/OPedK626ajscjjHEQRMgzLgQhzKHau/Cdd94x3mFnvbeBYMh5a1TAeeFQr9N96623fu93vsEYs9pQSltJW2uJOXfec8IbqQRnVLCyrvb3x5wy64w3Lk6Tsiyvri4Gg4HgtK6l4NQ5h7xfLeeYBCutKQlacWCQj3iY5WsG7Q1rLQjYwZ+SEOI9MdqBigAuGqj8McbQHYU7HdJNyLQgCdvaJsLssC0V/M6dO0AGI5g658fD0QTjJI2tNkpLh6hVPmAs5cn44WfW6/UqL7xzptGL9SbgLEwC6+lHL2e1lMY565RzLooi4yyAut3eAMXhN9//oJu26kru9+JX92/FYRIFiZF1O2Sbk4a0OrzTQUaX02lxPp21nu8+oJrgvG7OXzz/1f/v3xGMlfmaYvzgzv237r/aStKjo1vvPvrgSpnFdNpo3T/c++j8ZS9O4N5QTb11J4X6vWxqKSVMBmuHEdwtUEF6giEdjIPQmmtGXxBGCKGiyMm1WcbGewLaJqUM50LKGnsPZChjTBCKKEq0lt57TDzjBGGHieU3rZ1erzeZTOIwujg7O9g/tNYijLrddlUUImxRJD56/0leVZqg8a2DmEdJ0sIYI4Sd84yx9XoJoqlBr2e1FDTyGKlGVtQZY0QQJWFUNnUrTgaDAQy4v7q8ZIxgLDabTRAn15wshLzHl5NZGIb9fhdjzHlAKSWE+ZuJ2u00yfMSY8IJM02NkaK0okEsko5qFCFJ2k7Zdtq7MWY4HG7da43xgJuAWx1A98B12kJRQoh+v885n06nN2IAZInbTrICgSig02maFkUVBjHkOmcXV0krts5RzuaTqyTqyFVJqCQKR2GraTRjrK4qR7xznmBcl41zvlIYUWo95jzgmIRhyMOACO6dmxer269+Gj19fD6f9rrdx2dXBpGQi0FvZ3ewQzGS883ZP//tvf195H1xOTFluWBnUburGZ6sVt///e/LzeaNT3+KMZbGrU+/9maHpYLxycm0Gw9ms+8h5/qDHuc8jONNnlmrkfcUU0+pNgZjTASHUh26ShhjyFpAX8wpQwjB7YycpwjrRjaEME7CMMzzbG9vD+xFsywDN+StCAQqd7isMCJaa4DKYfamUg0hhBAUBYIg463zCG82m/F4rJVCDgetoCgzZQzy1XS1XqwzJkImqAiDdtKG1iwlLIgi8DrA2OdFdu/ePcaYNdZbF0ZCaxWGoXUOE2KkiqLo05965/z8/OnTj7ctHmOMRwRj6j1yzjvvGyXLWoaNIoRQ7CixcZzKquQ8KPMiFLjb6lprmTQ0Qdl6QZGpy8rYKm4NGPVlWTMgfVJKwSgeesqwLmVZYnINIH+SD0EIidMISHTAjTU3dsbQDgEQFIA9qJOKomiaJgiiuq4wxq1ut64lpqQdJ/lmI6WKAqeUijhTXnnrrbWNlPDBaiWDQFhjjUXWY4Q5CRijiGJknfPWOG06SbpeFpdXUx5GInLaOGXwPCuEEEGrQ/JcxdSUhdH+/MlzgrGvJCc4X2fL6dSHYj2br6fzpirfeO31JEnmk/nzZ89+7PVPW2uV0pu6zvM8SpI0TSXylFLiLZxAjHFZVVEYSimLPP8k3Qkh5NG1NQPGmDJMKRUBk6qGK5tSip1vyoox1ooTZF0YRo0zRV0BQ++TojmMEaXUKAWTfBH2GBPvMWwLup1xb733HuR4YRjKWnmMnXPWgPE0rStJCCOEUUbzTXaws7d9I4jQwObxHrVaLW8QodRZFIjIymukQqtrj98oio6OjpI4PD09nVycK6WyLCMUNKvOWmctMQYpI0HTF7CAUhowzpKEIMoIlVICP5ARzogTQmhTI2uNUs5q7J1SigUhLcumqrNur1OWJcLGGB2G4Wq1Hg8PvPdX57MnT5688847nDOrdCRSToU2TmrlkFdKlXVFOcuKvNfrNUoKGjnnQKH34sWLu3fvgsAFYyxlnaRx2dSrzcoxk1d5pNJFkSmL3HyVhlGuZa/Tn29WShpCCPbEGU+YKJVxiFhvG6uxtwghwzkhJAnDOAxanc5f/Ev/E01Qtzf4x//wn7z48MliMqUsna8WgVDSnjqtAkFiIfq9TkSYt07WyhhF8+WL1TSO46KqTh8/Qgi9+frrHz/+qFyus8XmedxBiJyen1/Op3ldHdw+cpwss1UYR14bL2USRUrKoMsppQejrleOIrwp8sobxpC2hmNhnTPaEUrDJOq2upvNJhGJoGKv393k2V/49/+9//L//n/jnP/SL/3S7/2r39lsMooJplhJBSCOcxYTTxhmnCDiRZI0xnij4jBA3hNCGll579ut1GG/qbL90UCqKk7b3iJnlXWNNq6elQ4TR1rTvDpf5dLbiDjh6Lg/CEJCmKCUO2+rqijzutqUSUqiKHbSWMPDkDdOUeZ5FDZSpmm02qx2d3eXi8VkchlFwe7uLkIuCshms4liMbmaG++s90wwz4hHqCzrqigDzjHX7bRVFVUURU3TUBZ4aRwhQghV1Nq5spaEkqQdS6mdzFebWdwZ4t/59f/3arUC8dqjR48AE4UrGzsGdxboPAEKdc5tslWv34+TEAaygCcMY2w0GimliL8WhkOfCSAnAHgx8da5Rhvn8X/91V8z1lVVQwhxFiHlOnHaarWR83VRa6kECHoQpYIighEiDiNCedXUPAghbB8e3irzvNsfLFcb1mr/7Ff+2B98/4ery/mw218sLo1srFFRwJq6tFoFDO/0umB3zSmr6xpTtFkvMcZFURij3nnnM0kYnZ6e5qu1rNVkNkUIiTBe5dn+eKyczZvqPF+XViWdrtVGN1oQOitWsqlCyglC7bQlhMiqUmqlvRMW3KIopTQMQ6VMGIaIUJiAigiW1vA4xDfzzOeTqXNOI5gfck1wxtiHYcg4AUZIGIZJGHjkvDVaa+dNEoXGahFS7NHxrT1sTCsJtNa9Tlcps1iunfNF1Vgaffj8VDkCvZVbe+MwYHsHY0pCRAmlNAmTD9/7uM6zVmL+yBc/9Ut/8t/SCidJUlWV89eGSILHQRA3ZZllWVlkzhkuaFXkk+mVrGqE0MePP76az4Io9Bh5goQQgpLdnWGapu2QR0EICDq6IWiDoiaIwmy94Zwbo5u6xM4LIZRqEOcsy7LDw0Pgb+/t7WGMV6vV7u4uYyxg8WQyabfbL1686Pf73vvZbAaDhdbr9XyhwjCEmbvwh0KINE2N9KvVCkAHrXW/3wf8nxCCifcI8dAVZf3OO+98/bf/hUdgT1ITRLvJ8GqzlFLGPHLeNArUTpIZsM8UAINHUaSMdcgTRmfrZRRFd1975Zd/4Rf/0W/8xoM3Xvn617+OmX//+aMI44AzbbS3CqPrBK6SqmpWkIeUZckFcw5Zq2gU3b51L0mSxWrpGTlbLupaOoGdReezy/F4bDEqqhIzqpQiFEdBGLQDr125zrpxXGNMPULOt8J4dzS6mE/LuhJhGPHQaqOUwRgHYYgxvZpOMCJSyaZpDm8fvTw/E2HgCQITVmDxNbpxzlGKb7r5YOiArbVxHFulm6ZhlFgnEUYMY49cFEWU+n6/b63tpmldZUIIQnlVlw5h42wYJU/Pr4yzcdLy3gMO886nHmJMPTLEU4a5tR7BBEqZ/exXvogQsfba/tgjXlVVkrSQJ8Zowpn3ngdCa6+UYoIzxgpnKaW3bx+VdVnKhhCCKUHWWYSlMonH6GY4m9SKc26NbkUh01xbwxiDk7BarZI4JBhd0w6xY8PhECT0YHgHDnV1XadpenZ2BhaKw+EQyvPRaOS918a3222PLLBDRqPRcrmEGU39fh/6T8456IUC3eRaJYKdCAJPqEfsx3/8x3/jn/1mkl5zRuO0c7WYYeetdtZU2GEETsTGhIg6j5WuuTHG+SAIvEMeJgCEgXPu+O6dnf3xn/5TfzLPy/HeaHY1SdsJ1to6Q0NW5QUliCFPGZdKGWOkNpxzTAnBFFFChMDOEcLyorq4nCTtlg95kER5njtvx7duxSIoyjJIYuNdEASFapBznNJQxCHli49ngtNWGGOEeq12Xdemlk4Z7RvmSBAEYZQggjnnL07OtHcO+bc+/c5wd/S1r30NhixyyjCjWZZVsvHeGwcN52tHnRsXCUQptUoihBjnRklnDaWUMIKsQ9Ywxq1sJHImCAKRYIzX6ywvq6pSFmGp1TrLAx4JTqGHsre3FwQhwpQyRAiiDBdZQbAnDHMhut1uU2nGOHinIc8ZDZ1FGHuMkfeIBYH3PoqiIttoI7udXlPLdrtdbOb9QZtkyFrfSF2XdZIkeV5453pHe5gS4yz7xNAVAEOm02nARRiGo9FINhVCyBurlIpYQIB3CD8ECCpot5umGQwGcPtMp1NQHQHlBDqfRrt2uw3te3dji5BlWRAEd+7cGQ6H0J2C2FDXNTiYXV1dwTD3Vqu1v78HBwXGiCNEpNTGGIwpxtQ7jDwhmClnrTfGWesdwNcI+07a2hvtOmmy1earv/r3/sKf//P/2f/yr/9f/ov/4i/8O39+cnmxvz9OuqnFSBtDhTDYN85cLefaO+eQ0U6wwChrPbLWOm2c8y9Pzv7oL/zCPN9cLmbrolRGE0IJIdSj6dVEhJGIQot8EASMkDiMtFSVbKy1r9x/sDsceYLrpmm0Qgh1u90kSXaHozSKq00+v5pcnJ69ePFCa91K0lCI05cvv/GNb0Dg8d47Y6WURVkqZ9XNM9tK8LTWgDExQkXA4yi0WhkNOQAKOGOMBEGgpCw2GSPXwymttcv1RmkfxIn0+Hw277R7adq6tbdvlSTe3bp1y3vMGQsEI9incfLow4/ancS65rVXHiBrjXZhGBKCbtiovGnU9ZZiLAiCSlZV08RpIoIoStK03cqyDCEXcCYYCwULGQ0D4ZzDHslGKeOqpuFBEMaxxzhttxEhIgzDOO52u0D7uvZcrq4V6r1OH3/rt/5/d+7cef/996Mouh7GoxRge3HQYvBRqgr+B6DeZfm60+1yQYHxBH5jhJD9/f3VanX28hL6eDBFxVoLg1qstYvljFAapa35YnU5naWt9v/z//X/qapmPB5nRW00sloSRLFHIQ+xR8D9owI1Vd1udymldSUFY9Y4awyl9O7xncvpRCTROssIQoJxQsju7u5yuWyUlFJSjJMockYziiMRMEJNo4MgUI3mnIZhCAwgQoh1erleNlIa74QQq9U6IAw5jL0fdvtxJ1mX60ZqhVFRFUnaSZLEWo+cFwEjBGVZVhVlWZb9fh/dTEkcxJ2AC4cwwnhd5lWj/t2/8O99/OTpk+fPnp6fgCMITMLdFDm8tfc+FBxfTy7klFLvbRiGmHjsEacEOe+MpgwHAacIc3ZtH86IC4IgjgKrTStJCGNKmeUqF0H03tOnCNMwjKMkJZQDz3I8HodhSAmJ48BYRUn44buPrS3TkP4Hf/F//Mr94/Xax3FMKXYOMcZlo7WREMiVdkKw87PTpmkGg56sS+eMbJqzsxPi1WI2X61WnPMsK5TWjmJwZ7hzuA85OrTTATYG8itBqC6rLMuE4M5q7ABz8HlesrIsYXduOR9pmqZpenl5yUlY1/VyuQSHN6h70jTd3d2dzK4A2oDpQ7dvj7Ise/7sJQxMArgUmvVVVa3Xa3Dk0UamrVaWZVVVtdvtoqy+8IUvnJ6eI4TySitTa60ootiRIMDYY20tpiQKI+u90ppY2zSNE4ITFgQB9mh+OdFWIU6Hg4HVpi6LJIpfvHiOCEbOp0nS7/fX6zVivJBVXtUIIZs1GJNQBGmalFWDiAd727QVI8GcUQEP//9l/WmsZWt6HoZ98/etaa89nnOqTk237jz1QPZAdjcpkmpakzVQjChRlmM4SIDYCCI4hh3BlgjI8Y8oFqwg9o8EkSIPMazYkQKIlGRKpFoDJTW72eyR3XeuW1WnzrjnNX5zfrx77y7SGxeN6rrn7rP3Wt96h+d93ucxvXnphYfL5Zpz7l00DG/nNy6aiGlvbZpmTV0rKSnhAYdtVSWpJIIXkxHm1BGklFJC9FfXTdOwglZN27Tt6YN7ZFv/3/7L/8KEkKZpjzxXMpUqhLDdbtu+c957gg4gHec7VwYhOKzfCSFICFRQzBllmKIYUUAIMU4YI6lUnHOKsVQqYtx1um46hGnVtpipohyggKSUvba3b9+eHM1UmnJCBUXe2kFefvjBU2e1VMG4ejoaBk844zGCiKTGiIHsKMJeCOG8CQEVgxIKP8xoLlSSJNfX18GENE03yxXDZDYezVfLTVMTEgAehoqWS4Ep6Y2WUgbvQwgcEwCJEYrBW0DyhWBCpTtLTAiEy+USREv8c7anwFEC7hL4K4CZFVDzLy8vt+t1DBYRPD0+IgIvVkuM8Ww2y7JsMZ93XWesHQzLpmkYxUII5ENRFGkx2Fa1VOlkMvva137LxYAo4ULhiLS22jnd9SEE7R3iEYUYfIAhIUHIWouBu+6D56Rar9fbbZGlgMu64KVK2k0F8R7k6wlB2lmKKGGUc95Zi4z2KPhgY4xtsF00jFGVZrbXs9ms3jad1tq76+XqE598y124vjOcE+QRChGACxSJlBIz2hrtvUeUyCxtdS8Zab11BG2svrk6y/JBQ/w3P3gnSfOWIy7TJoQYESYEM7perauqChgFsmv4YY+JI7SX0AkxRqm4c04QFFyApMkkZ5QrSYHFG4yO0TvnESFWW2NMjNTF+PTsMh1NCeZMsrqui3IwO5okWRZCwIxi5Cjh1saL82uKndXVl3//l1QigqeMsa6rk0R670EwhjEC5PlIdjxXLqR3OsbY9yYEV5blZmGUSmGJKsuTcphTjmWaCKmyLOVcRIystQcZnIPwB0wfrTVV33NCd3TNaFiSFTHGs/Proiis8bdOpufPnrV1kyQJkbRIE5j8AmYUMcKUIEeqzYZLwZQcj/I3XrolFckH2dlqcdNW06O7puqX1Wa7rqI1lOHZdCyLYjw9atarxfUNE3xdV/P50vogkny5XFPCci6D9kRKZz1s+xRFjjG2pu+NjjEKxkMMwQZCSERBUDYY5KlKFqsl8ub2rTvIu7quO2OGedb0nXHOep8J4UKklGBMdd/FaFNGrDGEEBTAscBhjLXTxpssKymlMs+X2woH7L03ffvpV19BvSkHk6azMVCPTK89IaRta4qJ1lXCFSKYUYJQRMgjH2CNLhDcYquGaVC8GCZxu7Uxgisx5yyEGDBqjd62jYsBEYIR4pRRQplSjLEYg+CUUopxpDhijDDBCBNtes6ITNPoHSFIKTWblFVVCcYJ4tbrpqqlypVKTaTPzm9efvtTHz6+NH2UCqm8SLKsKMvtejUej43RiOPxbPrRo6dUkkzJVA5+9NOfdIHoziQMkeBd3ylBJY29a1RSOkej88wZGzwluK7aLMs8EFK8870RTDAUR+OyrmtruiRJsvGIcU4pZd5b13gmMRGMCUKIty6B1QCHPPF91znj0qQIzhJCnDPROeZMNx5P8lRdXl4e3TrBjI5mY+vMcDhY3dRwLoHqDGjRYDBghHbVtjPa6IARv7q6evjyPe2sp3h8MlVMem8xY7g35XjYNW3T1Y6Etu0l5mmaykSZ4K2JhKM8y7/+9a8HvxM6lIQGjPKisMa44CkmAUVYRIEBLKHIR2+9Na1uTDsejjzyMfrVZnn/7r28HLz//vvV4oYSzvcV80FMAXa9276ilHLKW13HGBEiu3bE2arRXFDFRfSBMzY7PkqkGgzytm1NH3pw+srTp8/OkkRyyRDG3ofOaIQQIjhihBBqjdbOwkK6ULJpGtu2kN2cjyEEHDDlnHMRQlgul63uOecYIQC/4l4dHGMcgyOEcIIdDoRgSqmxLk1TJbjWmmKUpmo6OrQXmjPGKUVJ5hFdzldU5VIl600FMEtRFOvV9etvvOKtgZagyHKr28ePnz55cuac6druT/zhP3h0dMJFtt0Yjz3nPAQHHQjGEWaq4I+eFXlv+r2WAtZaR++Ojo7Onj4GSFEIYXTbdV1ZltF5rQ0th4Qw62MMAaIyQTFVilLa9j0KEWEMMmOHJQWMMXOov15cYMyNC0mSee+TJMM9dm6X+iEUX1xcgN/3er0WjPOIXfBYyr4zR6OxNsQZt6309P5R30aaKaakJW2HosgHBLO2bcfldHlz3bb18a2TRMjhMIOVQhS893E4GWPOVptmNpst1+vewWoORSgKzISgo9EIlJIIpREhgnHA+Hw+Rwi99PDFpmk+fPxEKVWMxsv1ynrPMaKc7QRqvJNS6r5HCKGAIorWaefcTo1S7BWQEY6URsowQUKJwWTkrbtZrfu+D4gRyYWSq+2q7rvAsW0DIYgRTlHAGB8c2ynltre2twEhyom1ViapDR75kKvUhRAQJgh57zebDXyGw/4d0J0iRpRSpSSUoZRgIQW0jIwxilFEPlNJlivJhTFmu2kIwUIyFEJvbPA4HZTHMn18PreIWeOr5WVRFMvl4s7p7UTsWhPYSJ4Mjz765ne7pk1UmJTqR3/0R50NMpFKIeQRxkhK6bzRWhPM4ANYa4uiQATjvY2TlCkhpO+t133Yq6TANhVjzGpDEU6VijEKLiinIe7UA6BcxBinSQ6uYxHDclWE/ViMMcNRZVkmGIvGV4sbwDUYE9ZGAOcfP358enoKfX0IYTKZdE17/87dEMIHT55wzkVaXpzfOOSCRJdPngxE1rSNj2E6Pr04W5jKci6jCZIk49GUMtb3fYixaZoiL09vHf2RP/Dlr/yzf76tVjfzFZfJ5fUFwowJ7pz31mAcg3MY4+V6TQip21YmSjub5zllLFrLKb24uaKU9lav622SJD4GmSind57HUH3D2qrWmlEWoXPkgnDW9j3QOimlPmDsvPOBYIQ4fnT2DOFgtSERRcx19Kbv5putRj70PWOMEcpYZGSvEIYwIYQixAiFmxRo9M653gjOicQxRMl5QMQ519Y/tCeFGA9Rh3MOmrsUE+RdCMExjDXFJIYYGKKcQyvvnbHeGWticG44HDpnMcaBYOfppu6apsNcDbPiZrUeFFmMLk34Sw8fOGellChEjAml5GaxWc2Xea4IXv/BP/CHtdZJknVdhzHiTBjTe2/Z3hACLhRgFJxxHHaL0TAsFIy321XbtihYYAst5lchBE5ZQIijCKoChKteG4wxoYhLAQqeh+1thJBQ0mrUG0MoyvKMpDHp5nU9X5Y5PzpSt2/n9faCM5ck8unTp1dXV6PR6OLiArRCgWBvnJ1vlxfLi2IgR6Py4uqcEGY7e/f41lSItx6Otpfv//Nf/Tt19bQcsPnq2nt769bxRx8/0tagvZYLDr7arJZXNy+9+PAXfv5PzsYj5+x6vV6v1whHoQRi2KGICEuTnMuEUo4IDYSKNJMqrZtutd5GTKhUi8220aY1VqRZIJSphErlY2i6VigplAQMVVtjnPUhgIteREgbFyJ2PoaIQ8SEkICitqYzerXe9t5umnZdN0+vr59dn5vgPcWTkyMipAnRRxwIdQjXxrXO2Rg9xgiR3rjeeR9QiBgjyhibTCbD4VACYklI27bb7bbpd1Af2ZuJ4b3eIsKBMkxQkFIWRcGAgU+wZFQIgXAI1uAYUHAkovVyJYSIyBOG0yzzAVOpPGbax4jIfLFKk9x7l6RyPBwoJZXkFBPGGKVcCPHd73yfkkh8R1z/Y5//dNdpRqXu+x2Z0BjQ1jwQ++FBAjEYlaWQkY3RWmtQowArjxCC9w7wL/gWUB/6vV8UtEed7h1MjDjjSspEMcG99z6G0WhkrfUxsDLlxEfMwnSa375za7upieCLzbaMDLIPDB7G43HbtuAVsl6vRcJijHmeG62VTBeL+Xe+861irCbTQa7E5z/7idPb084069p0XbWtkuPJFLpR55yx/WAwyPJciWS53mjrBOfT2dhazVhiTM8I9d4LJgKhVhspZTAaU+KsHQwGwXlggrZ1wyXHGI9GIzB9AtEe2Ck9bAgBcHPQ+fXWxv3LOTccDgkhQCowzsa9TFfAEQpBLgV3llOWZopQ7lHM83y9Xhu3k0+DftbBOjxB8XnZJrxbog0ItGoRzNviXlj0sP+J9wuicGu994hRQpD3HiMUnNEuEEIkVziiw0OOUEiSJE2SNEsC8gQTKRMXyMXFRZpm89UyzQYIoRCd1vrW8URrTQnCGBOEkA+RkBCdZMR7/cqrD4NHGBOECMM7VXnIDOi51X741YAZOQcauSSE4BGCJWnovuEoA5maUwYnJ0kSwijeuwKFvbFOjBEh7L1HfqfbpXXf6R7QdDIpm9dfLT7z+RcT4X7jN/7pt3/nWy+/+dbo1m0sEri48NCDmwdUgUdHR23f9bY/uzjjUla1Hpazz372xx+9/+jX/8GvfvT975QijjNyOs4VRaNB8vDBSauXL7x8G4SJu66rqs0H77+7Wt+sltcn01HC6Sdff2MyKPJUZUomSeItPGoRY1K1DUBFWuvpeNy37WsvvVwk6dtvvjkYDGD5Ic9zq839u/dMr5uqZoQ+3xvBIwvaXeCsEFD0MXS6b/uu0z0EV2v87h9rnfN102njCKPj6WQ6HQ8GA5UIeGIJ4512ddPXVedjcCFaa3tjOt0DScATZKK33hFK277b1pW2pjdaW+uCd8HDnfb7FxxrwPU4pThEzhglKAaHY4T587AsOCOH/7DvW6O7YTlAKGzXm651be+1jYvlNsvLhy+/plSS5FlVb1IpxsN8Nps6q0lEbE/Gu7m8wjQK5YUIX/79Px0sEzxHiDDGBGWAWcK+LoDq8KthdgibP2EnIY1ijAeHOM455xyCCJTXcW+bgTEGdgi8CaQUaABc8J3RMUZEMDwAMlHee1acpk+fXDx79wcBJUfHL/W9WSwaTtWTq49vz24RQvq+P6gMAFsUtDGEEMNsgCwuymkqRCT0jdNZItHV5cdnl0/e/tQnvvKVf4FIeff4ft9aqD9CINPpNMbJcnX9mc985tnTs0FeVNt1lqrj2fTtN9/4xrd+Rwqmu/54dqSt6ZrWIsu42G63GBHORFM1uUwl4djF7WKthHr15Ve/973vwRG8vr4ZDAZCmK7rCWfgT6q1zQbFIVvtDi6lNoSA0M1yCbSBzphU7OT/4KW1DhhpZ8fjoXE+9r2PAbi9IQSMQJY2+KajjAjGKaWRYBRDiLtnw7guYmmMMy4g7CJiLuzDkvNQzx3gBQj2dOctu1OFiDFywYUQKHrnHMUEh0gJlooJEndVNUuEED6Svjc+sHI8qzrz3d/5wWg8nq+WSon7D+6enBx37a9nLgAAZXFJREFUbZ1lGYkImNTRh48++gix2DSLz7z91unRbYp5lqjgESxnkr1dTgx4r3cUoEPvui4fFHQvNC6EkJxYbdq21loXWVI3Gp63JEmyNGuaBlAagXeUYsgYUH9TzqGECPuwLZT0XYDFbfI//r3f+MGj+fHpW4PyQcRl1Ubs0HZ1fTwpDz1mnufb7dY5xzm/f/++tTaReVkM5zfrpjWMy0BIkpUBi21tTSwm01fef3fxybe/eOvoRYbTrnJFPkKIpWkaAkrTNEY8v74RQhljjHF10/S6/eKP/fibr7+BEIrezuc36+XKWgsCG2ma6r63xhRpVmT54ma+uJnjiM7Pz3/wgx+AP+7hWaKUHh0dQW0HrP4D4RqWTxBCbO+sBwtVdm+0FbyPIQARBbarKWdt26NIut60Tb/dbq+urhQXBMJJAG3CnTCT9c4G31tjvHMxEM7AGYxzTij33nsUrfe9MYe4cmjeDxVeCI6QHQfcOccYQdF7bzHo5XrHGMlUkuf50dHR6enp7eMTzvlmUwWPQMnWejeZTTf1djgss1weHR3BqOXgV8uleO+D9733UrCXX3zw+37yCwgR5wLZWzvAMYI4d1jpgQgFxhigdQPaM4wxgtlgMPB7rVOEEBAIYbAJBx22MACQh5+s67ptW2MMYdSFQPlOYxUoed77iBB7/aXPGu3mNxXGpNo24B2hZBFCKIoCgC7gc8Dqj7U2S3JOmdZWCNHq1qzX9+/fR4gGxFb1Flt5drkRnPMmqCxVCGHmnenH4/H15RUssgiuqrqnlKokAx/bzvUqoZ/57Cfu3D35p//sN3rruEqapgkosuiRDzEGIVTVNuPxuGkaLNiy2vTO5Hm+3q5DjN7aJEmur+YIoe2m1t55jBCjVArJ2Hw+DzFgRhHCgVAbEWIcUWqtdT5wlXRdozChlIYQvI+K04g8Ci5aHwnWMayqLee8adtiWGqtqdAgYxRCYJxjxkKMzjrPEOecUBop9c71wUEsBPo9BUie0N5jQnkIAUWPwaoGI4SQIIQLCfuZu4dKOxM9I9TGEAkSnKiEZTljlFrXhygvblZcpEIYSoWSybPrZaNtCNs8U7avX375RZihMswkF5QxRqVx4f0nT2bDceaWf/Rnf+5oNGOMhKg5YyTg4BlnqTZ9jFEIBVoHMHSt65pSmnDS1Rsm8kTIrtW614HgNE1RFN47FwOlxDnXNFWapkWWIyTbtmVMpEluUTC2dc4Rwops4H1sOq0wEkJEH3yMBGOtDSM0UXnbt0zw1NnWWuOcLvISTiTf2wND2Qv9F7Dy2rZVQqqyjJidHN/ZVmuWJ5vVFSHI+IAiAkMx7/18sdpsNrdv34ZnFwI+5xy4SxDeF4vFjpopGaHsZDqdjEfj8fC733vnO9//QZ7ILE2229aEeHJy4pxbLObb7cb2Os/zUVlSqWA1hTPRtm252+KvR6MRtJMHvM1aC1qbUiZ4L5YEjzXE0TzPsfcHk9x1tU3TNGIaEbI+uLaFCAf7rkBZAp4R3avfQGm4A5hgLSmGw6oMxsSH4P1B0P53uYk6ZwljGEfvLYvcWhexh1ToQ9hv1QeEMKM8RtzWfZYrKeVyuRqUE2v90a178+X6vQ+fJOUwL9K+b2OMR0fTssjRTjoqEMxt79fd8nd+8G6RJlyG3/f5L43LIaUUIeJdcM5RwmP0xnbeO7y3BweeEd/5JznOeVPV2GwpE03TCsa7rgXSepIkzgFW40DBDxiDUsqm6ZxzNnrrfui4d/gVhBDvA8aYM2aMiT5Ya7UxrK7rPM9DiNbauq7BxyhGfHp6CjcDNuPatj0/P8cYF0XRNtVmO8eEO485x67f5inDGB8XQ+3oNSbAcprNZoQQAKp2S2SMgbAj5LW+74HoBBDGcrnsXJcl+TBXP/fH//Cbb7z85MmTr371663nRVE8OX+WSDnIU91bIURT1dHHPgRM+a44Jvjy5no2m6lMnV+dA3ALVCygDgLZOYRdHwqFNaQbrbVSIoYAzye0n/AJB4MBqC0owQH2g0of763t/V7hG+1Ntg8G8b/7gGKAVHY5HWMUCSaRYIIQ4nDzGaGY+GClFJLxffZHz9WmmHNe5CXFkVFqDTq9fX+xqVVafv/djykXmEup0tVqUZZlXa1P794hO68PjBAKCKdJ+e1vfd+ZXuXYdNVP/74vtk2fZUXwmFLsbEDMO2dCNDHQ556fnYszjMQSzghm2jhdd0dHR23dYIzbtoneSikJDX1roSH23g8HJQjDex+11ojTw2DCew/WeIxRY4zpNch4FUXR1o2UkgtFwJFjNBq2bQtyDKDOsFgs4H3n83nXdWBwCCIiCCGEzXpzE4IpBxnG+q03X3jj1XuvvHIPIzselevlCgJkkiQnJyd+byECt229XoMTdZqmt2/fBrYeaJFyQq3uMEKub168d+tnf+ZL/+Yv/qn7tyaKoekoy1LmUDTRRkbK2ZRnGfASnHN1vU3zjHI2Xy4CiuPpBLitUN4B1Xrn14FQkiRApgZxHjhSCKHJZAI42q1btxBCo9EItKdRiPCBd2iUMWmaQmyAHhztNbPxXtgb7iW0uwEhH6NxDsaYQJvHGFOGOedSCikF2GSBoSrYe4L4+qGR2r2Cj94JoTiXwUnG8u//4KP5vHn3vSdUZlVr82I0n8+zPG275nOf+5yPMew8cXbd9Afvf9x13XBQIt/+B//+vzOfX4HFm3OBYMa5lDKB7kfrDiyHoECHBx6qZIwxQiT4CA9/WZYIxSyRlO7wNRg6or3qOSSrA0xxUPCDLj7uldgOZS4Q3jHGxvYMIUIpWi5XcJJALjTLMiiNYZ0UY3x2dgZmXzHGNMkJcbdOptutvrpcehuvrzbz+dxYfHaxfPGl1/JBAV0bWDtCWgcwcjqddl23Wq1ijPAb79y5s1qt0iRJkwRTtN1ubx9N1ut1UQ6cNdNR9md+7l//9a/8k3ff/4gyIVU2Gg4IT4yxbVsrxpyurTVlOag2m4DioSpPkwTMo6DVe/z4cYxRCAHuWW3bIoSg84OevW37Is0woZjQqm4QJpdX14SQpu0451XbHFYfD9cajg48BuE5iVC0N044hE94EUII2enJI4wBqeWUIhwYo9Zq5w5yt4Hi3SwRIcwpg+AdSdTBXK8WqUiW8411QaiCI6SKybpuuEz6vk9VUmR5dpQxFIN3mIsQHKXS+eiNfvzkw8FA1PX1j//YJwQnQROA05VSTadJoPDhhRCbdQO75rBhAWgjPO0kiDRNbQhlWTZNQxDWupOcC8EYzbfVUgiRJFnjLRw7UELVuoXAmec52o92QUmZYgI1urXWeA8Pf9/3hOKdgC0hIcuy1XIDQ78sKy4uLkajUZqmxhhwkwHgNISwWCwGeTmdHlXVRV5kg/zWarMJqAg4ZANsnAUnADBdOD09pZQul8u7d+9u15umaUD0HnA1UGhhjIXo5vO5ECJN0/Pzs70ippyOJ96GP/oHfvbHPrP86OMnv/nNb4Vou2rFuQjerNa1EpIrulnPrfF/6k//wne+9/35fLFYLBpeQ0l0YGDMZrPFYqFUCkkAxBAPZVCMsWo7G6INjoZorAshMExijM7YW7dugaEUYHsHkBmKE7hoZO+aegCqEKbPH9CIUAiREEYZA1khKUUIDkUUQiAYxRgxCjFihIIxFtbKdpIyzhFC6rYqisF6U9fUhcjTMouEI55cLFbe9c654XDgvLl/eltKHoNlFEshFGdV2ylRfvW3v4aoa7rqaDr8Qz/7+1mM6bCAxhwT33bb0XhKCAoBSSk5NyBNivYbyQDSMcacDTJVbd1wKfq+xxit1+s0UxRFjDwkTCHEGkJvXgBmAm8SKK6qCiGUpjmU+5RShKL3XgoBNDwIvVLKbb1l0CXM5wulEpi1KKWWy52RMmwpATEeGp0Qwmw2QwHP59fjSWGt1cZpR9abjieKSgEFKOd8PB4754BczRi7vLw8ms6ePXtWliXnfDgcPnr0CCRc2raViciLIoRAKP3Up34EzB6dCzFazmSMeDqd3rl37423Xl9tN2fnzwBXunN8enZ29vVvfOv/9J/85UE5/A//wl9kKoE8AKSY5XIJO4Sg1dH3fYxYa30w04Y1c8i8EFbDc+bEeL+E0DQNlDegAIr2GmDwv8+/DugM3Cf0u18/fHNGnLMxOr7z5PQIU4RDDBFjYoyN3kspOeeEYCaYM7bttSxGVWtOjo7ybLBerzFnzy4vRZr2fX/n6HixWKBgX7x/T5BIgieYMMKiCZu2GgyG3/3ee13XZYm9d3r0C3/qTzqrU54xLJ23SimEYkSO0IOpiACNY+gj4ZMfIFsbXeg6Qpi3oSiK5WJeN1upqDZ9qhTywTjjvQXrYTjTECx8QPkg9yHCuT8MKWQifs8lgiuvVMLath0OR0VR9J0Zj8d9319fX4eAnHNg3TkcDkGdwRgDRrZN0wzLLM2T88tnmDLF1bbeZkXWub5z/XBy3NUNKNg/T4Ow1r777ruvvPIKaOIdRpHwK1rbZUkyn88RQh89epwlaZYWEKWcAyfx0LZb7M3xZJBwdHQ8wxivlvW9u7e2m+qXfukvRkRffu31+arKsmw+n2OMp9Ppiy+++L3vfW+zrZM0N9YTys3e3RXyLNo33RAeIF+j58Z6sHfV9z0MQtDeHfUwz8R7b204oNAOQvBzPv6eA0r3HiucUsp2kDWJKOBAKIoRI7zD5yELhRC4EDFiH5F34fqmmkwmXe+21cV0evTRxx+Xk5FxLkkUocH5/u23PsMJllxYZyjI4oRQluXNzfzpo6dFqTDq/9f/m3+LUSSi5EQghIqiCCFU1YYxglBw3iNEUMRlWUJFDgKGUGPQnU0ZrqsWPGHMpmeMZFkyyNO+xQh7uETeu6qqlFI67lwYY8Rt21MlnIee3acqgXf20THGCMJSSorxer321qVpKmTCBC/mN1tj3Hg82Gwa732eD0HAG2AU2IObzWaXl5fr9RrOaF6OQwgock6E1nY0moA5A7JRtx3Mx4QQ0F3B90yS5Pbt2xBTlVLn5+ewLw/FnPSCEc4Ix5EIxsuy/OCDD0IIDx480LbpdAuFEaVmMpwpnrreFUVeFFhK2Wj30dOnVd1zJt9848WrqxuM5NXlR/PrC2vtH/0TPz8Yjh99/OS9Dx799m//tlA4SbhAoe/7EH2MOBhDCeGcRky994jEiCITsClGtO299zG6gBwhBFPAp1iIOIbICMOIEcojIjFG50NEBJPDhP2HVekBk4fOgMRIwg/t4ylGITqgl2FsMOaIEo9ijNF1xrvoXLDWY86Hw1EIIWLy8ZOnSiV93aVpIqUc5OmDz3+WEMJE4kNgPKUcB+cFF+9/cH5+fj6eZm1984t/7udyIawxggvF8+h1tdpqaznnnPC+rimllIQYG46VIwEhkqaKECaU1L1lXGqtHUGBoL5pJBfWhdVipRhpm62UvNvUyDsaSfCYUxFcJJxRyp0LVdtKpVDEFJO26wZ54bBvqhUnNJGKJwRYlBCwoGPxfc1Wq1UIAawKIZGB9l+WZdvt5tBRguE21Mgxxvl8Dr0Y+IbACsThrYuigIx59+7dy8tLKIThYT0/P4dAdevWrUN1OxgM4BYWRQG7e9vt9ujoKIRwdXU1HI+UUmAWCu8D8pk79XulTk9uffGLX/za1795dnaW5JOsyO/df8GZfr2eb7fbv/0//g9379/7uZ//xcnszh/747+QF8mHH77/t/+/f4sQrfs6OhOiwzEyyhhPASHC+17lAHAihGFEAiy1GH8IdhL8wzB5iKbxObfjwwEF5A+6AUxJCMEHty8SwAosIAQnlTHCfAAsJhhjjLZSJmmaDgaDR48ewS1TSi2XyzRNf+RHfpTTEGNs6jpJUIyRMWKMT1W6Wq3e/f47gzLTuv7RH3ntZ37qS9Vm7V3ElHsfcYxt2wagBOHdrh/hzHlbaycStd3Wg8FAa4v2YMKhBF+tFrdv3fKt5ZQpgQlB0VugKXnvQ/CA2QHzIaAImQGeWJhQGmOUUplKUIjheem/roOOFqFIAHMBqZ3BYBBCGI1G8NagWn9zcxNCaNsWdj+eJ+PA9QImgX3OGB36HowxFKBlWe7BCHR6ejqdTsE/BOpaaOrh8wGpD+BS6P0PQ0g4K13XbbdbqPMAiocp14N7d1988QVj+7arg/MEx0Ex5jxhjPlgz8/Pr6+vBeXL+Ur3hjGRZgWi1Pq9Qds+Xx9a70PK/l1dzv7kHY4jNNqHHyZ7o4/Djx3eDe0RgMMLhn6HkQEhhDHBmEA7DwMSI4Zay7uAEIGoAdsNUBqt1+uyLGFXbL3eGuPgfBy+C6Pi7Om5lIKRGHz/+hsvad2FgBgTPu7gMGho8N4dD4arAJHBx+acP39l9oAoIKMmREco2mFGmBlnEcFMcMoZfo5fsqupnhvwNl0LKRogArp3m4BrcuifyEEZdLFYHC43KEYD0ApoJcw54TchGN7sfVIA6AZk3nvfNA0sv8MHWq/XjLHhcNg0zccff1xV1dXVFXTWcMQxxpvNxloLf9+27dOnT2EUC4AX6JQMh0P4wkqpxWIB6jEhhMtn501TjUajn/p9P3FyNFktrpt2Y113enorL/Oma5yz1ur/6m/89bbZpoqdX170Wr/6+hvaOCVTwgTjkjKBEPEueheDRwQzjChs5TMqwHMB/c+m52jfxR8CZ9gTlJxzUIw+/4JLB9eWcEG4oFxwqbhULqC2N3XXNn2HqXIBd9o1nWk6o22wIaZFPp5NT05OHj9+PBwOARgihDx48OCtt97qum5QjLS2wSMYuFtrs7R45513bm5uioLWzc0v/ht/4rOf+TTBAiMeA/M+ImrRHhqDzwZRw/uIMYWpIaCNcOhh7A7R6urqKi+yqtpYq2P0hOy+F4w5IO1oa3z8YQiA1+EqQWQE5B/4eIfwCSzSXUxN0zRNU3BtguNorYWeF/aDIRE3TVMUxc3Nzb17905PT+Hn4XN778GJAd4aUiSc3bquh8Mh/NbhcDidTiGPwxcAYLXv++l0KqUcDofe++12e+/ePWik4GGFYHxwsAW653g8Ho/HptdZlsXonbdS8n/73/63vNemrxjxZ5ePTu8c3b59EmO0vfWu/+W/+7d+5ztffeO1F95++5VykN6+dcwYw5RFzClRwVOYW0J49nvpw323TgmhGBOEcIwIYxJCDCHCluMhlR/i6AErPfzbQ5A4HGJjDKjUMsakVEopwRVnUmtrdOhaa433btd+QYn//e9/H4Be7/1rr732Uz/1U977L33pS7/yK7+SpjmjUggVA44RM8bOL87ef//9QZEw3n/59//4j3zizUQNcJDBMiXzSLy1fds18E27rttsNkD4Ojxa0ONCkEvTFHicIPvatm1ZFlW9iaDl5hyULkU5kIlCBGNKgBXuvDfWun2WgHeGcwLcKIQQ3Zv4ALx46GKdcwTCwNnZGSEEDA4BT9Fag1QT2ltX9X1/cnJycXEBefb6+hqiLNwbeKCBlxr201t4Jiilh51mqATAfQHUSgaDAZSwhBDQxIOpLkRomH1prefzOfyHwOKBXzQcDilBUghBSVNXeZ7+yKc/iVCoqk2a4+12eXx0dHR0JCXLM2W66p/8k//pv/4b/49mNf/TP/fH/rWf+alRkQtMg3VCqIjAxgqcAiOk1xgxQgQhghF9/h8UCQifxLA7cJBbDlfjUHQ+XydA0IWvhjBFmPoQrAvaOG0NItjH0BvtPGiJx4BixAETkuV53VaUk9FoBGLWAFFzzm/fvv3tb3/7y1/+coxxNpv1nc3zEiN6eXn9ta/9y3KkZBL+yB/68pe+8PlcDZCjfe+kTAghsAEHn+ow9oPFvUNCgDMHJpfwNbuuq6oKtAet1UWROW84h+X1AOPNAxMXOLjQsEP8AiOo53NLmqZFUeR5DoUinFEor+FaMfCSA5ejp0+fDodDOB9CCCk5sNHgUIJOE3zQl19+2Xv/7rvvgv8pnMXj4+Pz8/PRaATHHwYPm82mLMsd3Ggs0DiSJIHJvttLkyZJslgsoOoFBfHBYADpQFsDfRUhBIzatdZAaxeMOc8koyh6SrHR7ac+/fbZ2ZlxtjchBNe29Z3T41Wqzs7OcEQJFxdPPvq//xd/7ctf/tl/79/794eZ+hv/r/8mTRLnPSYM71siKIDcfm8L7VFAtIeffk99eSjK4c8HQOowbUJ73ue+gdjF3BhjjD1CiAtqzH5vLlCMIegihAihyLp+MBhgjE5PT1er1Ztvvvnhhx8eHR2dnZ2dnp5+9atf/cVf/MW/9d/9rTt37hzPTuY3648//GC9WRyfjPMc//zP/6FPvPailAmJgiBpzDbPUwAVECIAjId9UQdBgYAodu+Gw2HXaZUkTdOVaXL+7HIwGNzc3HDOb906MabinHqLKMPBeu9RCMF6b8E1ipKAUdu1h/QCCQSOVqf7NE0xoyEEF6JnPHrQ9HueeEARiiyEAHAP3Tv1QgqGA54kydXVFZjUcM7BphdEGGOM8H1gmAuVKIDz2+3We399ff3yyy+v1+ubmxuwRby6uLx9+3ZVVQB3j0aj9XrddR1w8GBv88APAlwdghN8Qzi4s9ns0FjUdR2Q587JLCc0VFVVFsOf+IkvffOb37qaV5zJul7H6PNUvfjg/tXlsutsVzWc0H/66//43e//4K/81f9sNpv93V/+5W9++1uUMW93uw3wS+FjwBl1zh/O2aGWOpzR/dXcncIdC8n7A8Z0KADggMYIiAqoT4QYY9QR4V3EhfaKEIKJxxgTgqFrEVIsl8tXXnnl6OiIcw4ZqW3b11577Z133nnttddubhZV1bz//vsk2CJXQvi33nrp3oPjPC3bRnNGOl1hYn3QKETGVK8xIoRSGryPGMGk0BhD+A78YoxhbGDkAUNjEIk33jutvfc4Rqmo9xYjBJCi8YYQclAzBrV4SJtt36o0g+QOqCrMSjjjIQRKCCRhuPgxxr7vpRT4n//K32ya5qWXXjo7O2uaZjQaxRhBsgagInDkhW7Ow2IeIVlWdF3X931VVbPZDGIMBDareyCVgvYQxlhKuVqtBoMBeW5/Dxzq4VvB38DzAK4jsLsDaQJGGuDPfO/ePeji4Yi4YL33nEsuk7btI6WSybY3g2H5d37l793c3Fhr1utlkmTD4XCzrm9uFl0Ps7UgFb1z5/Tzn//8n/yTP///+zu//Df/5n8dkCeMxui160MIFGMcEY6UROLQ7+p4Dnj+ods7NOZmT0aOMboAoPf+mMboXRtj0FpjSuF6QsSF4Eowwxh7SzjnhKCILEJBCE4Z/uIXP//iSw9/7dd+Qwk1GU4U5bBz65Djkv/Gv/yXr77yqWfPni2un8zGqWR+u13/tf/rX7mZX986uc08s9bDx6YUS5n0fU8pvb6+zjPOOJAu+Gg8revGWpsmg67TEQdKOTxRlHAp1WazoZRKKd/78IMsEU7XFCOOw7ZaAxB08CIEtgcUrwghyrkxpmkawtHJyQmKRGsjeCokCcEzxpRKKOX+ORI0hMh6WzGo82CxP4SwXq+Pjo4uLi7gCMe9qvehb4DMCwMAOIiQoKGOhmcCKP5vvPHGRx99hECpBmOwCgW6HWMMeFYA6fd9L6WcTqdnZ2eTyeSDDz6AXO+932w2k8nkYAZ8fX0NBJmDgyUsdQiVIoRGo1G1ruAJ/pmf+qm///f/PiI5UEa6riuHhTHG+T7GGKLX2n300SNC6L/8l785Gc/u3r17fX1pPZhrEISCj5EgRKKP0SO8K9fw3ucYHtpD/jqQ7p4HqiilMeIY8B51CiF4lTCVZJhIGDIf+FB4PzhlklMK0QUTwkE3ajKZPX1y9vbbb0surNZPPn7isT87O7v/4MHDW68Myul7773DCBoOsq5d/fwv/i/yPG2q+t7pPa2tDx6Ct7WW83Sz2Xjvy7K0VjORhhAwIZSz5XIZIyKEAaXGOA13nHMuRbJYLGHB4ezszHsXAsMxWqsjCUIISkmaJsC+gBfwkuCrRYzhzzIFClgYDAbeYUKilEJKSQhFiEDYgg4J+FM+Bta2bZqmNzc3ZVnCn4FOBkUhlFNpmh6k1+u6ZoxdXFy0bQvh7eTkBDSaQS40SxSkNuj94Xz3fT+fz8fj8Xw+h2Ic2p3NZlPX9cnJydnZ2WaziTE+e/bs9PQUKnfgUgF4HkJgjE2nU6312dnZdDqllFbVFiFUliPO+cmsWFUVnCHnXKrEj3/+s99/570rfCmlqqqaEFqOy0ho13XbbSNYghD54P2PrbWjz06UEiezk6v5VYwMY+pocKZFCEUcfIwI7WwMDhgFBHjo+g8l6aH3h9eOWQfmhShQgpKUJ5Ll5cAYGkIw2vkdxyoeagOCPUKIcdixRCe3jlarVQiIEMEZqeu63m7fevvtENDdew/u3HvwV//zv1YURcp9KlX05v/8V/7T+eUzSnmR5G1llEwJi5A3hRCbTSUEg+s/GAwYE33fEkIxot47zjml0Cnummj4YABUw1deLpeYhDbqhFNCSL3dFEX+3GoNgvLsUHcSQuIe3ICuOni03W6VzBmnXddprTkXew+QXSiEKIMjYgchPMAaAXe8deuWc269Xud5Diju06dPrbV37tw5Pj7WWkPwX6/XUkoQFoV5/XA41F0LuoGQvuP+BnDOF4sF53wwGGw2G/hFlFJoqu7evQsHUQgBbwvYL1wXKSWAA03TZFl27969pmkOFmRSSoTxTniDcuNC23cIh6LIbp8cPXr0yDhHCFmsNoPBYDwptBYx2qa2MZKm7rIs+6f/7Ctf/OKPe0Xnc+eCpwhLLhvnY4wB24A93vc0h/89tOr/817q8GM+gkywx0BLxu71V1/97Gd/VEr54cfX/+gffsUYRynDmDKGDwddMu+cm81meZ6HEJrt5qUXHv7213/rC1/40s1iIaUiTPpIBWdPzs6+/967J7dGPmhsKoLDL/3S/5EEPxpNlJBdbaTKqqoWnCEUoBllbCea13WdUilAXoAfee+FUBhRQogQStsWiivYceVcNE0DcVElFKHgrMPBU0pATdYZK7l0YVcLwSGDAGcBFtV6xEpCCCWUc5GoXCXMGCA2EOdC13UQ8iBFW2s5pwQeEQAdIc5tNpvr62sAUUGXHkLXCy+84Jx79uxZ0zTL5bJpGgjIcDKA0XxYAgYYFY4gRBqIr9BjwfEFBAAhBMBvXdda6/V6DVAfBKfxeFwUxeXl5dHR0WFBCpCIEIJSarvdQs4lnNVVVdc1QSFRImFMUDabzSgwMwgnhK3WW4Qc5/j0zu1ikME7ALb39a9/rWlXX/jijw6HKpEMR08xIQgzIikRzw+cnn/FvYM8bJXEPQ5wSHMYhVQpznAi6ac+8dZP/sQXUYzf+87vfPVffJ1SmueDPE+V5IwxLqjzOngTon3hwd27d26lSaKkJIha7V5/9U1vfSJ5CGFQTM+fLf7xV/7F06cXjx+/L4RJ0/hnfuGP/Ef/4Z9vNkuCYtf0dd0a5zmXlPIQHFwucBSCTw5asFCMcs6d80IoRgWlXKoUenx4/KCOOnQ2GOMYnBRcKSEVV0pt1xscke76rmuhooMm6bBZIDkHChhcJbiJcN78flUYIQQN/qFPhaedZVkGReTNzQ2cAHD0AqIa7A8Bge/AGgGPeFh4uLm5mc1mcD+EEFVVMYLrugbsF6yVIUzCYwH7SaDUXNc1FK8waM3zHMBOxhiQoaBRQwgdHx9fX18jhKD2ACILBNcsy9br9fToBFR+YkTGGOOdKjLbd87YBw8ePHr8hDMcMAkxLhYrpZRS9OTkKFHNer2x1sYYhGTvf/CRC/6zn/3ss2fPLi6ueIeEUKtN7XpEKXk+gv6ek0p2bgf4edQTIZSkkiDkg87zbFQWP/3TP/2db31bCOE9gltICMYkIoRQDAjFLEuyLDualErIvreEEEp4WY5ubm6+8IUvfuUrX/nEJz9pfPjGb31X9+boeNRUG8n8uGB/7t/804XMzs/PX3/5Fd0Zxljb9HfvP7iZz/N8IIWEglhrfSDj7pKpizwR1nrOZdyv77VNY60FgxhIu4yxum4gxTNGGUMhOO8dxVhwGhztug4mF3DLAI0J+xWXA1eTEs4YkyIBDqd1XdxP56GAPIjHQ99CEGJVVfV9//Dhw/F4DKMj4DNDyQJo/Hg8hvNKCAHe7nA4hi4esCeYrUNBPRkNMcYAf0opQTaHcw5/AIs6mFTBB4KOuCgKUKWCRwc4foDzA/vk7OwM/u1ms4Gi9nAsAAeglHIutHFt20aCvbEx+FzJ6J1k3FrPqIiC9J22pqVESEkHpUoz1rbtZlNpHahMn5xdnp2dffn3/+TDF259+OjxZlM5J7TWUA7+nvP3fKL/PSd198Mx+mgEI7pr/oP/9D/5zX/1tdWyquv2ww8/4mnGCfXBhhARQoSGJJW3b98WQrBI27rjnFsTpZRCEK0NGPr843/8zwJC0+l0uTq3PYmxe/Hu6Z/8439wLItOu9dff/PxR4/v3L47X1y8/PLLZ8+ejWdTRJFzDshAgEYfYAfI8pxRQLgQ2sG3m81GCuU8HEeGEAKMHXKX955gxCn22oWInNWCMixkjJEFaYOB0wn1K9RgLgQRAqVUKUUIPtBknXMhwG4gQWin9ASZlnNeFEVXWfydf/53qqpqmiZN0+VymWUZiDFBQQmh8fT09N133wUNWwjCbdvDlwF0F+pfQkie5zdXlwCywDD9tddeI4R88MEHL7zwwnK5hPIcKhKgn8L0BQ7oeDzebDaHlZTDDx/2VJRSaZoC5JQkibY9xlgIxYTKsqKuG+vCdrsdTSd6u/E+1L2+uJnf3CweP7vQDvVGM0Kt09ZqTIJzpixLpZSzsWm093yxvKIYE9T+5Jc+q0233dSrTbtYVjfL5nkcNIYfHlPy3C7586GUEEIZShMuJf9zf/bPXpxf/ZN//E9v3brz7W99F2OKBYk4eO8Ep4yxPFejySjLshgCctwZC92G905KNR6Pv/nNb6ZpWo6Pnp0/DVG/9upd13fj4fAnPv/5+7fuPvrgo1c++XrTNKPR5Pri8ujoCCHkYuBSEE6o8VALgYoRrARCrayUkpKH6Ky1nEmlUozp1eW1lIl2DQApzrmu1XXd5HneNHVVbxjRgiFkbPQheosxJZhWbYMJ88itVqt9fiCwx9sbA1Hg4UsPoAIMIVIine8RioQQxjjGNPxuJm6apiR4AgAV0J+ACQ+3vygK+DUYY5g3QDsVQoCYD0rh0PEA2FnX9W7FhFKor6fT6eXlJWBDUOYekNG4n17AmzPG1us1hHMoZOEQo+dmhoe/h8brgEQCHAYHBcID1BIxRhgQQKt3mI+juJPZjtFDgRGiz7IsSbLZ9HgymcD0NUvSvEiFYJNxiSNCIeKI4J/f88K/e8j0PGIaIv70pz/9qU/+iNb6pZdeefLoYyEU1Pco/HDyCYCd3fPMORdoNz2mSqmrq6tDlUYpLcuCUvTW26++8eYr9+/etdY9uP/SYrFCiEDNA9u5jDHtNHQIh9kYwD1w8RljMXpM4gGZhtsNvyXsh2d+v/IGOMNB5/CQPQ5ENjj0AAiTPdUV7Rc4D6UFvNB+aHQY2sH/jTvpmuic45QyuNne++FwCL8DmhIpJaUccNfttmZMXFxcpWnaNM3RUcIpCc5aa1Ml79+9Y4zx1kjOnNHee4DloX4FhigFL3VCYHcUng94c+DmrDdzLnDTbrz3va4hUo5Go+VyGQxCCBeDjHOOSUxSiUnUusvy/Oq6klImSdo0DaXceR8j9hE7G0iSUhsDDpPxoO3qJGUeB+0CogiFyCVzznEibdffbKtxOZQpCVHnimDMIx/dXNS6xMPheFbi5XL55ssvvv/++5RJwiiltDfGxRgwwiFiFBCKGEeEA8IYIexjwIhSKnDKrAuf//xP/9bXv/ve9z7OUmW8HU0TzGjhJSQE6FFCCDRwgilyhAnlnCNCmuCpFE/OnzXbajKZ1HVdDkTTiOlk8IXPfe5kNk7TPBDqkbWun44ngAdxzqEXxgGN0lJr3e5qRymlXK/XWVbA49q2rVC8bjohBONJ1TUR0W1VC5VAlrDGW+v6xkgpY/C2b9t6TVH0ODrnWECUUERElqrd8Hm76azprHbeMMZ8jIplxuNt1VnrmeA1KHcTSggCw2E4uD5Slaam6yPyjFBndCq4s6bVmjnnBoNBXddA1ICACuTlNM1DCLBrBkomVVUNBoOyLJtqCxHikDh2V5lS7/1gMID5E0zw4cgeRqPQuAHcZa2FkmU4GjAmkkSGEGAK4FwAGnlRDObzeTEYBo/artZaX16ewzoKBIy+7wflCAKzcwFq6BgJl7ujXJblIMut20oufAwhBMmFt9Z7B7E2yzKMkRDMGMcYQ5hSLi8ub1brzemdE+Pn2phyNGm6FkhihDEQANNaO2OdcyE6hAKmhGBGGPcuDAbjZ9ebv/gX/uP33n3/t7/2tbu3bn304btCCEQJ5SzG3ahda2uth09OKQ8BYeoxITHim6u5McbYNksV43h2NFwuz770xc997jM/kqfp6mbebBuD+lE55BkHjuLR0dGBMwr3ArpeAFKgK4V6bLlcIoS03s1yjbHOOdDULYrSmB46Kq0NpbSqqrqupRg43asiDcAGodj7ELzv+14762MA/XljDGW79TprbfDkgHjAb4ec+TwOhdGOJhKio5iE4BClzjnBGIPJOIRGODfwlQ7pGJhHw+FwsVjcu3fv8vIS3hS25uGbw7GGRj5JEgCxoE4AmtLO5Rch2ECFRABPDwzlr64vsizrOtT3vRCCUgQdmNgbgM/n80Rlg8EA2jgwtgOqVJ4P5stVUZQRYYRwjBEGB4wQhIJgdFQOlVJusaSUOu1QjBRRRANjzBnjvVuGdZEWTEjGWAwYEdoamw7K6+vrPkap8rbZYM4kSpquY5JJhrNUYYzThBPC9iN1H1A0xnGhFovVH/tj//o/+rWv2t7+1te+fu/OnSePPmAJRwY7i1SiqEIwFIVUC/fPwj6T887Hi2eXwbkQghA0TSjy9cuvvXp6++jzn/0xHPDyZi1ZloiiXq8Wi8V4PAZPVNioKcsSri0AJjCfBPm3PM+11jBM4pw3TQeSFpDfge9qrXbOGts3DbHWJjJJUrVeecBEm6ZhPIKMWQwxYGSt8TH4GCjnMlG96aHYybIMI2b3IUklAiIU29snwAElhOCwR/EiMFMD2otRkkMZBA8KJOXz83OoBhhjQG6C+LpcLqfTKfQxAFsCnXSz2SyXS8751dUVeEofSNpwZOEwAdsqhACJHmZ9jLHFYjEcDpVSMHk6sNfi3v+uLMuyLKE2OrTtoPcHX+NAW4HtNill37bRB4Iwp1xKmapEcskJFZQxTAKYGESCKWVCBI82VbWptrsml+DL6+u2N5/+7Ocurm8ub5bHJydS8ZdefqhNF4KjlHqjq/Xq4uzZkycfz+dzrTvvPUJRKVVVm7fffvOrX/3q/+p/+W/8xj//yq3bR027srFvumY8Hm829XZdO5CWhFqNsYhx2At4NE3z7OwMBUcxUoInnJVZ8tO/7wtf+LFPvfnqS5JRTtmD0/tKpfWm5pSdHB8zTqH7uby8hCsPJR0EApiJr9drzjmwziGIAnQdQqjrOgbcdrW1tu+7+fxmW605p13fOG8A8fHee28hGEcfOKUo7Ci/EHHglAOADcUofB28JynDdBqeExiy8P2MHhoPqDahAIXSWUrJYHdisVjArDLG2LYtTG7m86W1djKZMMYOEgwQt2EhJMa43W6h/oDuCnxqoPq8uLgAQSXI42mawm8xxkynU2strIIA5GSNr3VLKe9aHUJIk5wzb43XvU3TdLVatW2dpqmiou/7NAVZPJMXJZTnaZp6HwFwrdveGBOcK7Ic4rE2Dr55ry1BFGHMqUAsMM6NxXtOcXTGdl2HMK6rGmMslPzqV7/62c9+9oMP3ru6PDPGGdOPx0Nrdd+3WVakKSWEgLscxpgwyhjtOlMU5Xq9fXD/ha9/7V88ffruvTu3np2fIRSUUtaHvCi63iSJxBg7BzM9C7gepLm2bSklGBFCIsFuNp397Jd/8tWXHzT1tlC5aTuCxbObp23dASWyaVsuGNKaEHL37l0IcvBsx70lg/e+KAohxHa7Bbx5OBze3NxMJuOm7ghmCKNhOdamAwpzURSAfxV56axtuw4OU0CeRWS1TnhW7Wm7ILYaMY4YAemJCwopjlEJpGwAeQA3RAj9nixKGN1RngmLfnem4WlhgL1DD14UBVxr0C0aDAaj0ej999+/c+cOIErwbZVSfdtApQgbdtD0AfSTZdnNzQ0MnyAwQz3U9/2DBw+gSQTu/Q6ttTuhB0op8E2BeFrXNazjJUnGGBsMcs75ZrvarWLtKQWQJpiQq9XGGIMxiTGWZbm6mTfbKiLftDZEPByUit+0jAUcPbaCsYhxjJ5RgZDR1oYQOaVt1zHGikEmJPvwg3fu3r37m//qn/3Mz/zU2eMn221NkE+VmC9rjCkXinOZD8YFiTBqRwh5WL/xeDE/y9JBtbx+6eH9jz76AGPsEem0q9sNZSJi9PTsnOzXkvBeXhlG+VKoGB2KfjRMlRKf/9yn3njlpc1qwTlvaluW5eXllVLq6GS2Wq2ariaEqHRE9gs2IJ4DWQiORd/3UGdXVTWZTICzCycYY0Ipg82fJJXbSsPmo1LCe5pllDGOor9ZrDEB3QkPRlPOWBgvMca0NVQI76x2FrKulAIh1Hem67Z952DrCCrDQ+cOf4CQvNo0hBBOqFIKR9R1DUEIYiopiuLDDz884I4Y49VqBeZdMHgEwSZjzM3NDZBFYFsNhvVFUYBOLGNsNBptt9sD3GqM+fDDD0HSaLvdbrdb2ICDzxdCAHaI936xWPS9AUfOqmrW6+1wOE7TXGs7HI4BMQ0hAP4FKR7eB8S9pJQgAa4SCfEYrgUEe4rwqBwmSZJmCfwlwcy76L032llrfcScS8qZccGjqJ1tmooT/Im33rC6nU3Kxx++d3X++LWX77/5xiveG8l535ubxfLq+ubs4urqcr6YrxaL1XK5bmrtXeza/t/9d/931aZeztdN1a1XXdei5bxbriprYtu2MR5EGXYDKshdcP/6vheMFbkalulolIyHWZZI09t603tHzs+vpJQh+POrZy7avCySPGNy54gFvSznHPg9kN9hVxZQFHgekiRZr9eTyURrK6V0brdEpZQajUaU4r7vEQqHtinLEkIQEOYxxpwya22epJzsNpaqqlosFsAag4QORdpwOJRSEsKEUNDSQOKGPd4DwR6AMOAfQvcCFYINnkFCB97efD6/ubm5e/du0zTHx8fz+RLQf9Aw8h6aTQf5GrCM5XLpvT85OQF2KiwNQ70IdcbTp0+llCCLdZjSQpFaFMVmszmAXqDlxBg7Pj6+vLwEBR/oJLbbbQhOCNG0FTR2wNDLssw5FyMGJmLi075vMZVQcgjKtpWbHp9Y453px8PRelVtdYsQ8ihyygkhLlhBaSS4N4hSbnoN8azrunBjJRevvvaq1f3pyYwQbHV7fDS9ul6G1lqPXEQoIrJbPXIYY5XEEPEv/MKfUSpljL3w4MV3331XyXy13MgkNc5wgQlGhCD/u0ejcc8KjTEqKYtCZSm/dXta5CpJ5UcffJiqgbOo6wxCFFFCCRaBF0VRVc1oOA4xAtYLlxeyJ/wvXN66rgkhQgigL15cXBRFAUJAXae7rgWvI+895zsLVohKbdsShDFFxhjnd2SDvmma7cbmeYxxPB5vqi1i1HqHnJVFjjDIoDOuOOy0HYZtAJpCeQZzJsCwm84WRVFVVZYn3jpKcYQhPuOEMREjzrLCGKdUmmVFmuZCqBBQksjVakEpFoJxThkjTVNlWaKUoFzUbedCJIwPhqO214vVmjA+X67my8XNYn55fbXebrZ1NRyPytFwOB4hsuMHEELgYG02G1isyfPcOaOUmE7Ho1FZVRspOaU4SSTGMclSFzymZFtXo/FYSGm98zH4gHqDIpbWY+9jlmXL+U2aCMnienGZpsp4Q6UIEWPBIsFKUikwosiTgDlz2IMrD0IoOs8iZSEkkjNGrDeBoy662ujvfu+9bWWLvOSE1uvqwd0HWhtGZZIUmSwkk5FiLlLJUkE5wyjLkpOjW//9f/vfd6sFl8wiV/dbmVFMeyE8o1FQRhHOkjxNEqk4YTgSL3P1Y1/8XJarFx7eGw8TjFyWyHIwunvnRYYTa1DddNr2N/OnIXbIeymT6PFyvi7yIaUcM249bnvnI2175wIxDlGe+EiZSLWNiAgmUm39alO1uudK2uBVlu4h8DFjnDOFEUWRESzquo2OLm+2zbZbLZZttQ3BFWkmGeWcKiXycpCPymSQOxSEktZaHBE2Nuja9V0MAWMSPA2eOGulYIlilESEYggexofOBWNcjJgQVhZZcAaj0LcNQUFy5ozZrBcoegYrUdCnQxsFPJfVagUCfFAjTiaT4+PjwxYzVIrwzEFWAlBwMBi0bQ3/CtpwYHtsNhutdZHngEl57yEpQ/cNGOqBpg7RFxjaXddRLrquy/Od6UmWZdZpjLGSiQsBmjzY72uaCoh/MSLI7zvdHkogj+R5vq5751gIIQQcUQwhwDZxBHWrsCNMaGtijARhH9HFxcVmef3GG28cH99tul5yZcOONJZS5SlPaeJ7RwhZbxYvvfbwb/yN/ydnbDBIl6sFsBMhtkFZAh0ADjiEEHBodW8tdka/8tLLH7zznu76GJxgNE1kWZQk4NVqFV0QTFZVNT0eQgXV9/r4+Nh7z4Xqus703WQwNEbHvY0B9BwwKz6Mkbz3wM456F/0zU40He4j3AJCiDYOY7zZbKQSfd8naeGMiYIySq21nBAAv6WUne5BBcQYo7UpB0qoxEfqQ4R41Pf9eDxOU+WchSkrY6Dw/cMNMPi0jNAYEfRGwXnAOhlU0ISQk5MTGCZBnaf3dp+HelFrDZPrxWJRFCVoQw6HQzg3cMRv3boVgluv1yEEqDMgY8LoyO1ttTjnYLkEFCTQLwFOPnwe+A/zPO+67nq+ePnlly8unkHLdXNzA9O5pmlWmybP86Io6rruui7PBzHGyWTWtu1mXSmlKCbaeWfMuByuq60QYjIsa1Yv1lY7F5+jGDPGQnC7zEhJShPnnLcOemCC0G/8i98clqOXX3s1IiQkc96jaAlGzhvjEecS+SC4nIxn3/ve9+7eunX79vTJ0ytopQ8b5YfpX3QeUeJMQIgwJn7+53/ut3/rG4M8zfN8eVVTgiejcd+0m+VKcFakOUKoHOZSyqatJpOJdg4gPGOt1np2cqybinHGuYjR9bpGCGGCjG1jjAQLYLYjFJMk0dsOriR0IYwxAPZh4In3K51d3zKOrdWMk0TyrgkxOO8NRQKTnUZV3/e96bU2bdtaawlGXddhzCPBISBjd/GLUkwpZkTEGPu+B+yCcwkVIzpY2zhfVRslRJomnW2hKdptOoP85wcffAA8A/LctheQpcHwBchEEAag5Qe20WQyAQbxcrns+3Y2m+H9jgHolCyXy81mMxmN4dPUdQ2faT/1idBjwq879A2LxeL4+JgJCQU+sJ+01gD5okjSzK/Xy+n0aLlczmYTznkIUfduu2kGeY4xNs7leW69Qwj1fX/75JYQK8FZ27beaEyIBd3UEEJACCMlVUAIMDXwz4RD7CMijGsffvDOe5xzRGkIFmxhKWbOOMbSztq8HH71X/2rVKnlci54DCFQjOH5hw9wuKqRYIqJ85pSSjgb5OWTjz9+7dUXP3r/vUyp8XA4yHKEgpA8S9LJbLq8mdNcrVar2fTY7fWOMcZAXLTWxoisdUUxgLgFpcvef8LDae771ns/nU6VUqA277QD0BA4uADsA1B6fvaUcWK8TwTfbjdppoKzaZZ0dYsw8d4HFLXWLniMUUQB4ZgmKlEqz/O6M23bQ08sJUcoeO8pJlwIjDGlDO/X7eF2CyGUUmVZpqnq2xZynfNGCEFgYx1K5ul0ChEl7knjwNSESZff67ZBwKvruizLuq5vbm4AawXxj90y197bAWN8c3MjhIBAC4w+aJLquj46OhoMBqA7Ap/4oEkJiePq6gqadzi7QNGH4qGqN2maOm9Wq0VRZISwpml1b+Hc920Ho4HFYkEQ3mmfTEbO6NlknCcyTZRzjjHw2sIIIRdQ2/d931OMlRCMELxvODwmVKqAkYvBo+hDIJQSgmL0OOAQvEd9b1pjekYiRvqFh3e3dWWcA22C3VCEEBjVwHgBQJZEpn/pL/zFb3/rW/fu3bu+uByVQ0bCqMg5ZRwTZ5zT9ubySkgGtEZILyDABvUMJDqlVFEUBx5G2HONAQkGuh30PTBpgxO8j6wgfbC7s4QQq/vJdGSt9t5yQaVgkgsUvO21Mdp5QxnWpkM45HlaVRvnDKV4OBxyTsFfD4CUpq04p5ggxinGOylGYE9DKQVNPfxeoMADVgC1wc31goBCJ3T7MNeBWgS+CaA2UDxBRgC4FL4ewHhHR0egIwDiHzA4hegN7wPcVcA1EUIg8H6A6CCaQoUElatSCib48N+ivbwH/NhBVRpjrHVnjMFkp8Bz6Bb5Xu4/eo9CMMbkaQaEHQC5BoOBEIILduDpIIQC+qEC0aG/RgghsvvDXkzhh3wlhFDwmBDiow/RWNfHaLMkEYwBznKYo6Ad3WmXQBHBm2prjPnSl76Upvn11Xw8HHVNq6SkBCWSJ5KHgKRIEEI+WEqp9ztVCPjucKmhygRcme03N4A9BPg0FN/hOaebHejg3PNVx+FbQ6I/7EgqLnCElXZwrvcE7/AHeE8gjx/iUZ4PAPaGDwM/SQjBOB5gtbBX7oBzBU8RCMYAmgSnawf1Q30KxwLAhaurq8lk8vTp0wcP7o1GI9C1g4ObZdk3vvGNoiiEUMleYxsW36bTKUwpvN9pucNI8+TkZDKZrFYrgI1AeSHGeHZ2dufOHTiRIJMLFTDU6fC4A0TitfHeAx4RQphOp9fX15A+ttVmNjvu+94an6Z5kqQxIsBTYSJnel1kedt1s9msbpuuaU9mI4Sx1QMDKmVNq6OnlGpnXfAYBRJxxBiFSDHBFMcQIsWRRIQRCsF7RAjxDvIMxtHjQJjiLvSRWBvC8WgynY4ZJ9ZH0LhB+w3Pw30NIWCKB4MBJfxzn/uxX/3VXx0Ph9cXl8NBafv+/p07RZ4F56GnaZoqzaSLLskz0MwHiA3SCCjgFcNScAUPvPeeUbGTh+nrYTkG4+o8z2HnHpI7QqiqKrxnvAN2jtBu/90Yh2LDKU5V4rzerhdKSe8cZqQoMtPbg/5r13UEY4iClFLvIiHUamutjchzThmnJAZvLCGEMQEi6BhjYzTGGCS24dMyIQ+fJ0kS541SyY7iDvTBwzyDc/7iiy82TXVzc3N0dAQdEqzvvPHGG4SQm5sbcE6CiD0ejzHGcM4Q2ulFQcnFGIN9TiB/wEPDGINKAIB9QgiM8nfLqVIyxrbbbdd1EK3hOMIzehC1AwY+51xr2zbVdHrUdX1d9WU50lr3rYYxmhAioNhUFY5oPB6bvmGMccayrHhydulDJFpTFrWzPjqgCYcQfESEUYoJptQFjyjGmAQcEJj+WosxxSDdHSlh1FvHORGUTceDPJWrzVaKTGsHzfshciBEEPIxxrapQ0C/9Jf+8ld+/dfX621ZFourZ7ePpyhYLijnvGkaH2JVVVmWHB/PRAKqorTv+9FohCgFKsJwNBmPx9rZvt+Za3Euu64jhNV1gzG11gMAjhByzkDXC7QHhFCmst1yBSGEIOdd29UxYIRCdIhT5r0zps/zfL1eJUoghOrtRghFCPIxOGe6vtFdX4yGRZH1uhU0xRjkauNhsYzuZXu11nbvYae1tdbCsYaHbbe+61xd16lKfLDloCCtbtfVmiue5+licaN1Nx6Wi5vrxc31eDy+e/fubDYTQhwfH2OMz8/PlVKXl5fT8WiQZ7PJuCzyVMmbq8tnT5+kSnpr4HhBxgcOVJ7n5+fnMcYnT55cXl7CfipsMgFacSj2AfKAaD+dTsuyzPNccGytxpQnWWkcqRrtbcCRJDK9d/u03W6wd8NBul7OEfJNu+x1FZExwdd9hxjdtk1AyIVgrZVSam2USmyvp9PpgxfujEeDvMhUwiXHOAYcSYzYeNz66DHxEXtreEQSCYaYIEowiQImGAdvQ3QIB4N1JEHwkKcMex9MiI5ulhtjW0pxCJ4Sw6jFMdBAiMccCeYwcv7P/sKf+vjDdy6ffTQa8CeP3ptMC65wUqp11fbOqiKt2i3hWCSSMFFtdaJKlReR0qwse23Xm2pQjjiXTdMFG3qjKWcBRRe8UBKU5SJGlDPvkNE+BBQjzbKCEh4DdjYMy7H1HuNYlsX85hIjX283glHTbzFyGHkfjNUNp9gZg0IkkZLIURTOGecCw0qyLGFZlqTjIk0olZyhSIz1yAeru2hNtA52roQQzgXvgwsEMW4Ctj5qrQmKnBJKUdfVXV8zTpWSkGYxonVd7Vznx+MxFDeAhiqlXnjhBWDKPXnyZDAYQGwD15ijoyOM8Xa7BcVumL9DhX4gKAAMAQQ8tCdMDYfDsiwhv4B+DgTCAxUVvPHg30LpCR0VjDRggHR0dAQ8FXjmYHyFEKIUE4pAsRuaLShiYJkVlAT7vod9QJkmzpksy8qyHI+HeZYoJVOlhPgh+Q2U5eDTEkIwps9VVPgAanLOJeNgVQPiXrA4RilFmFuDmtY3nbXWBuSdM9tq0dm6LIvHjx89+uiD09NT7z1CYTQeAq9+UBYwDSnLIsuSwSB3zgyHA4wjyGqcn59fXl7evn07xvj48WPIV1DwQShqmgbkKSGJd32rTQ+qCqvVSgjlXIDJnzO9994Y7b1fLG4wxnW1yfOdNStCQWvdtg1QbQ6loBAieNS2vXOhHBbANG2ahvPdbA/KGLgFwLiAyg0Q7hAClHkY47Zt1+s1kJI554dBOhwbbToGCp2r1apIM6AXzWYzuMRQDcCwAVItjMVh4R0W2OEIAtsD7mvbd8BKBPlZaDbLsnzy5MlkNLbWAmfv/Pwc1EQOPQoMZOFaAG0UysTZbGas5xFDFlssFq7fKfACPuWcY5xDj8UF3WxWdd1OxkewF0p3rj86z3NY/XPOSsG2VZ1lRZJKoTiiSGvtrUMIxegIRSRS51wkSHJKMMKYYvTDXofsXyEEY3opEmcZInaQ5RjjutnGGF2wMTJMCSLRh1jbTmCGEMI0ZFly/8EdxjHFFGMsZBKjjzHWzRZjnBeJD9Z5NBqXUsrRuHTOUUI556enp9vt1hhzenoKOXowGCilnj17Brj9oU9HCIG40NXVFcW7dd6ua0IIo9HIe0sI0VrDuvSgyDHGBOMsVfMbQ3HM0wRFjxAWkvW9reqNECLi3cqHNhYeTufcfD7H2CtVUMLj3o0EEMnn1jlACYJba4WQ3vvxeLxZrYPzQgpKqY++bVtKSFEUigvGmO56gKIIkEIO1WFRFOv1GlbagQRAKX327NnhXINoNzSwMLMCMFbvBfcP7FdQ3zPGrNfrGCNIOIGOM2PszTffLMtyu90CvjCbzaChQwhBPUophd0dUNaFvg+mIyBRBg8MxFGAcowxzoa6brXW1mlje8qwD3ZbrXvd1s2WC9r0rQ3ee6+kiNFzyhiluUpOJrPJqBzkeZGknDL6nIoVeA+H/QvO6AFJ4VRwKqLzkqu+7WyvtdYg6MoUDrh3sY3YcSmoVJHQ2fHpnQcvDgZ5VW2E4M4ZrTuov2FTCnSr8zwvyxKuIZBjQKGTEALJqus6qNefPXsGRR5AIoAZw+mBCgqKPK27LMs4p3W9xRhbp8EKJ3jX9z3FJE1V27bBmabZQYqAwQFmFWPUpmvb9ubmRmtLCLFWe2/SVAE+IIRQKoVYcwCz8V5jFTKq27lS7bdD95roME2E+hh4xhCwAEImkPsggNd1DRwluBDwPV955RUIpZAmYCfEWgv5FCEESouXl5fAR4bqE94BEM22bbfbLVBDQLUGRkdHR0dJksxms0ePHl1fX19eXoL6COgxwW2DPYHpdAoJGri3sEGbJIkQCnpbEF/dbrfWasbIoUWllF5cXECdEGMkBEnJnXO3bt3yzqZS+L7P02w8Ko+nk/GgyNMkEZKCaxra2bkgusMyD3DMbv6BSXRWMYoRGgwGbdMwiiUXLqIkS7kkKudQWDPGxuPpZHp0+/Ztay2lPEkyrU2SZLDsBqB6nucxYmfDdHLU1B1GlBLOmYSw7d1uukEpzfMcWHNQO8Ed6fv++PgYSPVCiMePH0O8dM4Iwby3WZI4b4Rkfd9JxmLwRZFdX19ShgVl1WbFOU8TuV7deG/7tjVGIxw4584ZjHExyGAwvtlsN5tN3/dA8AP3IsCwodLbw44GwhZgl5xzWLmBEAvNBhwtSghEQPjkbK/pSaAdgb8CtByUTAghUBwAy1UpVVUVVKWw3QF/BmLeer1eLBZHR0cQ4eDhgGMBtUiWZXmew4QJPJCePHkym82ePn36/vvvQz4aj8f37t27e/cuxvj6+pruzNDZgXc4m81gpR+YAHFnobdzlwM7EYgxWusYgzHaOVsUuRDcOYsx8n4Ho8I+Z4xRmz5JkjxRktHZeHR8ND2eTTkjJAb2nMwnwLEHnBjv+cUxRkoIEPU5wbrtYJoAUeTy2aKr+u26auvOaEsJOr09M3oTffvn//f/h7fe/HSiCmcDZxLk8WBbJk3TJMmqqhmNJicntxEibdsPBsOu048ePYJirmmayWRydHQE4pUQWYBbdHl5ORgMzs/PQZoAogY85Iyxqq3oXqcTdh/augnOUxRX60Uile66pmmKoowx2rAb3yulhsNhmioINwSzoigGgwLhAEg+Jdy5EPyORwE7bZCZgcZ/SEGm14wxwH8gugNcRfeq83B2IR9qrZkQ4vz8vO/7PElhXN53/Xw+f+GFF6qmHg6HMIyBtQ14FAAagGMBCRrSOmwgaWsopTCdgpwFEFXbtsu2A4IINBnb7XY0GpVlGUI4Pj6GnAUtmlIKagmE0PX1tTbOuJsQYt/31uqT6QR+WIhiMBhUVTOdzR49esykYFTUdX1Yva+qCjb9AaKSUs6XNwihQV5YqwkhFJNIIopBSQnTnVSqxXLjXAgoeO8DwsbvTL3QXjr0MBfAhHJC+64bjUaSC4SWTd2VkymltLc2VSp6L7lSSo2G00+8+dY7735XCDebDtN8/MKDV5Bn773/DqC/AKpLyQmmIcS+14TQ7baaTCZd1zvnz88vjo+PoSMWQsG4EtLRer0+PT0FWWBgx5L9ug/nfDwaQrjCJGqtx+ORMbqua6WErhrvPd/5LiKgHkRj+76XKScU2+idRzgGxpjpLcYYIaLkbuoRUYzId12HxzjL8qbttDYIobIsKcN1XYPWJOdUa42pCCEA67KqKm8dAF5SSuPMdrudTacxxhBc3/fBecAjydnZGeccGqOnT5/CiX799dcxxkDuzLLswYMHkF6zLBsMBovFAiIzdCHwn7/55pvT6XQ4HIL34aGDBp4y5/ztt98+OTmBiRw8NPw5+wiMMQRawFMxxmVZwgh0NpvNZrPlcgmLiIwJkFGFXRRjHJStsKgkpRyPx8PhMMYIjykhJM/z6XQKkbgoMkqxC75tW6WEZDyVihM2GpQIIcF5kiRAcIaHMIQQ/E65zu8F5+NeAyeE4H2c3ywfPHiY53mW5sbZ4Whw9/YpRShNUZ4zTmN09pUXX7q6uBwPJ6Niqlv/l/6jX1Iy/+Y3v/vwhZc26wpQASEAisYEM6NDDJRRhSKTImsbMxrOABI+7MQdmGIgewHaLYCmQRSHpmK1WrVtG5FXSrG9HD3nlD83OIRvtFqthBAIheFw4LwlMQjKDuVvtn+BOghlmFI8Ho8PHS3ZS0gD7/gA4Pi94TY8h4BqQ8g7zOSyNIUpA/T+AN1jjEl0HvngtGnb9v79+4SQ0WS82qxXm/V2s8gzOb86r6tN9CE4zwgniL708GVQYYAkCMT49Xq9XC5jjNV2PZ2MxqNykBeCycnoqCzGN1fLpuoB0lqtVuAZAllpMpm0bbu6mUshzp89A0y0bhsYBspETWezptOD4Wh8dBQIOT655TGjNGU88YGt1rV1yDlEKRdU5GmaJUlb14Iyb2yZF6brOaGpVImQDBNnYvSEEQ6pE1GCGR2MShvcYJg1XTUcDhhjMVCGeXTBxxiQh1JvVwg+P8Fj1EfsPKs2TsnRcHpct31XN8ej8os/8vpbr97/kU+8ORyUr7782qc/9SmCPMdBdz3DIlH4r//1//JLP/n5b3zzt27fPVF5Nh6PHzx4mKpkOh1fXl8Mhpn1JitSwrD1zgUvE5UkRZIUWvuu05tNpfVuwtf3vbeOYtI1bZGl42GZKqkkTRPuXR+cpjg4rUlwBEevW91sx0VKomvbthhkkcQQXdVUVbus6gUXdL25STjCKDirUYgo4ETmnEvBUxSJcw5hn6YqRk8JL/IRpbJtW61763SWJyE66G8oZZQysGfu20b3HRc4ONtUW281owgF60xHIxKUjQal08b0utpsgQTijCVlWR4dHfV9DzHmQLeDOm8+X57cPtV7q7gQ3GKxqKsN9JgQnLIsm0wmEL2llE3TLBaLxWLxve99zzm3XM4h4/zgBz+AVokx9vHHH0MjBevOx8fH0+k0xgh2H9ABwFqzlPL73/8+1HxnZ2cXFxewncgYw4hCdQWcCagpvfdN0xwdHY3HYxgTQwBACB1kUWBOAwjDoYyGKH737v3xeJxlBUy9MeNKKZnsGJZ78tjOegZG2CpJEWEffPRxr+3s+AQTPl9uzy/nVdUdn9wtBtP/7b/z5//jv/iX/97f/4dMCkSwUJJLFpEfDPJ/+I9+9f6De9fX13fv3h2NRmSvQA02hKDYClZpoPcLuN4eW6BhZ0Vewzol9EnAZoJ5ptYa+muMcTksnHPG9JRSKUQIYTVfVPXGe88Jdc46Y1GIPjjnbJqmjFDJBfxj7S4iQmLM87xpmvl8DoKMYCIDWMHB5BJoG0BDqaoK7nuaptH5JEnKsgzBQdEZY7TaOGNXiyVBGPItrEhIKQmQKaHa01pvt9ubm5v5fD6bzbhIR+Ppar1VMh0U2aDIgrepoE73EOcAQZhMJmdnZ5RSkG5M01TrzjmTF2nXNUWRMYbW6+VsNhkOh8DkeOGFF+DowHIcSN5576FJr6pKaw0SP5vN5rB8NxqNgEkNPfvl5WX/nJeX2evcQpzGGMP0H8xuYAwIjzVQTuGWTyYTACsGg4EQilPuHDo5OhaUU0o5oRjR6MOBV3Go7Q7HNGBEpeyN+d47790sN7fuPugcSvIh5fnHT6+TdPzKq5/85rd/MB4d+UBdIJgmiCjAOqpq8/HHHw0G+QcfvAdst9PTu5vNZjabgRMrPBKQ0K+vr+EbgUo1FMTAhoYtWWCpwVL4oZ+DH+Ccw3VTShGMQwgk7kgnpmsj8s65utlGFARlhBCVCCC/wYQPUJ226cHdBXwHuq6D5RD4qMCdAyEZ/pxDC+zcjkYjCH8Qaw6rzFAAHDhN0BODVHdd133fEq013ouua61BqLsoihgjYUIkijGWZDughCKsEgGseyBHQUfS931ZloPBoCgK53Y7KDFGLmhEPkllnqfG9EC6gWYFfhj+F9LlDsjc9+OHE3Z1dQUMiaZpYGUR75WOocqBJxViBhD+0d5q7NBrH8BLqIpASAJWzOjeSdtqjyIhCCuVxhgZoYQQjIAdt1NAPnCC4LCinWWvw5RHjC+ub9rOOI+W22a5qTlLfu3X/4n38dd+7dfqrg0hRIR3EJXf+Q6u1+urqyvGydOnT0EOCGZjQKKDUhLtWbkHrTX0nFguFG0QKYHYAHO7sFdNgsbO7gR+KTwJcPVI3EHX8ObOuRi9EOx5It8hXcA1hF83mUxu374NssOHcwZI5+GJInueIfTmB7wT+nfo2yDcQu8P25fAgwGZsBgjOTCSEEJ3794FznwI4cmTJ5gSIZNGGwjjwPvSXc8ZkVLCkYfNjdFoBD6fRVHcvn0bep2iyGL0QtDxuExSPpmO4CGGUvXs7Oz6+jrLMqXUfD4/HNy43/ACzJVSCgioSBRY2MASHzB64FY1TZMkyfHx8YFZUtd1CFFKRSnT2sBybQiREDoeTyaTqXPeWleWQ84F/H3fa4wo5xIEsIeDAReUUUwIUULC6HJ3yZ7zTUMIEUYQwS54TJl14dnVfL6uLq5XN6vqO9/9/r/2B/7I/+U/+6ubzUZK7pHlghCGKScx4tVqs16vEQptW2OMnz17Cg/hw4cP/d7oDFpMuBRw29h+UznuHZEBfYTtVoTQaDQCqBKyEBxW0AjabDZCCB+C1roz2lqbJBKcAZ3VSSLhUcQIERThbYFEBhg5sMXp3oNgs1n1fQ/TbyjYTk5OxuPxAYaDE7ljOXlPCIElC845DDxhtgltH9mpkNZAZDs/PyeEYBwJwD1QwcCKJmyxxRgXi8W3v/1t02ugNjZNc/v2bZCmllLC4/LgwQNASb33wK7vOv3CCy+cnp6C9p21zhjLGNtsVkA3Pj4+XiwWL7/8cgjh6urq8vISwi18OPBQ5JyPx2Pv/Z07d+CdU6mklN77y8tLGKAdAj9MEKC1d871nTHasb0VGMBYB8wfIQQLhMPhcL1eQ0wyxty6dWtcjsGjLZG8yFMSI2MUodjtzSQPwTjs5Eg5xtg572MIMYYYjbFMcEKlkJnz+JOfemu1uuj6rZS0aVdC0q6vfNAIhaIYfPzxx4AxG2PquiKEbdaVNZ5SChqoIBQMU7Q0TUEUw+03VcCY75CsIGkCRxveEyx7ILgwxrpW71ZulIKAB4FTSZ4lyfzqerteplxaY2L03kcoQqB/B1hgOBzCug7gJ0VRAhoIR3+xWFxeXkLdf8jUwM+HxHgYc+zEHTgHWB4iFADA8H0ButnNkyCGHQB9yJVCCOh7xvfvee+7pkXRbzabqqryJH3y5MlodjSZTKqq2mw2IYTz8/PZbAb+4NttbU3sWqu1T9MUoe4gz6RkDiYyfd9fX18/ffr09ddfBwx5s1iCOCjjvCxLLkXf90+ePKmqajgcbOum73sfUV3Xpy/dMsYETNbrNbBDsiyDYAwlB5A8nAsY09FoQilnzMeIGBNKpatNlSY5ZzJLi+BRXcHWDluvtpnKqqoKAZVlcf/B6XI7b/pORkpoOKT1Q3UB+R1jjALCiIRIUMDOOaloxIhJ5kP38aP3wx2TpZwijyjSXcUpzdOCEv6DH7zXtq1KZYiu122eZyiGxWIxyIt33lnDsKMsR9C6wRIi3GwA7waDAUA5h3uXDdLnce9DTUIIcdYyxoztjY6EIO8cEPAwxk29ZYzFGLy3XdOWg0yq3Fsj8gz61P0+o9xsNhjtNDiAf2NMD6Pptm2tdYyxYpBrvZthMsZAPB6GJvA3MUYfwmFQxBklMSCEKGaAz0NvpPaWB31vmHMOSKOgpQ0VNyzLN/W6yE8QYhfnZwSzPM/n86u33/7k3Qcv1E0DVcFByr6qqhdeeOHDDz8MgXSdRihAYXrr1i2EECyIQlmDMb5z58719fXDhw/D3oMG5oEnJyfzxUJK2fZdCOH27dtQSmdZJpN0va0QQo8fP2aU3Do6NsbMjiZQuIDWLhgz972ZTCZybz26e67yHAIPZJnRaASbKlBa1HV969atel0VRZGmqYuhLNLJuMyMqnu9XC4pV3DXD8Uf4KBCCO8QQgjvtTOFEBzTGH2SJEEHRlDwIYQgFaeEzMazcjD+2m/+1sX5FaVM6248KbMss1ZzzBeL1XK+unU8Ojk5gdxtrYXh54HlDg7HoAQD6RJWGKDWBH4QoAGj8QhYXdG53XqnD0KIm/Ua5ibW2rapKEbG9FmWSinSNOWUjEYlTLPpXmIDliDyLAd8QGvtvFVKAe8pSZIQOihFQNAUSgXoFrz31hq4YtZaH2JvHCEkgdI5hhCC2Rt/QYKF7TRKKYqMbLfbk5MTaAAnkwksb4CUA2Nsu12vFnO8Mx3DDx48XK/XIWJo3KD+hV98//79x48fH3YJKOUPHjx4+vTpd7/73WfPnoGW4mq1gmMKy5yMMTAwBUjVOQeHBl5QUcEHg4JsNBrduXMnz/PRaAwGuMYYyIMAfEDZBFArVCwgiAL904ErLaUcDoewghJjhIEWY2xQlpTSiPxiNc/L/PXXXvnJn/jiq6+8VA7yA10LXgAUcM6zLJOcMYIZRoyRJEkYI1JySAuDwTB4lMqUMRZcFFRILr777e80dQ1FEZQfjJHwnJYgSGIB+n1o4AaDAcAUUPqv12u0d88Gkh4cSijGNpsNDK6hXvfeg9WE934+n0OgAgqO1j0MI+pqU5al5AJOVZYnB6Kj2KsIHkrSPfxuwaaa7TVtoPuGQgtSZXhOBRfQgMMiBtDW4JMfiCOMsbNnT0J02nRppsqyZMa2Vb0aTwY+6Hq9Bi1wlagkSc6fXfNAiqKMhC8WC8oZs733fr1dPXr0yFr7hS984fzi4uHDh5iS+XIhExVQVJQGZHvdVw3PB8VwPNpsNm2n8zzvWg0jBzhbbM9QppSu1mspZYgRDBGzJN2uN5xziklTtX3Tee2ElNV8HTFGgbqCRk9Gg1FXmMViIQRLU8U5ZYwxKggh62o9mUw29YYwFr2PmDLOAiLBU8wEQv709O7FxUVd12mqgIkHEGzb94lUXdedzI5QiHcmU/zwxScXN65vCInWOy4opQjqUY4ZToDj7GOMKhWUpoiAUQFNlMJUNKZ3zjFCX3jx5b/3d385SZKizAaiKAb3GA0388umWUnGy7IgiHsXIhURcy6z1Wo1nRaMieiD7npGKLBhEqmaqoaT5IxlhOZ5/uTJEwioeZJbrRkh0bmg7fFkcnX1TEqVSNWYuq87zvnN5ZU3NlCSZqpqV1JRQkTX1kxMEMEhEIx4kjgBG1E+Gu1QJMbYJEmlDFfzi66ph4NcCU4IcaYnBFGKr64vRqPSO40xKooCEeK6pjMtIpEJFqLp2t55bPeKzOvVZpANYnD5QDJCMUbW6TxNTNMwhnWzTYuUwRAc6s66aoWQjIqm7jbrCuIKFDqAXAJl+OTk5OTkxBiz3W4fPHgADb7WGiaTkHCBywxmX3DhmqYZDAbAXQAjOedc3/eQSth+4QscvAFsh7ap67okybS2UqnRqPztb33n7t37bduDSlvbtk27PT29BY1h29Z5Plgv63VVgfTundN7FxcXEFyNMUIo2DIrigyUNUNwEGzCXtYa1Hsgh0xGY2MMS1OHrHMOjH6N7fvOU0qLIo8xpnkWiW+artchehp3HGemTWi7KlEqIhoI+2//3//dg/v3syyLMXS6vbrats0aoaAS+fDhi6e371+cXc39HET4gWIHcVF3/YHGBvxGiKagHXSIQDt8ipPhcIAxdh5FtCNMAui9Wa27prl9+yQEY50OgW02myyTzrnBIC8H+YEaNxqNthXWWoOP434plDpnofEQoEeHkRDCOgtaDExIjMFL1zjnVJrGGAOKpu8ppdPhCGGCfIS2iXOeDpTptA8+Gk+E8MF671MlA6Wc75xj2dXlzWg04pznWY4iadv2+vp6NBolSQJ99Hq9BkwRMLCyLMfj8Wq1AsIbIBewu+ycOz4+7vseaHWAfvd9D8PMpmkoJlDkQQEHBrLQTUPHDegVzPohpY5GI4QDiiHLkjRTDx8+/N7332ma5vHjR0IoyTHnVAg5v1mPJ0NrPWOiqTsA3larFTC0wSsMXkri57dMB4NB1zWAJ0Nt7pwDPzvIZU3T5Gm27arXX3kZvsVmsxGigGsym83KcrRtaqE4wezianl9vUWYIEowjs77T/3oj/3t/+H/wxhrmyrG+OjpWdvWQogiU1kuhVDD4WA2m92/90KWFSQyzrl1/WazGZT50eykqqrgXL2twFoNvgtgRlDqQd0JqdYYIwTrugYgHiWF1mG1WhVFQimxBpTbsLWaUIRxqJt137fDYZ7kGdQtAAwBgEMIlSKBVszY3joN9RuhKE3TVMkiS3AMSZKcX16plFZVJVSy2WycC1KlhFDo6KFEUUrp3hLGCGGMCR8spaTrGqcdlJuYcIJ2SDOIo0Tke2PIyclJCAE21CBAHh8fl2UJmxhw/sqyhF7y6OgI5m/QkfV9DwXiaDR6+PBhURRPnjwBzht0kdPpFOIBcBYPpC+4oNPpFGY/g8EAfjUsAEDogmHBkydPACsGCJpS+vIrLzJGvLcY+4AQYUJlGeF8ud4KlVZNhyiZHB3Be06n08ViASsfCCFAxZfLZQhhvV4fHx+DlCYQzwDGB76VtXY0GsEUqigHSsg7p7coQffv3fn85z7z4P7du3dun94+mU3HqUpOT04no+mt49tlXu5mpIggRIz1f+gP/5Gq6QkXUqWYMLsXrm2aZr3aHs1Ojo9vDQZDKaVzYTQaAV/JeRNjDNERii4uniklOKfWaim5UsJ7KwSjFHtvsyzx3lKKY4yw2iv23qqQfAhBkAOhx8/zfLmaN03tvOm6riyLw/iNYNG2fV23nHOEA6XUGAukKphpYxxDcErtJB6WyyXwgIGEPp1OR6PRvXv3wIUChn9xL4eIEMKYOhe0tk3TQTZwzqWpEkIkiTyMGACHdc4pmZbliBHCtLZCKCkTjKvpdATtPEIEbirar/eDQxzgGsBEHI1G8/kcvjb4F04mE4Ak4R4/e/aMEAJbIkmSOMpgdgzxGEZ2cNBPT08PSzMhhMlkst1uQwgvvfTShx+9P5lMFoubuCTDcpymKcZLKMNhsRMeU+cNIRRkeNu2KYoC3lDstJ4xUNFGwxHwXEGnwDl3fn4GcAGUAcDhAiQPepG+7aazSbXdzKYTIUSSKCmOYvBCCMEFwSxVaaMbY3ulhBCqs4GggBGOkVAiuVQYERsi59w6LRSnDEsqCSE3NzdJKp1znPDgY5akMN3O0hSYrzFGYDsA2AzBMs9zgKW99y+88AKAMAgFKWWv2zxJjdExBoRJXW9ns9lqtWgbobX2weLgQ/Bt2xGCuCBQa+EYR6OxM+b09DQEF7wbDofO7uxSEYpV3VBK03QnoCellIJTHKN3wHrbwZ9CwhSqqlutd2ZzGGPOFfADbxbbrjdad8bE+eJaCe60wSRikk7GQ2Nong+qzdZ0fZZlUiabumIQaQBmKsvSOffo0SOl1Gg08nsN+WfPnkkpgSsACmEQcgCK13v7Q8h6SqmbmxuIQHA4YDJUVRVBGKI31FiALcNQHma7oOqY5zmUGYyxH/zgB9PZ8Nmzp0VRUsaXq3meZwgHY/sYY5YPhJKYMmPtaDQlBHW6995yzoNDJycndV0fzU5ubm6EUMYYWI+Gof/x8TFYBUAvD30SFANwxUG0H6BvQXgmU0hGzaa+ffv2ZrEGCe0iG26rJhiXDwcrWlNMBCVEcOOd0e5/+gf/YDgote5sr4miBEXOqFISOZ3lCefi2dml9/78ycV0OuWUCcnrZptnEubgIQRBSdfvVmKcc9ZhTGLTVrC33bQVF9R5M5mOlsslY7RtG0DudN8WRXFYpAGUQJsOoRCRjzHE6Hvd+eBSmWw2m1vHJ5vNWkjWtlUIqywtk0RBdgUACPD/A+dwNBrprp3P55QL62Pcy4Jq3UmVxojavj9UbiEEpRLBjTa+abosk0qpssj5Tt8YW6sRwiG4wWCQTGeLxSoEhBFl5+eXUiZN0zx9+gz4z2+88db5+TlCxDmNMb6+vr5169bB6BMhBHW3c+7p06cwTk3T9Pr6Gvxluq578ODBcrkEEUrok2BykKpksVjAmhsUiJDREELj8biua6CDgOci4DjHx8fWda+88opzwVgvhJLGvfjii48fP+77frOtVqsNZ1KpxGhQnmZCUs55ZzTcmP2mKMEYF0XRNlprPc5HAIjAvz3QneB7IYSgt4P65N69e0+fPi3yAdR2RTaw2o1Ho6IoulYXabZeb2OMfK9+jzHFCDnnXn7ppW984xuCMhNjUWS6rwUngtMYXJrKPM/HwwmAGJLx7XYbaYgoDIfD+/fvA3LZ9z2JCORX4Tk/Pj6+uLgAbAiSz2w2g2QCrSqJKETXtd1kNMY4bjarwWCIUKjr+XJxk0jatlVeKIQpjjjPU6VSHJAUyWZdRRSrbTMosr5vm7YCnCjNlDVeSkkJb9t2sVjOF/Pj2XS73RIUT05O6rZLKWeMbesmTZPl8pxQDiC/DDs0Osa4WVfWOq016PeG4KzTSmSQgb11ZVlqbbu+98KORqPVcl0MSwKDzclkAlEEyj4oCIAxBMseh/IZYiRM/wBchXMJ/RBUmXCbj46O4JzBCigUgiBYt1qtXnzxxYMc7gHA++ijj66vr0HCCkgJQAV85513mqYBwE9rfXp6CmSCu3fvSinruqmqqqqquPd8go4HwvzNzY0xBkoRjPHByAGehKOjI845DGYAeoQxd1nuqkloSoBkDQxwgGxv37qTJru1E3Bf9t5HgqH+ttYyTPq+vTh7ulzNKcZGd5xzIdlgkOd5CkPXtq1hkQb6QkIIo/zo6AgiHzRDIJh1UKCFH1N79S+47LCPNZlMYAyDMR6NRvCXEAtgIVZKeXNzA2bBo9Ho6HgKRL7RaFQ3W8YYIRSIS0op0IXNsgxAg+Vytdls+167/Yow5xzUCGE0CAUhJCiguYDKEkyYhRCr1Wq5XMK6OXzy2Ww2Go0GgxwkCsG+VoBbvbYIofl8/v8HbDk3Y7+2uZcAAAAASUVORK5CYII=\n",
            "text/plain": "<PIL.Image.Image image mode=RGB size=224x224 at 0x7F723009FF28>"
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "with torch.no_grad():\n    logits = model(x)[0]\n\ntorch.argmax(logits)  # should be 243 - bull mastiff",
      "execution_count": 9,
      "outputs": [
        {
          "data": {
            "text/plain": "tensor(243, device='cuda:0')"
          },
          "execution_count": 9,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "%%timeit -n 10\nwith torch.no_grad():\n    model(x)",
      "execution_count": 10,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "88 ms ± 6.37 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "with torch.no_grad():\n    torch.onnx.export(\n        model, x, 'resnet50.onnx', verbose=True,\n        input_names=['input_0'],\n        output_names=['logits_0'])",
      "execution_count": 11,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "graph(%input_0 : Float(32, 3, 224, 224)\n      %1 : Float(64, 3, 7, 7)\n      %2 : Float(64)\n      %3 : Float(64)\n      %4 : Float(64)\n      %5 : Float(64)\n      %6 : Long()\n      %7 : Float(64, 64, 1, 1)\n      %8 : Float(64)\n      %9 : Float(64)\n      %10 : Float(64)\n      %11 : Float(64)\n      %12 : Long()\n      %13 : Float(64, 64, 3, 3)\n      %14 : Float(64)\n      %15 : Float(64)\n      %16 : Float(64)\n      %17 : Float(64)\n      %18 : Long()\n      %19 : Float(256, 64, 1, 1)\n      + : Float(256)\n      %21 : Float(256)\n      %22 : Float(256)\n      %23 : Float(256)\n      %24 : Long()\n      %25 : Float(256, 64, 1, 1)\n      %26 : Float(256)\n      %27 : Float(256)\n      %28 : Float(256)\n      %29 : Float(256)\n      %30 : Long()\n      %31 : Float(64, 256, 1, 1)\n      %32 : Float(64)\n      %33 : Float(64)\n      %34 : Float(64)\n      %35 : Float(64)\n      %36 : Long()\n      %37 : Float(64, 64, 3, 3)\n      %38 : Float(64)\n      %39 : Float(64)\n      %40 : Float(64)\n      %41 : Float(64)\n      %42 : Long()\n      %43 : Float(256, 64, 1, 1)\n      %44 : Float(256)\n      %45 : Float(256)\n      %46 : Float(256)\n      %47 : Float(256)\n      %48 : Long()\n      %49 : Float(64, 256, 1, 1)\n      %50 : Float(64)\n      %51 : Float(64)\n      %52 : Float(64)\n      %53 : Float(64)\n      %54 : Long()\n      %55 : Float(64, 64, 3, 3)\n      %56 : Float(64)\n      %57 : Float(64)\n      %58 : Float(64)\n      %59 : Float(64)\n      %60 : Long()\n      %61 : Float(256, 64, 1, 1)\n      %62 : Float(256)\n      %63 : Float(256)\n      %64 : Float(256)\n      %65 : Float(256)\n      %66 : Long()\n      %67 : Float(128, 256, 1, 1)\n      %68 : Float(128)\n      %69 : Float(128)\n      %70 : Float(128)\n      %71 : Float(128)\n      %72 : Long()\n      %73 : Float(128, 128, 3, 3)\n      %74 : Float(128)\n      %75 : Float(128)\n      %76 : Float(128)\n      %77 : Float(128)\n      %78 : Long()\n      %79 : Float(512, 128, 1, 1)\n      %80 : Float(512)\n      %81 : Float(512)\n      %82 : Float(512)\n      %83 : Float(512)\n      %84 : Long()\n      %85 : Float(512, 256, 1, 1)\n      %86 : Float(512)\n      %87 : Float(512)\n      %88 : Float(512)\n      %89 : Float(512)\n      %90 : Long()\n      %91 : Float(128, 512, 1, 1)\n      %92 : Float(128)\n      %93 : Float(128)\n      %94 : Float(128)\n      %95 : Float(128)\n      %96 : Long()\n      %97 : Float(128, 128, 3, 3)\n      %98 : Float(128)\n      %99 : Float(128)\n      %100 : Float(128)\n      %101 : Float(128)\n      %102 : Long()\n      %103 : Float(512, 128, 1, 1)\n      %104 : Float(512)\n      %105 : Float(512)\n      %106 : Float(512)\n      %107 : Float(512)\n      %108 : Long()\n      %109 : Float(128, 512, 1, 1)\n      %110 : Float(128)\n      %111 : Float(128)\n      %112 : Float(128)\n      %113 : Float(128)\n      %114 : Long()\n      %115 : Float(128, 128, 3, 3)\n      %116 : Float(128)\n      %117 : Float(128)\n      %118 : Float(128)\n      %119 : Float(128)\n      %120 : Long()\n      %121 : Float(512, 128, 1, 1)\n      %122 : Float(512)\n      %123 : Float(512)\n      %124 : Float(512)\n      %125 : Float(512)\n      %126 : Long()\n      %127 : Float(128, 512, 1, 1)\n      %128 : Float(128)\n      %129 : Float(128)\n      %130 : Float(128)\n      %131 : Float(128)\n      %132 : Long()\n      %133 : Float(128, 128, 3, 3)\n      %134 : Float(128)\n      %135 : Float(128)\n      %136 : Float(128)\n      %137 : Float(128)\n      %138 : Long()\n      %139 : Float(512, 128, 1, 1)\n      %140 : Float(512)\n      %141 : Float(512)\n      %142 : Float(512)\n      %143 : Float(512)\n      %144 : Long()\n      %145 : Float(256, 512, 1, 1)\n      %146 : Float(256)\n      %147 : Float(256)\n      %148 : Float(256)\n      %149 : Float(256)\n      %150 : Long()\n      %151 : Float(256, 256, 3, 3)\n      %152 : Float(256)\n      %153 : Float(256)\n      %154 : Float(256)\n      %155 : Float(256)\n      %156 : Long()\n      %157 : Float(1024, 256, 1, 1)\n      %158 : Float(1024)\n      %159 : Float(1024)\n      %160 : Float(1024)\n      %161 : Float(1024)\n      %162 : Long()\n      %163 : Float(1024, 512, 1, 1)\n      %164 : Float(1024)\n      %165 : Float(1024)\n      %166 : Float(1024)\n      %167 : Float(1024)\n      %168 : Long()\n      %169 : Float(256, 1024, 1, 1)\n      %170 : Float(256)\n      %171 : Float(256)\n      %172 : Float(256)\n      %173 : Float(256)\n      %174 : Long()\n      %175 : Float(256, 256, 3, 3)\n      %176 : Float(256)\n      %177 : Float(256)\n      %178 : Float(256)\n      %179 : Float(256)\n      %180 : Long()\n      %181 : Float(1024, 256, 1, 1)\n      %182 : Float(1024)\n      %183 : Float(1024)\n      %184 : Float(1024)\n      %185 : Float(1024)\n      %186 : Long()\n      %187 : Float(256, 1024, 1, 1)\n      %188 : Float(256)\n      %189 : Float(256)\n      %190 : Float(256)\n      %191 : Float(256)\n      %192 : Long()\n      %193 : Float(256, 256, 3, 3)\n      %194 : Float(256)\n      %195 : Float(256)\n      %196 : Float(256)\n      %197 : Float(256)\n      %198 : Long()\n      %199 : Float(1024, 256, 1, 1)\n      +0 : Float(1024)\n      +1 : Float(1024)\n      +2 : Float(1024)\n      +3 : Float(1024)\n      +4 : Long()\n      +5 : Float(256, 1024, 1, 1)\n      +6 : Float(256)\n      +7 : Float(256)\n      +8 : Float(256)\n      +9 : Float(256)\n      %210 : Long()\n      %211 : Float(256, 256, 3, 3)\n      %212 : Float(256)\n      %213 : Float(256)\n      %214 : Float(256)\n      %215 : Float(256)\n      %216 : Long()\n      %217 : Float(1024, 256, 1, 1)\n      %218 : Float(1024)\n      %219 : Float(1024)\n      %220 : Float(1024)\n      %221 : Float(1024)\n      %222 : Long()\n      %223 : Float(256, 1024, 1, 1)\n      %224 : Float(256)\n      %225 : Float(256)\n      %226 : Float(256)\n      %227 : Float(256)\n      %228 : Long()\n      %229 : Float(256, 256, 3, 3)\n      %230 : Float(256)\n      %231 : Float(256)\n      %232 : Float(256)\n      %233 : Float(256)\n      %234 : Long()\n      %235 : Float(1024, 256, 1, 1)\n      %236 : Float(1024)\n      %237 : Float(1024)\n      %238 : Float(1024)\n      %239 : Float(1024)\n      %240 : Long()\n      %241 : Float(256, 1024, 1, 1)\n      %242 : Float(256)\n      %243 : Float(256)\n      %244 : Float(256)\n      %245 : Float(256)\n      %246 : Long()\n      %247 : Float(256, 256, 3, 3)\n      %248 : Float(256)\n      %249 : Float(256)\n      %250 : Float(256)\n      %251 : Float(256)\n      %252 : Long()\n      %253 : Float(1024, 256, 1, 1)\n      %254 : Float(1024)\n      %255 : Float(1024)\n      %256 : Float(1024)\n      %257 : Float(1024)\n      %258 : Long()\n      %259 : Float(256, 1024, 1, 1)\n      %260 : Float(256)\n      %261 : Float(256)\n      %262 : Float(256)\n      %263 : Float(256)\n      %264 : Long()\n      %265 : Float(256, 256, 3, 3)\n      %266 : Float(256)\n      %267 : Float(256)\n      %268 : Float(256)\n      %269 : Float(256)\n      %270 : Long()\n      %271 : Float(1024, 256, 1, 1)\n      %272 : Float(1024)\n      %273 : Float(1024)\n      %274 : Float(1024)\n      %275 : Float(1024)\n      %276 : Long()\n      %277 : Float(256, 1024, 1, 1)\n      %278 : Float(256)\n      %279 : Float(256)\n      %280 : Float(256)\n      %281 : Float(256)\n      %282 : Long()\n      %283 : Float(256, 256, 3, 3)\n      %284 : Float(256)\n      %285 : Float(256)\n      %286 : Float(256)\n      %287 : Float(256)\n      %288 : Long()\n      %289 : Float(1024, 256, 1, 1)\n      %290 : Float(1024)\n      %291 : Float(1024)\n      %292 : Float(1024)\n      %293 : Float(1024)\n      %294 : Long()\n      %295 : Float(256, 1024, 1, 1)\n      %296 : Float(256)\n      %297 : Float(256)\n      %298 : Float(256)\n      %299 : Float(256)\n      %300 : Long()\n      %301 : Float(256, 256, 3, 3)\n      %302 : Float(256)\n      %303 : Float(256)\n      %304 : Float(256)\n      %305 : Float(256)\n      %306 : Long()\n      %307 : Float(1024, 256, 1, 1)\n      %308 : Float(1024)\n      %309 : Float(1024)\n      %310 : Float(1024)\n      %311 : Float(1024)\n      %312 : Long()\n      %313 : Float(256, 1024, 1, 1)\n      %314 : Float(256)\n      %315 : Float(256)\n      %316 : Float(256)\n      %317 : Float(256)\n      %318 : Long()\n      %319 : Float(256, 256, 3, 3)\n      %320 : Float(256)\n      %321 : Float(256)\n      %322 : Float(256)\n      %323 : Float(256)\n      %324 : Long()\n      %325 : Float(1024, 256, 1, 1)\n      %326 : Float(1024)\n      %327 : Float(1024)\n      %328 : Float(1024)\n      %329 : Float(1024)\n      %330 : Long()\n      %331 : Float(256, 1024, 1, 1)\n      %332 : Float(256)\n      %333 : Float(256)\n      %334 : Float(256)\n      %335 : Float(256)\n      %336 : Long()\n      %337 : Float(256, 256, 3, 3)\n      %338 : Float(256)\n      %339 : Float(256)\n      %340 : Float(256)\n      %341 : Float(256)\n      %342 : Long()\n      %343 : Float(1024, 256, 1, 1)\n      %344 : Float(1024)\n      %345 : Float(1024)\n      %346 : Float(1024)\n      %347 : Float(1024)\n      %348 : Long()\n      %349 : Float(256, 1024, 1, 1)\n      %350 : Float(256)\n      %351 : Float(256)\n      %352 : Float(256)\n      %353 : Float(256)\n      %354 : Long()\n      %355 : Float(256, 256, 3, 3)\n      %356 : Float(256)\n      %357 : Float(256)\n      %358 : Float(256)\n      %359 : Float(256)\n      %360 : Long()\n      %361 : Float(1024, 256, 1, 1)\n      %362 : Float(1024)\n      %363 : Float(1024)\n      %364 : Float(1024)\n      %365 : Float(1024)\n      %366 : Long()\n      %367 : Float(256, 1024, 1, 1)\n      %368 : Float(256)\n      %369 : Float(256)\n      %370 : Float(256)\n      %371 : Float(256)\n      %372 : Long()\n      %373 : Float(256, 256, 3, 3)\n      %374 : Float(256)\n      %375 : Float(256)\n      %376 : Float(256)\n      %377 : Float(256)\n      %378 : Long()\n      %379 : Float(1024, 256, 1, 1)\n      %380 : Float(1024)\n      %381 : Float(1024)\n      %382 : Float(1024)\n      %383 : Float(1024)\n      %384 : Long()\n      %385 : Float(256, 1024, 1, 1)\n      %386 : Float(256)\n      %387 : Float(256)\n      %388 : Float(256)\n      %389 : Float(256)\n      %390 : Long()\n      %391 : Float(256, 256, 3, 3)\n      %392 : Float(256)\n      %393 : Float(256)\n      %394 : Float(256)\n      %395 : Float(256)\n      %396 : Long()\n      %397 : Float(1024, 256, 1, 1)\n      %398 : Float(1024)\n      %399 : Float(1024)\n      %400 : Float(1024)\n      %401 : Float(1024)\n      %402 : Long()\n      %403 : Float(256, 1024, 1, 1)\n      %404 : Float(256)\n      %405 : Float(256)\n      %406 : Float(256)\n      %407 : Float(256)\n      %408 : Long()\n      %409 : Float(256, 256, 3, 3)\n      %410 : Float(256)\n      %411 : Float(256)\n      %412 : Float(256)\n      %413 : Float(256)\n      %414 : Long()\n      %415 : Float(1024, 256, 1, 1)\n      %416 : Float(1024)\n      %417 : Float(1024)\n      %418 : Float(1024)\n      %419 : Float(1024)\n      %420 : Long()\n      %421 : Float(256, 1024, 1, 1)\n      %422 : Float(256)\n      %423 : Float(256)\n      %424 : Float(256)\n      %425 : Float(256)\n      %426 : Long()\n      %427 : Float(256, 256, 3, 3)\n      %428 : Float(256)\n      %429 : Float(256)\n      %430 : Float(256)\n      %431 : Float(256)\n      %432 : Long()\n      %433 : Float(1024, 256, 1, 1)\n      %434 : Float(1024)\n      %435 : Float(1024)\n      %436 : Float(1024)\n      %437 : Float(1024)\n      %438 : Long()\n      %439 : Float(256, 1024, 1, 1)\n      %440 : Float(256)\n      %441 : Float(256)\n      %442 : Float(256)\n      %443 : Float(256)\n      %444 : Long()\n      %445 : Float(256, 256, 3, 3)\n      %446 : Float(256)\n      %447 : Float(256)\n      %448 : Float(256)\n      %449 : Float(256)\n      %450 : Long()\n      %451 : Float(1024, 256, 1, 1)\n      %452 : Float(1024)\n      %453 : Float(1024)\n      %454 : Float(1024)\n      %455 : Float(1024)\n      %456 : Long()\n      %457 : Float(256, 1024, 1, 1)\n      %458 : Float(256)\n      %459 : Float(256)\n      %460 : Float(256)\n      %461 : Float(256)\n      %462 : Long()\n      %463 : Float(256, 256, 3, 3)\n      %464 : Float(256)\n      %465 : Float(256)\n      %466 : Float(256)\n      %467 : Float(256)\n      %468 : Long()\n      %469 : Float(1024, 256, 1, 1)\n      %470 : Float(1024)\n      %471 : Float(1024)\n      %472 : Float(1024)\n      %473 : Float(1024)\n      %474 : Long()\n      %475 : Float(256, 1024, 1, 1)\n      %476 : Float(256)\n      %477 : Float(256)\n      %478 : Float(256)\n      %479 : Float(256)\n      %480 : Long()\n      %481 : Float(256, 256, 3, 3)\n      %482 : Float(256)\n      %483 : Float(256)\n      %484 : Float(256)\n      %485 : Float(256)\n      %486 : Long()\n      %487 : Float(1024, 256, 1, 1)\n      %488 : Float(1024)\n      %489 : Float(1024)\n      %490 : Float(1024)\n      %491 : Float(1024)\n      %492 : Long()\n      %493 : Float(256, 1024, 1, 1)\n      %494 : Float(256)\n      %495 : Float(256)\n      %496 : Float(256)\n      %497 : Float(256)\n      %498 : Long()\n      %499 : Float(256, 256, 3, 3)\n      %500 : Float(256)\n      %501 : Float(256)\n      %502 : Float(256)\n      %503 : Float(256)\n      %504 : Long()\n      %505 : Float(1024, 256, 1, 1)\n      %506 : Float(1024)\n      %507 : Float(1024)\n      %508 : Float(1024)\n      %509 : Float(1024)\n      %510 : Long()\n      %511 : Float(256, 1024, 1, 1)\n      %512 : Float(256)\n      %513 : Float(256)\n      %514 : Float(256)\n      %515 : Float(256)\n      %516 : Long()\n      %517 : Float(256, 256, 3, 3)\n      %518 : Float(256)\n      %519 : Float(256)\n      %520 : Float(256)\n      %521 : Float(256)\n      %522 : Long()\n      %523 : Float(1024, 256, 1, 1)\n      %524 : Float(1024)\n      %525 : Float(1024)\n      %526 : Float(1024)\n      %527 : Float(1024)\n      %528 : Long()\n      %529 : Float(256, 1024, 1, 1)\n      %530 : Float(256)\n      %531 : Float(256)\n      %532 : Float(256)\n      %533 : Float(256)\n      %534 : Long()\n      %535 : Float(256, 256, 3, 3)\n      %536 : Float(256)\n      %537 : Float(256)\n      %538 : Float(256)\n      %539 : Float(256)\n      %540 : Long()\n      %541 : Float(1024, 256, 1, 1)\n      %542 : Float(1024)\n      %543 : Float(1024)\n      %544 : Float(1024)\n      %545 : Float(1024)\n      %546 : Long()\n      %547 : Float(256, 1024, 1, 1)\n      %548 : Float(256)\n      %549 : Float(256)\n      %550 : Float(256)\n      %551 : Float(256)\n      %552 : Long()\n      %553 : Float(256, 256, 3, 3)\n      %554 : Float(256)\n      %555 : Float(256)\n      %556 : Float(256)\n      %557 : Float(256)\n      %558 : Long()\n      %559 : Float(1024, 256, 1, 1)\n      %560 : Float(1024)\n      %561 : Float(1024)\n      %562 : Float(1024)\n      %563 : Float(1024)\n      %564 : Long()\n      %565 : Float(512, 1024, 1, 1)\n      %566 : Float(512)\n      %567 : Float(512)\n      %568 : Float(512)\n      %569 : Float(512)\n      %570 : Long()\n      %571 : Float(512, 512, 3, 3)\n      %572 : Float(512)\n      %573 : Float(512)\n      %574 : Float(512)\n      %575 : Float(512)\n      %576 : Long()\n      %577 : Float(2048, 512, 1, 1)\n      %578 : Float(2048)\n      %579 : Float(2048)\n      %580 : Float(2048)\n      %581 : Float(2048)\n      %582 : Long()\n      %583 : Float(2048, 1024, 1, 1)\n      %584 : Float(2048)\n      %585 : Float(2048)\n      %586 : Float(2048)\n      %587 : Float(2048)\n      %588 : Long()\n      %589 : Float(512, 2048, 1, 1)\n      %590 : Float(512)\n      %591 : Float(512)\n      %592 : Float(512)\n      %593 : Float(512)\n      %594 : Long()\n      %595 : Float(512, 512, 3, 3)\n      %596 : Float(512)\n      %597 : Float(512)\n      %598 : Float(512)\n      %599 : Float(512)\n      %600 : Long()\n      %601 : Float(2048, 512, 1, 1)\n      %602 : Float(2048)\n      %603 : Float(2048)\n      %604 : Float(2048)\n      %605 : Float(2048)\n      %606 : Long()\n      %607 : Float(512, 2048, 1, 1)\n      %608 : Float(512)\n      %609 : Float(512)\n      %610 : Float(512)\n      %611 : Float(512)\n      %612 : Long()\n      %613 : Float(512, 512, 3, 3)\n      %614 : Float(512)\n      %615 : Float(512)\n      %616 : Float(512)\n      %617 : Float(512)\n      %618 : Long()\n      %619 : Float(2048, 512, 1, 1)\n      %620 : Float(2048)\n      %621 : Float(2048)\n      %622 : Float(2048)\n      %623 : Float(2048)\n      %624 : Long()\n      %625 : Float(1000, 2048)\n      %626 : Float(1000)) {\n  %627 : Float(32, 64, 112, 112) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[7, 7], pads=[3, 3, 3, 3], strides=[2, 2]](%input_0, %1), scope: ResNet/Conv2d[conv1]\n  %628 : Float(32, 64, 112, 112) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%627, %2, %3, %4, %5), scope: ResNet/BatchNorm2d[bn1]\n  %629 : Float(32, 64, 112, 112) = onnx::Relu(%628), scope: ResNet/ReLU[relu]\n  %630 : Float(32, 64, 56, 56) = onnx::MaxPool[kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%629), scope: ResNet/MaxPool2d[maxpool]\n  %631 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%630, %7), scope: ResNet/Sequential[layer1]/Bottleneck[0]/Conv2d[conv1]\n  %632 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%631, %8, %9, %10, %11), scope: ResNet/Sequential[layer1]/Bottleneck[0]/BatchNorm2d[bn1]\n  %633 : Float(32, 64, 56, 56) = onnx::Relu(%632), scope: ResNet/Sequential[layer1]/Bottleneck[0]/ReLU[relu]\n  %634 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%633, %13), scope: ResNet/Sequential[layer1]/Bottleneck[0]/Conv2d[conv2]\n  %635 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%634, %14, %15, %16, %17), scope: ResNet/Sequential[layer1]/Bottleneck[0]/BatchNorm2d[bn2]\n  %636 : Float(32, 64, 56, 56) = onnx::Relu(%635), scope: ResNet/Sequential[layer1]/Bottleneck[0]/ReLU[relu]\n  %637 : Float(32, 256, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%636, %19), scope: ResNet/Sequential[layer1]/Bottleneck[0]/Conv2d[conv3]\n  %638 : Float(32, 256, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%637, +, %21, %22, %23), scope: ResNet/Sequential[layer1]/Bottleneck[0]/BatchNorm2d[bn3]\n  %639 : Float(32, 256, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%630, %25), scope: ResNet/Sequential[layer1]/Bottleneck[0]/Sequential[downsample]/Conv2d[0]\n  %640 : Float(32, 256, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%639, %26, %27, %28, %29), scope: ResNet/Sequential[layer1]/Bottleneck[0]/Sequential[downsample]/BatchNorm2d[1]\n  %641 : Float(32, 256, 56, 56) = onnx::Add(%638, %640), scope: ResNet/Sequential[layer1]/Bottleneck[0]\n  %642 : Float(32, 256, 56, 56) = onnx::Relu(%641), scope: ResNet/Sequential[layer1]/Bottleneck[0]/ReLU[relu]\n  %643 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%642, %31), scope: ResNet/Sequential[layer1]/Bottleneck[1]/Conv2d[conv1]\n  %644 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%643, %32, %33, %34, %35), scope: ResNet/Sequential[layer1]/Bottleneck[1]/BatchNorm2d[bn1]\n  %645 : Float(32, 64, 56, 56) = onnx::Relu(%644), scope: ResNet/Sequential[layer1]/Bottleneck[1]/ReLU[relu]\n  %646 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%645, %37), scope: ResNet/Sequential[layer1]/Bottleneck[1]/Conv2d[conv2]\n  %647 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%646, %38, %39, %40, %41), scope: ResNet/Sequential[layer1]/Bottleneck[1]/BatchNorm2d[bn2]\n  %648 : Float(32, 64, 56, 56) = onnx::Relu(%647), scope: ResNet/Sequential[layer1]/Bottleneck[1]/ReLU[relu]\n  %649 : Float(32, 256, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%648, %43), scope: ResNet/Sequential[layer1]/Bottleneck[1]/Conv2d[conv3]\n  %650 : Float(32, 256, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%649, %44, %45, %46, %47), scope: ResNet/Sequential[layer1]/Bottleneck[1]/BatchNorm2d[bn3]\n  %651 : Float(32, 256, 56, 56) = onnx::Add(%650, %642), scope: ResNet/Sequential[layer1]/Bottleneck[1]\n  %652 : Float(32, 256, 56, 56) = onnx::Relu(%651), scope: ResNet/Sequential[layer1]/Bottleneck[1]/ReLU[relu]\n  %653 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%652, %49), scope: ResNet/Sequential[layer1]/Bottleneck[2]/Conv2d[conv1]\n  %654 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%653, %50, %51, %52, %53), scope: ResNet/Sequential[layer1]/Bottleneck[2]/BatchNorm2d[bn1]\n  %655 : Float(32, 64, 56, 56) = onnx::Relu(%654), scope: ResNet/Sequential[layer1]/Bottleneck[2]/ReLU[relu]\n  %656 : Float(32, 64, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%655, %55), scope: ResNet/Sequential[layer1]/Bottleneck[2]/Conv2d[conv2]\n  %657 : Float(32, 64, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%656, %56, %57, %58, %59), scope: ResNet/Sequential[layer1]/Bottleneck[2]/BatchNorm2d[bn2]\n  %658 : Float(32, 64, 56, 56) = onnx::Relu(%657), scope: ResNet/Sequential[layer1]/Bottleneck[2]/ReLU[relu]\n  %659 : Float(32, 256, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%658, %61), scope: ResNet/Sequential[layer1]/Bottleneck[2]/Conv2d[conv3]\n  %660 : Float(32, 256, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%659, %62, %63, %64, %65), scope: ResNet/Sequential[layer1]/Bottleneck[2]/BatchNorm2d[bn3]\n  %661 : Float(32, 256, 56, 56) = onnx::Add(%660, %652), scope: ResNet/Sequential[layer1]/Bottleneck[2]\n  %662 : Float(32, 256, 56, 56) = onnx::Relu(%661), scope: ResNet/Sequential[layer1]/Bottleneck[2]/ReLU[relu]\n  %663 : Float(32, 128, 56, 56) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%662, %67), scope: ResNet/Sequential[layer2]/Bottleneck[0]/Conv2d[conv1]\n  %664 : Float(32, 128, 56, 56) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%663, %68, %69, %70, %71), scope: ResNet/Sequential[layer2]/Bottleneck[0]/BatchNorm2d[bn1]\n  %665 : Float(32, 128, 56, 56) = onnx::Relu(%664), scope: ResNet/Sequential[layer2]/Bottleneck[0]/ReLU[relu]\n  %666 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%665, %73), scope: ResNet/Sequential[layer2]/Bottleneck[0]/Conv2d[conv2]\n  %667 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%666, %74, %75, %76, %77), scope: ResNet/Sequential[layer2]/Bottleneck[0]/BatchNorm2d[bn2]\n  %668 : Float(32, 128, 28, 28) = onnx::Relu(%667), scope: ResNet/Sequential[layer2]/Bottleneck[0]/ReLU[relu]\n  %669 : Float(32, 512, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%668, %79), scope: ResNet/Sequential[layer2]/Bottleneck[0]/Conv2d[conv3]\n  %670 : Float(32, 512, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%669, %80, %81, %82, %83), scope: ResNet/Sequential[layer2]/Bottleneck[0]/BatchNorm2d[bn3]\n  %671 : Float(32, 512, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[2, 2]](%662, %85), scope: ResNet/Sequential[layer2]/Bottleneck[0]/Sequential[downsample]/Conv2d[0]\n  %672 : Float(32, 512, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%671, %86, %87, %88, %89), scope: ResNet/Sequential[layer2]/Bottleneck[0]/Sequential[downsample]/BatchNorm2d[1]\n  %673 : Float(32, 512, 28, 28) = onnx::Add(%670, %672), scope: ResNet/Sequential[layer2]/Bottleneck[0]\n  %674 : Float(32, 512, 28, 28) = onnx::Relu(%673), scope: ResNet/Sequential[layer2]/Bottleneck[0]/ReLU[relu]\n  %675 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%674, %91), scope: ResNet/Sequential[layer2]/Bottleneck[1]/Conv2d[conv1]\n  %676 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%675, %92, %93, %94, %95), scope: ResNet/Sequential[layer2]/Bottleneck[1]/BatchNorm2d[bn1]\n  %677 : Float(32, 128, 28, 28) = onnx::Relu(%676), scope: ResNet/Sequential[layer2]/Bottleneck[1]/ReLU[relu]\n  %678 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%677, %97), scope: ResNet/Sequential[layer2]/Bottleneck[1]/Conv2d[conv2]\n  %679 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%678, %98, %99, %100, %101), scope: ResNet/Sequential[layer2]/Bottleneck[1]/BatchNorm2d[bn2]\n  %680 : Float(32, 128, 28, 28) = onnx::Relu(%679), scope: ResNet/Sequential[layer2]/Bottleneck[1]/ReLU[relu]\n  %681 : Float(32, 512, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%680, %103), scope: ResNet/Sequential[layer2]/Bottleneck[1]/Conv2d[conv3]\n  %682 : Float(32, 512, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%681, %104, %105, %106, %107), scope: ResNet/Sequential[layer2]/Bottleneck[1]/BatchNorm2d[bn3]\n  %683 : Float(32, 512, 28, 28) = onnx::Add(%682, %674), scope: ResNet/Sequential[layer2]/Bottleneck[1]\n  %684 : Float(32, 512, 28, 28) = onnx::Relu(%683), scope: ResNet/Sequential[layer2]/Bottleneck[1]/ReLU[relu]\n  %685 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%684, %109), scope: ResNet/Sequential[layer2]/Bottleneck[2]/Conv2d[conv1]\n  %686 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%685, %110, %111, %112, %113), scope: ResNet/Sequential[layer2]/Bottleneck[2]/BatchNorm2d[bn1]\n  %687 : Float(32, 128, 28, 28) = onnx::Relu(%686), scope: ResNet/Sequential[layer2]/Bottleneck[2]/ReLU[relu]\n  %688 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%687, %115), scope: ResNet/Sequential[layer2]/Bottleneck[2]/Conv2d[conv2]\n  %689 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%688, %116, %117, %118, %119), scope: ResNet/Sequential[layer2]/Bottleneck[2]/BatchNorm2d[bn2]\n  %690 : Float(32, 128, 28, 28) = onnx::Relu(%689), scope: ResNet/Sequential[layer2]/Bottleneck[2]/ReLU[relu]\n  %691 : Float(32, 512, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%690, %121), scope: ResNet/Sequential[layer2]/Bottleneck[2]/Conv2d[conv3]\n  %692 : Float(32, 512, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%691, %122, %123, %124, %125), scope: ResNet/Sequential[layer2]/Bottleneck[2]/BatchNorm2d[bn3]\n  %693 : Float(32, 512, 28, 28) = onnx::Add(%692, %684), scope: ResNet/Sequential[layer2]/Bottleneck[2]\n  %694 : Float(32, 512, 28, 28) = onnx::Relu(%693), scope: ResNet/Sequential[layer2]/Bottleneck[2]/ReLU[relu]\n  %695 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%694, %127), scope: ResNet/Sequential[layer2]/Bottleneck[3]/Conv2d[conv1]\n  %696 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%695, %128, %129, %130, %131), scope: ResNet/Sequential[layer2]/Bottleneck[3]/BatchNorm2d[bn1]\n  %697 : Float(32, 128, 28, 28) = onnx::Relu(%696), scope: ResNet/Sequential[layer2]/Bottleneck[3]/ReLU[relu]\n  %698 : Float(32, 128, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%697, %133), scope: ResNet/Sequential[layer2]/Bottleneck[3]/Conv2d[conv2]\n  %699 : Float(32, 128, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%698, %134, %135, %136, %137), scope: ResNet/Sequential[layer2]/Bottleneck[3]/BatchNorm2d[bn2]\n  %700 : Float(32, 128, 28, 28) = onnx::Relu(%699), scope: ResNet/Sequential[layer2]/Bottleneck[3]/ReLU[relu]\n  %701 : Float(32, 512, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%700, %139), scope: ResNet/Sequential[layer2]/Bottleneck[3]/Conv2d[conv3]\n  %702 : Float(32, 512, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%701, %140, %141, %142, %143), scope: ResNet/Sequential[layer2]/Bottleneck[3]/BatchNorm2d[bn3]\n  %703 : Float(32, 512, 28, 28) = onnx::Add(%702, %694), scope: ResNet/Sequential[layer2]/Bottleneck[3]\n  %704 : Float(32, 512, 28, 28) = onnx::Relu(%703), scope: ResNet/Sequential[layer2]/Bottleneck[3]/ReLU[relu]\n  %705 : Float(32, 256, 28, 28) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%704, %145), scope: ResNet/Sequential[layer3]/Bottleneck[0]/Conv2d[conv1]\n  %706 : Float(32, 256, 28, 28) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%705, %146, %147, %148, %149), scope: ResNet/Sequential[layer3]/Bottleneck[0]/BatchNorm2d[bn1]\n  %707 : Float(32, 256, 28, 28) = onnx::Relu(%706), scope: ResNet/Sequential[layer3]/Bottleneck[0]/ReLU[relu]\n  %708 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%707, %151), scope: ResNet/Sequential[layer3]/Bottleneck[0]/Conv2d[conv2]\n  %709 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%708, %152, %153, %154, %155), scope: ResNet/Sequential[layer3]/Bottleneck[0]/BatchNorm2d[bn2]\n  %710 : Float(32, 256, 14, 14) = onnx::Relu(%709), scope: ResNet/Sequential[layer3]/Bottleneck[0]/ReLU[relu]\n  %711 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%710, %157), scope: ResNet/Sequential[layer3]/Bottleneck[0]/Conv2d[conv3]\n  %712 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%711, %158, %159, %160, %161), scope: ResNet/Sequential[layer3]/Bottleneck[0]/BatchNorm2d[bn3]\n  %713 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[2, 2]](%704, %163), scope: ResNet/Sequential[layer3]/Bottleneck[0]/Sequential[downsample]/Conv2d[0]\n  %714 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%713, %164, %165, %166, %167), scope: ResNet/Sequential[layer3]/Bottleneck[0]/Sequential[downsample]/BatchNorm2d[1]\n  %715 : Float(32, 1024, 14, 14) = onnx::Add(%712, %714), scope: ResNet/Sequential[layer3]/Bottleneck[0]\n  %716 : Float(32, 1024, 14, 14) = onnx::Relu(%715), scope: ResNet/Sequential[layer3]/Bottleneck[0]/ReLU[relu]\n  %717 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%716, %169), scope: ResNet/Sequential[layer3]/Bottleneck[1]/Conv2d[conv1]\n  %718 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%717, %170, %171, %172, %173), scope: ResNet/Sequential[layer3]/Bottleneck[1]/BatchNorm2d[bn1]\n  %719 : Float(32, 256, 14, 14) = onnx::Relu(%718), scope: ResNet/Sequential[layer3]/Bottleneck[1]/ReLU[relu]\n  %720 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%719, %175), scope: ResNet/Sequential[layer3]/Bottleneck[1]/Conv2d[conv2]\n  %721 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%720, %176, %177, %178, %179), scope: ResNet/Sequential[layer3]/Bottleneck[1]/BatchNorm2d[bn2]\n  %722 : Float(32, 256, 14, 14) = onnx::Relu(%721), scope: ResNet/Sequential[layer3]/Bottleneck[1]/ReLU[relu]\n  %723 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%722, %181), scope: ResNet/Sequential[layer3]/Bottleneck[1]/Conv2d[conv3]\n  %724 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%723, %182, %183, %184, %185), scope: ResNet/Sequential[layer3]/Bottleneck[1]/BatchNorm2d[bn3]\n  %725 : Float(32, 1024, 14, 14) = onnx::Add(%724, %716), scope: ResNet/Sequential[layer3]/Bottleneck[1]\n  %726 : Float(32, 1024, 14, 14) = onnx::Relu(%725), scope: ResNet/Sequential[layer3]/Bottleneck[1]/ReLU[relu]\n  %727 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%726, %187), scope: ResNet/Sequential[layer3]/Bottleneck[2]/Conv2d[conv1]\n  %728 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%727, %188, %189, %190, %191), scope: ResNet/Sequential[layer3]/Bottleneck[2]/BatchNorm2d[bn1]\n  %729 : Float(32, 256, 14, 14) = onnx::Relu(%728), scope: ResNet/Sequential[layer3]/Bottleneck[2]/ReLU[relu]\n  %730 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%729, %193), scope: ResNet/Sequential[layer3]/Bottleneck[2]/Conv2d[conv2]\n  %731 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%730, %194, %195, %196, %197), scope: ResNet/Sequential[layer3]/Bottleneck[2]/BatchNorm2d[bn2]\n  %732 : Float(32, 256, 14, 14) = onnx::Relu(%731), scope: ResNet/Sequential[layer3]/Bottleneck[2]/ReLU[relu]\n  %733 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%732, %199), scope: ResNet/Sequential[layer3]/Bottleneck[2]/Conv2d[conv3]\n  %734 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%733, +0, +1, +2, +3), scope: ResNet/Sequential[layer3]/Bottleneck[2]/BatchNorm2d[bn3]\n  %735 : Float(32, 1024, 14, 14) = onnx::Add(%734, %726), scope: ResNet/Sequential[layer3]/Bottleneck[2]\n  %736 : Float(32, 1024, 14, 14) = onnx::Relu(%735), scope: ResNet/Sequential[layer3]/Bottleneck[2]/ReLU[relu]\n  %737 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%736, +5), scope: ResNet/Sequential[layer3]/Bottleneck[3]/Conv2d[conv1]\n  %738 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%737, +6, +7, +8, +9), scope: ResNet/Sequential[layer3]/Bottleneck[3]/BatchNorm2d[bn1]\n  %739 : Float(32, 256, 14, 14) = onnx::Relu(%738), scope: ResNet/Sequential[layer3]/Bottleneck[3]/ReLU[relu]\n  %740 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%739, %211), scope: ResNet/Sequential[layer3]/Bottleneck[3]/Conv2d[conv2]\n  %741 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%740, %212, %213, %214, %215), scope: ResNet/Sequential[layer3]/Bottleneck[3]/BatchNorm2d[bn2]\n  %742 : Float(32, 256, 14, 14) = onnx::Relu(%741), scope: ResNet/Sequential[layer3]/Bottleneck[3]/ReLU[relu]\n  %743 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%742, %217), scope: ResNet/Sequential[layer3]/Bottleneck[3]/Conv2d[conv3]\n  %744 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%743, %218, %219, %220, %221), scope: ResNet/Sequential[layer3]/Bottleneck[3]/BatchNorm2d[bn3]\n  %745 : Float(32, 1024, 14, 14) = onnx::Add(%744, %736), scope: ResNet/Sequential[layer3]/Bottleneck[3]\n  %746 : Float(32, 1024, 14, 14) = onnx::Relu(%745), scope: ResNet/Sequential[layer3]/Bottleneck[3]/ReLU[relu]\n  %747 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%746, %223), scope: ResNet/Sequential[layer3]/Bottleneck[4]/Conv2d[conv1]\n  %748 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%747, %224, %225, %226, %227), scope: ResNet/Sequential[layer3]/Bottleneck[4]/BatchNorm2d[bn1]\n  %749 : Float(32, 256, 14, 14) = onnx::Relu(%748), scope: ResNet/Sequential[layer3]/Bottleneck[4]/ReLU[relu]\n  %750 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%749, %229), scope: ResNet/Sequential[layer3]/Bottleneck[4]/Conv2d[conv2]\n  %751 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%750, %230, %231, %232, %233), scope: ResNet/Sequential[layer3]/Bottleneck[4]/BatchNorm2d[bn2]\n  %752 : Float(32, 256, 14, 14) = onnx::Relu(%751), scope: ResNet/Sequential[layer3]/Bottleneck[4]/ReLU[relu]\n  %753 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%752, %235), scope: ResNet/Sequential[layer3]/Bottleneck[4]/Conv2d[conv3]\n  %754 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%753, %236, %237, %238, %239), scope: ResNet/Sequential[layer3]/Bottleneck[4]/BatchNorm2d[bn3]\n  %755 : Float(32, 1024, 14, 14) = onnx::Add(%754, %746), scope: ResNet/Sequential[layer3]/Bottleneck[4]\n  %756 : Float(32, 1024, 14, 14) = onnx::Relu(%755), scope: ResNet/Sequential[layer3]/Bottleneck[4]/ReLU[relu]\n  %757 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%756, %241), scope: ResNet/Sequential[layer3]/Bottleneck[5]/Conv2d[conv1]\n  %758 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%757, %242, %243, %244, %245), scope: ResNet/Sequential[layer3]/Bottleneck[5]/BatchNorm2d[bn1]\n  %759 : Float(32, 256, 14, 14) = onnx::Relu(%758), scope: ResNet/Sequential[layer3]/Bottleneck[5]/ReLU[relu]\n  %760 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%759, %247), scope: ResNet/Sequential[layer3]/Bottleneck[5]/Conv2d[conv2]\n  %761 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%760, %248, %249, %250, %251), scope: ResNet/Sequential[layer3]/Bottleneck[5]/BatchNorm2d[bn2]\n  %762 : Float(32, 256, 14, 14) = onnx::Relu(%761), scope: ResNet/Sequential[layer3]/Bottleneck[5]/ReLU[relu]\n  %763 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%762, %253), scope: ResNet/Sequential[layer3]/Bottleneck[5]/Conv2d[conv3]\n  %764 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%763, %254, %255, %256, %257), scope: ResNet/Sequential[layer3]/Bottleneck[5]/BatchNorm2d[bn3]\n  %765 : Float(32, 1024, 14, 14) = onnx::Add(%764, %756), scope: ResNet/Sequential[layer3]/Bottleneck[5]\n  %766 : Float(32, 1024, 14, 14) = onnx::Relu(%765), scope: ResNet/Sequential[layer3]/Bottleneck[5]/ReLU[relu]\n  %767 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%766, %259), scope: ResNet/Sequential[layer3]/Bottleneck[6]/Conv2d[conv1]\n  %768 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%767, %260, %261, %262, %263), scope: ResNet/Sequential[layer3]/Bottleneck[6]/BatchNorm2d[bn1]\n  %769 : Float(32, 256, 14, 14) = onnx::Relu(%768), scope: ResNet/Sequential[layer3]/Bottleneck[6]/ReLU[relu]\n  %770 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%769, %265), scope: ResNet/Sequential[layer3]/Bottleneck[6]/Conv2d[conv2]\n  %771 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%770, %266, %267, %268, %269), scope: ResNet/Sequential[layer3]/Bottleneck[6]/BatchNorm2d[bn2]\n  %772 : Float(32, 256, 14, 14) = onnx::Relu(%771), scope: ResNet/Sequential[layer3]/Bottleneck[6]/ReLU[relu]\n  %773 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%772, %271), scope: ResNet/Sequential[layer3]/Bottleneck[6]/Conv2d[conv3]\n  %774 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%773, %272, %273, %274, %275), scope: ResNet/Sequential[layer3]/Bottleneck[6]/BatchNorm2d[bn3]\n  %775 : Float(32, 1024, 14, 14) = onnx::Add(%774, %766), scope: ResNet/Sequential[layer3]/Bottleneck[6]\n  %776 : Float(32, 1024, 14, 14) = onnx::Relu(%775), scope: ResNet/Sequential[layer3]/Bottleneck[6]/ReLU[relu]\n  %777 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%776, %277), scope: ResNet/Sequential[layer3]/Bottleneck[7]/Conv2d[conv1]\n  %778 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%777, %278, %279, %280, %281), scope: ResNet/Sequential[layer3]/Bottleneck[7]/BatchNorm2d[bn1]\n  %779 : Float(32, 256, 14, 14) = onnx::Relu(%778), scope: ResNet/Sequential[layer3]/Bottleneck[7]/ReLU[relu]\n  %780 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%779, %283), scope: ResNet/Sequential[layer3]/Bottleneck[7]/Conv2d[conv2]\n  %781 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%780, %284, %285, %286, %287), scope: ResNet/Sequential[layer3]/Bottleneck[7]/BatchNorm2d[bn2]\n  %782 : Float(32, 256, 14, 14) = onnx::Relu(%781), scope: ResNet/Sequential[layer3]/Bottleneck[7]/ReLU[relu]\n  %783 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%782, %289), scope: ResNet/Sequential[layer3]/Bottleneck[7]/Conv2d[conv3]\n  %784 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%783, %290, %291, %292, %293), scope: ResNet/Sequential[layer3]/Bottleneck[7]/BatchNorm2d[bn3]\n  %785 : Float(32, 1024, 14, 14) = onnx::Add(%784, %776), scope: ResNet/Sequential[layer3]/Bottleneck[7]\n  %786 : Float(32, 1024, 14, 14) = onnx::Relu(%785), scope: ResNet/Sequential[layer3]/Bottleneck[7]/ReLU[relu]\n  %787 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%786, %295), scope: ResNet/Sequential[layer3]/Bottleneck[8]/Conv2d[conv1]\n  %788 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%787, %296, %297, %298, %299), scope: ResNet/Sequential[layer3]/Bottleneck[8]/BatchNorm2d[bn1]\n  %789 : Float(32, 256, 14, 14) = onnx::Relu(%788), scope: ResNet/Sequential[layer3]/Bottleneck[8]/ReLU[relu]\n  %790 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%789, %301), scope: ResNet/Sequential[layer3]/Bottleneck[8]/Conv2d[conv2]\n  %791 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%790, %302, %303, %304, %305), scope: ResNet/Sequential[layer3]/Bottleneck[8]/BatchNorm2d[bn2]\n  %792 : Float(32, 256, 14, 14) = onnx::Relu(%791), scope: ResNet/Sequential[layer3]/Bottleneck[8]/ReLU[relu]\n  %793 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%792, %307), scope: ResNet/Sequential[layer3]/Bottleneck[8]/Conv2d[conv3]\n  %794 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%793, %308, %309, %310, %311), scope: ResNet/Sequential[layer3]/Bottleneck[8]/BatchNorm2d[bn3]\n  %795 : Float(32, 1024, 14, 14) = onnx::Add(%794, %786), scope: ResNet/Sequential[layer3]/Bottleneck[8]\n  %796 : Float(32, 1024, 14, 14) = onnx::Relu(%795), scope: ResNet/Sequential[layer3]/Bottleneck[8]/ReLU[relu]\n  %797 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%796, %313), scope: ResNet/Sequential[layer3]/Bottleneck[9]/Conv2d[conv1]\n  %798 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%797, %314, %315, %316, %317), scope: ResNet/Sequential[layer3]/Bottleneck[9]/BatchNorm2d[bn1]\n  %799 : Float(32, 256, 14, 14) = onnx::Relu(%798), scope: ResNet/Sequential[layer3]/Bottleneck[9]/ReLU[relu]\n  %800 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%799, %319), scope: ResNet/Sequential[layer3]/Bottleneck[9]/Conv2d[conv2]\n  %801 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%800, %320, %321, %322, %323), scope: ResNet/Sequential[layer3]/Bottleneck[9]/BatchNorm2d[bn2]\n  %802 : Float(32, 256, 14, 14) = onnx::Relu(%801), scope: ResNet/Sequential[layer3]/Bottleneck[9]/ReLU[relu]\n  %803 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%802, %325), scope: ResNet/Sequential[layer3]/Bottleneck[9]/Conv2d[conv3]\n  %804 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%803, %326, %327, %328, %329), scope: ResNet/Sequential[layer3]/Bottleneck[9]/BatchNorm2d[bn3]\n  %805 : Float(32, 1024, 14, 14) = onnx::Add(%804, %796), scope: ResNet/Sequential[layer3]/Bottleneck[9]\n  %806 : Float(32, 1024, 14, 14) = onnx::Relu(%805), scope: ResNet/Sequential[layer3]/Bottleneck[9]/ReLU[relu]\n  %807 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%806, %331), scope: ResNet/Sequential[layer3]/Bottleneck[10]/Conv2d[conv1]\n  %808 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%807, %332, %333, %334, %335), scope: ResNet/Sequential[layer3]/Bottleneck[10]/BatchNorm2d[bn1]\n  %809 : Float(32, 256, 14, 14) = onnx::Relu(%808), scope: ResNet/Sequential[layer3]/Bottleneck[10]/ReLU[relu]\n  %810 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%809, %337), scope: ResNet/Sequential[layer3]/Bottleneck[10]/Conv2d[conv2]\n  %811 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%810, %338, %339, %340, %341), scope: ResNet/Sequential[layer3]/Bottleneck[10]/BatchNorm2d[bn2]\n  %812 : Float(32, 256, 14, 14) = onnx::Relu(%811), scope: ResNet/Sequential[layer3]/Bottleneck[10]/ReLU[relu]\n  %813 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%812, %343), scope: ResNet/Sequential[layer3]/Bottleneck[10]/Conv2d[conv3]\n  %814 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%813, %344, %345, %346, %347), scope: ResNet/Sequential[layer3]/Bottleneck[10]/BatchNorm2d[bn3]\n  %815 : Float(32, 1024, 14, 14) = onnx::Add(%814, %806), scope: ResNet/Sequential[layer3]/Bottleneck[10]\n  %816 : Float(32, 1024, 14, 14) = onnx::Relu(%815), scope: ResNet/Sequential[layer3]/Bottleneck[10]/ReLU[relu]\n  %817 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%816, %349), scope: ResNet/Sequential[layer3]/Bottleneck[11]/Conv2d[conv1]\n  %818 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%817, %350, %351, %352, %353), scope: ResNet/Sequential[layer3]/Bottleneck[11]/BatchNorm2d[bn1]\n  %819 : Float(32, 256, 14, 14) = onnx::Relu(%818), scope: ResNet/Sequential[layer3]/Bottleneck[11]/ReLU[relu]\n  %820 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%819, %355), scope: ResNet/Sequential[layer3]/Bottleneck[11]/Conv2d[conv2]\n  %821 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%820, %356, %357, %358, %359), scope: ResNet/Sequential[layer3]/Bottleneck[11]/BatchNorm2d[bn2]\n  %822 : Float(32, 256, 14, 14) = onnx::Relu(%821), scope: ResNet/Sequential[layer3]/Bottleneck[11]/ReLU[relu]\n  %823 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%822, %361), scope: ResNet/Sequential[layer3]/Bottleneck[11]/Conv2d[conv3]\n  %824 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%823, %362, %363, %364, %365), scope: ResNet/Sequential[layer3]/Bottleneck[11]/BatchNorm2d[bn3]\n  %825 : Float(32, 1024, 14, 14) = onnx::Add(%824, %816), scope: ResNet/Sequential[layer3]/Bottleneck[11]\n  %826 : Float(32, 1024, 14, 14) = onnx::Relu(%825), scope: ResNet/Sequential[layer3]/Bottleneck[11]/ReLU[relu]\n  %827 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%826, %367), scope: ResNet/Sequential[layer3]/Bottleneck[12]/Conv2d[conv1]\n  %828 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%827, %368, %369, %370, %371), scope: ResNet/Sequential[layer3]/Bottleneck[12]/BatchNorm2d[bn1]\n  %829 : Float(32, 256, 14, 14) = onnx::Relu(%828), scope: ResNet/Sequential[layer3]/Bottleneck[12]/ReLU[relu]\n  %830 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%829, %373), scope: ResNet/Sequential[layer3]/Bottleneck[12]/Conv2d[conv2]\n  %831 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%830, %374, %375, %376, %377), scope: ResNet/Sequential[layer3]/Bottleneck[12]/BatchNorm2d[bn2]\n  %832 : Float(32, 256, 14, 14) = onnx::Relu(%831), scope: ResNet/Sequential[layer3]/Bottleneck[12]/ReLU[relu]\n  %833 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%832, %379), scope: ResNet/Sequential[layer3]/Bottleneck[12]/Conv2d[conv3]\n  %834 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%833, %380, %381, %382, %383), scope: ResNet/Sequential[layer3]/Bottleneck[12]/BatchNorm2d[bn3]\n  %835 : Float(32, 1024, 14, 14) = onnx::Add(%834, %826), scope: ResNet/Sequential[layer3]/Bottleneck[12]\n  %836 : Float(32, 1024, 14, 14) = onnx::Relu(%835), scope: ResNet/Sequential[layer3]/Bottleneck[12]/ReLU[relu]\n  %837 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%836, %385), scope: ResNet/Sequential[layer3]/Bottleneck[13]/Conv2d[conv1]\n  %838 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%837, %386, %387, %388, %389), scope: ResNet/Sequential[layer3]/Bottleneck[13]/BatchNorm2d[bn1]\n  %839 : Float(32, 256, 14, 14) = onnx::Relu(%838), scope: ResNet/Sequential[layer3]/Bottleneck[13]/ReLU[relu]\n  %840 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%839, %391), scope: ResNet/Sequential[layer3]/Bottleneck[13]/Conv2d[conv2]\n  %841 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%840, %392, %393, %394, %395), scope: ResNet/Sequential[layer3]/Bottleneck[13]/BatchNorm2d[bn2]\n  %842 : Float(32, 256, 14, 14) = onnx::Relu(%841), scope: ResNet/Sequential[layer3]/Bottleneck[13]/ReLU[relu]\n  %843 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%842, %397), scope: ResNet/Sequential[layer3]/Bottleneck[13]/Conv2d[conv3]\n  %844 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%843, %398, %399, %400, %401), scope: ResNet/Sequential[layer3]/Bottleneck[13]/BatchNorm2d[bn3]\n  %845 : Float(32, 1024, 14, 14) = onnx::Add(%844, %836), scope: ResNet/Sequential[layer3]/Bottleneck[13]\n  %846 : Float(32, 1024, 14, 14) = onnx::Relu(%845), scope: ResNet/Sequential[layer3]/Bottleneck[13]/ReLU[relu]\n  %847 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%846, %403), scope: ResNet/Sequential[layer3]/Bottleneck[14]/Conv2d[conv1]\n  %848 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%847, %404, %405, %406, %407), scope: ResNet/Sequential[layer3]/Bottleneck[14]/BatchNorm2d[bn1]\n  %849 : Float(32, 256, 14, 14) = onnx::Relu(%848), scope: ResNet/Sequential[layer3]/Bottleneck[14]/ReLU[relu]\n  %850 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%849, %409), scope: ResNet/Sequential[layer3]/Bottleneck[14]/Conv2d[conv2]\n  %851 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%850, %410, %411, %412, %413), scope: ResNet/Sequential[layer3]/Bottleneck[14]/BatchNorm2d[bn2]\n  %852 : Float(32, 256, 14, 14) = onnx::Relu(%851), scope: ResNet/Sequential[layer3]/Bottleneck[14]/ReLU[relu]\n  %853 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%852, %415), scope: ResNet/Sequential[layer3]/Bottleneck[14]/Conv2d[conv3]\n  %854 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%853, %416, %417, %418, %419), scope: ResNet/Sequential[layer3]/Bottleneck[14]/BatchNorm2d[bn3]\n  %855 : Float(32, 1024, 14, 14) = onnx::Add(%854, %846), scope: ResNet/Sequential[layer3]/Bottleneck[14]\n  %856 : Float(32, 1024, 14, 14) = onnx::Relu(%855), scope: ResNet/Sequential[layer3]/Bottleneck[14]/ReLU[relu]\n  %857 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%856, %421), scope: ResNet/Sequential[layer3]/Bottleneck[15]/Conv2d[conv1]\n  %858 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%857, %422, %423, %424, %425), scope: ResNet/Sequential[layer3]/Bottleneck[15]/BatchNorm2d[bn1]\n  %859 : Float(32, 256, 14, 14) = onnx::Relu(%858), scope: ResNet/Sequential[layer3]/Bottleneck[15]/ReLU[relu]\n  %860 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%859, %427), scope: ResNet/Sequential[layer3]/Bottleneck[15]/Conv2d[conv2]\n  %861 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%860, %428, %429, %430, %431), scope: ResNet/Sequential[layer3]/Bottleneck[15]/BatchNorm2d[bn2]\n  %862 : Float(32, 256, 14, 14) = onnx::Relu(%861), scope: ResNet/Sequential[layer3]/Bottleneck[15]/ReLU[relu]\n  %863 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%862, %433), scope: ResNet/Sequential[layer3]/Bottleneck[15]/Conv2d[conv3]\n  %864 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%863, %434, %435, %436, %437), scope: ResNet/Sequential[layer3]/Bottleneck[15]/BatchNorm2d[bn3]\n  %865 : Float(32, 1024, 14, 14) = onnx::Add(%864, %856), scope: ResNet/Sequential[layer3]/Bottleneck[15]\n  %866 : Float(32, 1024, 14, 14) = onnx::Relu(%865), scope: ResNet/Sequential[layer3]/Bottleneck[15]/ReLU[relu]\n  %867 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%866, %439), scope: ResNet/Sequential[layer3]/Bottleneck[16]/Conv2d[conv1]\n  %868 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%867, %440, %441, %442, %443), scope: ResNet/Sequential[layer3]/Bottleneck[16]/BatchNorm2d[bn1]\n  %869 : Float(32, 256, 14, 14) = onnx::Relu(%868), scope: ResNet/Sequential[layer3]/Bottleneck[16]/ReLU[relu]\n  %870 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%869, %445), scope: ResNet/Sequential[layer3]/Bottleneck[16]/Conv2d[conv2]\n  %871 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%870, %446, %447, %448, %449), scope: ResNet/Sequential[layer3]/Bottleneck[16]/BatchNorm2d[bn2]\n  %872 : Float(32, 256, 14, 14) = onnx::Relu(%871), scope: ResNet/Sequential[layer3]/Bottleneck[16]/ReLU[relu]\n  %873 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%872, %451), scope: ResNet/Sequential[layer3]/Bottleneck[16]/Conv2d[conv3]\n  %874 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%873, %452, %453, %454, %455), scope: ResNet/Sequential[layer3]/Bottleneck[16]/BatchNorm2d[bn3]\n  %875 : Float(32, 1024, 14, 14) = onnx::Add(%874, %866), scope: ResNet/Sequential[layer3]/Bottleneck[16]\n  %876 : Float(32, 1024, 14, 14) = onnx::Relu(%875), scope: ResNet/Sequential[layer3]/Bottleneck[16]/ReLU[relu]\n  %877 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%876, %457), scope: ResNet/Sequential[layer3]/Bottleneck[17]/Conv2d[conv1]\n  %878 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%877, %458, %459, %460, %461), scope: ResNet/Sequential[layer3]/Bottleneck[17]/BatchNorm2d[bn1]\n  %879 : Float(32, 256, 14, 14) = onnx::Relu(%878), scope: ResNet/Sequential[layer3]/Bottleneck[17]/ReLU[relu]\n  %880 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%879, %463), scope: ResNet/Sequential[layer3]/Bottleneck[17]/Conv2d[conv2]\n  %881 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%880, %464, %465, %466, %467), scope: ResNet/Sequential[layer3]/Bottleneck[17]/BatchNorm2d[bn2]\n  %882 : Float(32, 256, 14, 14) = onnx::Relu(%881), scope: ResNet/Sequential[layer3]/Bottleneck[17]/ReLU[relu]\n  %883 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%882, %469), scope: ResNet/Sequential[layer3]/Bottleneck[17]/Conv2d[conv3]\n  %884 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%883, %470, %471, %472, %473), scope: ResNet/Sequential[layer3]/Bottleneck[17]/BatchNorm2d[bn3]\n  %885 : Float(32, 1024, 14, 14) = onnx::Add(%884, %876), scope: ResNet/Sequential[layer3]/Bottleneck[17]\n  %886 : Float(32, 1024, 14, 14) = onnx::Relu(%885), scope: ResNet/Sequential[layer3]/Bottleneck[17]/ReLU[relu]\n  %887 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%886, %475), scope: ResNet/Sequential[layer3]/Bottleneck[18]/Conv2d[conv1]\n  %888 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%887, %476, %477, %478, %479), scope: ResNet/Sequential[layer3]/Bottleneck[18]/BatchNorm2d[bn1]\n  %889 : Float(32, 256, 14, 14) = onnx::Relu(%888), scope: ResNet/Sequential[layer3]/Bottleneck[18]/ReLU[relu]\n  %890 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%889, %481), scope: ResNet/Sequential[layer3]/Bottleneck[18]/Conv2d[conv2]\n  %891 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%890, %482, %483, %484, %485), scope: ResNet/Sequential[layer3]/Bottleneck[18]/BatchNorm2d[bn2]\n  %892 : Float(32, 256, 14, 14) = onnx::Relu(%891), scope: ResNet/Sequential[layer3]/Bottleneck[18]/ReLU[relu]\n  %893 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%892, %487), scope: ResNet/Sequential[layer3]/Bottleneck[18]/Conv2d[conv3]\n  %894 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%893, %488, %489, %490, %491), scope: ResNet/Sequential[layer3]/Bottleneck[18]/BatchNorm2d[bn3]\n  %895 : Float(32, 1024, 14, 14) = onnx::Add(%894, %886), scope: ResNet/Sequential[layer3]/Bottleneck[18]\n  %896 : Float(32, 1024, 14, 14) = onnx::Relu(%895), scope: ResNet/Sequential[layer3]/Bottleneck[18]/ReLU[relu]\n  %897 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%896, %493), scope: ResNet/Sequential[layer3]/Bottleneck[19]/Conv2d[conv1]\n  %898 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%897, %494, %495, %496, %497), scope: ResNet/Sequential[layer3]/Bottleneck[19]/BatchNorm2d[bn1]\n  %899 : Float(32, 256, 14, 14) = onnx::Relu(%898), scope: ResNet/Sequential[layer3]/Bottleneck[19]/ReLU[relu]\n  %900 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%899, %499), scope: ResNet/Sequential[layer3]/Bottleneck[19]/Conv2d[conv2]\n  %901 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%900, %500, %501, %502, %503), scope: ResNet/Sequential[layer3]/Bottleneck[19]/BatchNorm2d[bn2]\n  %902 : Float(32, 256, 14, 14) = onnx::Relu(%901), scope: ResNet/Sequential[layer3]/Bottleneck[19]/ReLU[relu]\n  %903 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%902, %505), scope: ResNet/Sequential[layer3]/Bottleneck[19]/Conv2d[conv3]\n  %904 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%903, %506, %507, %508, %509), scope: ResNet/Sequential[layer3]/Bottleneck[19]/BatchNorm2d[bn3]\n  %905 : Float(32, 1024, 14, 14) = onnx::Add(%904, %896), scope: ResNet/Sequential[layer3]/Bottleneck[19]\n  %906 : Float(32, 1024, 14, 14) = onnx::Relu(%905), scope: ResNet/Sequential[layer3]/Bottleneck[19]/ReLU[relu]\n  %907 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%906, %511), scope: ResNet/Sequential[layer3]/Bottleneck[20]/Conv2d[conv1]\n  %908 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%907, %512, %513, %514, %515), scope: ResNet/Sequential[layer3]/Bottleneck[20]/BatchNorm2d[bn1]\n  %909 : Float(32, 256, 14, 14) = onnx::Relu(%908), scope: ResNet/Sequential[layer3]/Bottleneck[20]/ReLU[relu]\n  %910 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%909, %517), scope: ResNet/Sequential[layer3]/Bottleneck[20]/Conv2d[conv2]\n  %911 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%910, %518, %519, %520, %521), scope: ResNet/Sequential[layer3]/Bottleneck[20]/BatchNorm2d[bn2]\n  %912 : Float(32, 256, 14, 14) = onnx::Relu(%911), scope: ResNet/Sequential[layer3]/Bottleneck[20]/ReLU[relu]\n  %913 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%912, %523), scope: ResNet/Sequential[layer3]/Bottleneck[20]/Conv2d[conv3]\n  %914 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%913, %524, %525, %526, %527), scope: ResNet/Sequential[layer3]/Bottleneck[20]/BatchNorm2d[bn3]\n  %915 : Float(32, 1024, 14, 14) = onnx::Add(%914, %906), scope: ResNet/Sequential[layer3]/Bottleneck[20]\n  %916 : Float(32, 1024, 14, 14) = onnx::Relu(%915), scope: ResNet/Sequential[layer3]/Bottleneck[20]/ReLU[relu]\n  %917 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%916, %529), scope: ResNet/Sequential[layer3]/Bottleneck[21]/Conv2d[conv1]\n  %918 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%917, %530, %531, %532, %533), scope: ResNet/Sequential[layer3]/Bottleneck[21]/BatchNorm2d[bn1]\n  %919 : Float(32, 256, 14, 14) = onnx::Relu(%918), scope: ResNet/Sequential[layer3]/Bottleneck[21]/ReLU[relu]\n  %920 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%919, %535), scope: ResNet/Sequential[layer3]/Bottleneck[21]/Conv2d[conv2]\n  %921 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%920, %536, %537, %538, %539), scope: ResNet/Sequential[layer3]/Bottleneck[21]/BatchNorm2d[bn2]\n  %922 : Float(32, 256, 14, 14) = onnx::Relu(%921), scope: ResNet/Sequential[layer3]/Bottleneck[21]/ReLU[relu]\n  %923 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%922, %541), scope: ResNet/Sequential[layer3]/Bottleneck[21]/Conv2d[conv3]\n  %924 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%923, %542, %543, %544, %545), scope: ResNet/Sequential[layer3]/Bottleneck[21]/BatchNorm2d[bn3]\n  %925 : Float(32, 1024, 14, 14) = onnx::Add(%924, %916), scope: ResNet/Sequential[layer3]/Bottleneck[21]\n  %926 : Float(32, 1024, 14, 14) = onnx::Relu(%925), scope: ResNet/Sequential[layer3]/Bottleneck[21]/ReLU[relu]\n  %927 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%926, %547), scope: ResNet/Sequential[layer3]/Bottleneck[22]/Conv2d[conv1]\n  %928 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%927, %548, %549, %550, %551), scope: ResNet/Sequential[layer3]/Bottleneck[22]/BatchNorm2d[bn1]\n  %929 : Float(32, 256, 14, 14) = onnx::Relu(%928), scope: ResNet/Sequential[layer3]/Bottleneck[22]/ReLU[relu]\n  %930 : Float(32, 256, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%929, %553), scope: ResNet/Sequential[layer3]/Bottleneck[22]/Conv2d[conv2]\n  %931 : Float(32, 256, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%930, %554, %555, %556, %557), scope: ResNet/Sequential[layer3]/Bottleneck[22]/BatchNorm2d[bn2]\n  %932 : Float(32, 256, 14, 14) = onnx::Relu(%931), scope: ResNet/Sequential[layer3]/Bottleneck[22]/ReLU[relu]\n  %933 : Float(32, 1024, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%932, %559), scope: ResNet/Sequential[layer3]/Bottleneck[22]/Conv2d[conv3]\n  %934 : Float(32, 1024, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%933, %560, %561, %562, %563), scope: ResNet/Sequential[layer3]/Bottleneck[22]/BatchNorm2d[bn3]\n  %935 : Float(32, 1024, 14, 14) = onnx::Add(%934, %926), scope: ResNet/Sequential[layer3]/Bottleneck[22]\n  %936 : Float(32, 1024, 14, 14) = onnx::Relu(%935), scope: ResNet/Sequential[layer3]/Bottleneck[22]/ReLU[relu]\n  %937 : Float(32, 512, 14, 14) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%936, %565), scope: ResNet/Sequential[layer4]/Bottleneck[0]/Conv2d[conv1]\n  %938 : Float(32, 512, 14, 14) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%937, %566, %567, %568, %569), scope: ResNet/Sequential[layer4]/Bottleneck[0]/BatchNorm2d[bn1]\n  %939 : Float(32, 512, 14, 14) = onnx::Relu(%938), scope: ResNet/Sequential[layer4]/Bottleneck[0]/ReLU[relu]\n  %940 : Float(32, 512, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[2, 2]](%939, %571), scope: ResNet/Sequential[layer4]/Bottleneck[0]/Conv2d[conv2]\n  %941 : Float(32, 512, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%940, %572, %573, %574, %575), scope: ResNet/Sequential[layer4]/Bottleneck[0]/BatchNorm2d[bn2]\n  %942 : Float(32, 512, 7, 7) = onnx::Relu(%941), scope: ResNet/Sequential[layer4]/Bottleneck[0]/ReLU[relu]\n  %943 : Float(32, 2048, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%942, %577), scope: ResNet/Sequential[layer4]/Bottleneck[0]/Conv2d[conv3]\n  %944 : Float(32, 2048, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%943, %578, %579, %580, %581), scope: ResNet/Sequential[layer4]/Bottleneck[0]/BatchNorm2d[bn3]\n  %945 : Float(32, 2048, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[2, 2]](%936, %583), scope: ResNet/Sequential[layer4]/Bottleneck[0]/Sequential[downsample]/Conv2d[0]\n  %946 : Float(32, 2048, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%945, %584, %585, %586, %587), scope: ResNet/Sequential[layer4]/Bottleneck[0]/Sequential[downsample]/BatchNorm2d[1]\n  %947 : Float(32, 2048, 7, 7) = onnx::Add(%944, %946), scope: ResNet/Sequential[layer4]/Bottleneck[0]\n  %948 : Float(32, 2048, 7, 7) = onnx::Relu(%947), scope: ResNet/Sequential[layer4]/Bottleneck[0]/ReLU[relu]\n  %949 : Float(32, 512, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%948, %589), scope: ResNet/Sequential[layer4]/Bottleneck[1]/Conv2d[conv1]\n  %950 : Float(32, 512, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%949, %590, %591, %592, %593), scope: ResNet/Sequential[layer4]/Bottleneck[1]/BatchNorm2d[bn1]\n  %951 : Float(32, 512, 7, 7) = onnx::Relu(%950), scope: ResNet/Sequential[layer4]/Bottleneck[1]/ReLU[relu]\n  %952 : Float(32, 512, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%951, %595), scope: ResNet/Sequential[layer4]/Bottleneck[1]/Conv2d[conv2]\n  %953 : Float(32, 512, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%952, %596, %597, %598, %599), scope: ResNet/Sequential[layer4]/Bottleneck[1]/BatchNorm2d[bn2]\n  %954 : Float(32, 512, 7, 7) = onnx::Relu(%953), scope: ResNet/Sequential[layer4]/Bottleneck[1]/ReLU[relu]\n  %955 : Float(32, 2048, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%954, %601), scope: ResNet/Sequential[layer4]/Bottleneck[1]/Conv2d[conv3]\n  %956 : Float(32, 2048, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%955, %602, %603, %604, %605), scope: ResNet/Sequential[layer4]/Bottleneck[1]/BatchNorm2d[bn3]\n  %957 : Float(32, 2048, 7, 7) = onnx::Add(%956, %948), scope: ResNet/Sequential[layer4]/Bottleneck[1]\n  %958 : Float(32, 2048, 7, 7) = onnx::Relu(%957), scope: ResNet/Sequential[layer4]/Bottleneck[1]/ReLU[relu]\n  %959 : Float(32, 512, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%958, %607), scope: ResNet/Sequential[layer4]/Bottleneck[2]/Conv2d[conv1]\n  %960 : Float(32, 512, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%959, %608, %609, %610, %611), scope: ResNet/Sequential[layer4]/Bottleneck[2]/BatchNorm2d[bn1]\n  %961 : Float(32, 512, 7, 7) = onnx::Relu(%960), scope: ResNet/Sequential[layer4]/Bottleneck[2]/ReLU[relu]\n  %962 : Float(32, 512, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[3, 3], pads=[1, 1, 1, 1], strides=[1, 1]](%961, %613), scope: ResNet/Sequential[layer4]/Bottleneck[2]/Conv2d[conv2]\n  %963 : Float(32, 512, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%962, %614, %615, %616, %617), scope: ResNet/Sequential[layer4]/Bottleneck[2]/BatchNorm2d[bn2]\n  %964 : Float(32, 512, 7, 7) = onnx::Relu(%963), scope: ResNet/Sequential[layer4]/Bottleneck[2]/ReLU[relu]\n  %965 : Float(32, 2048, 7, 7) = onnx::Conv[dilations=[1, 1], group=1, kernel_shape=[1, 1], pads=[0, 0, 0, 0], strides=[1, 1]](%964, %619), scope: ResNet/Sequential[layer4]/Bottleneck[2]/Conv2d[conv3]\n  %966 : Float(32, 2048, 7, 7) = onnx::BatchNormalization[epsilon=1e-05, momentum=1](%965, %620, %621, %622, %623), scope: ResNet/Sequential[layer4]/Bottleneck[2]/BatchNorm2d[bn3]\n  %967 : Float(32, 2048, 7, 7) = onnx::Add(%966, %958), scope: ResNet/Sequential[layer4]/Bottleneck[2]\n  %968 : Float(32, 2048, 7, 7) = onnx::Relu(%967), scope: ResNet/Sequential[layer4]/Bottleneck[2]/ReLU[relu]\n  %969 : Tensor = onnx::Pad[mode=\"constant\", pads=[0, 0, 0, 0, 0, 0, 0, 0], value=0](%968), scope: ResNet/AvgPool2d[avgpool]\n  %970 : Float(32, 2048, 1, 1) = onnx::AveragePool[kernel_shape=[7, 7], pads=[0, 0, 0, 0], strides=[1, 1]](%969), scope: ResNet/AvgPool2d[avgpool]\n  %971 : Long() = onnx::Constant[value={0}](), scope: ResNet\n  %972 : Tensor = onnx::Shape(%970), scope: ResNet\n  %973 : Long() = onnx::Gather[axis=0](%972, %971), scope: ResNet\n  %974 : Long() = onnx::Constant[value={-1}](), scope: ResNet\n  %975 : Tensor = onnx::Unsqueeze[axes=[0]](%973)\n  %976 : Tensor = onnx::Unsqueeze[axes=[0]](%974)\n  %977 : Tensor = onnx::Concat[axis=0](%975, %976)\n  %978 : Float(32, 2048) = onnx::Reshape(%970, %977), scope: ResNet\n  %logits_0 : Float(32, 1000) = onnx::Gemm[alpha=1, beta=1, transB=1](%978, %625, %626), scope: ResNet/Linear[fc]\n  return (%logits_0);\n}\n\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "model = onnx.load('resnet50.onnx')\nserve = onnx_backend.prepare(model, device='CUDA:1')\n\nx_np = x.detach().to('cpu').numpy()",
      "execution_count": 12,
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": "CUDA operators do not support 64-bit doubles, please use arr.astype(np.float32) or np.int32 for ints. Blob: input_0 type: float64\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "logits = serve.run(x_np).logits_0\nprint(np.argmax(logits[0]))",
      "execution_count": 15,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "243\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "%%timeit -n 10\nserve.run(x_np)",
      "execution_count": 17,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "93.4 ms ± 459 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
        }
      ]
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "gist": {
      "id": "",
      "data": {
        "description": "Pytorch ONNX.ipynb",
        "public": true
      }
    },
    "kernelspec": {
      "name": "rik_ssh_arte_yolo",
      "display_name": "SSH daniel@arte Python 3 [arte/yolo]",
      "language": "python"
    },
    "language_info": {
      "codemirror_mode": {
        "version": 3,
        "name": "ipython"
      },
      "mimetype": "text/x-python",
      "version": "3.5.5",
      "name": "python",
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "varInspector": {
      "window_display": false,
      "cols": {
        "lenName": 16,
        "lenType": 16,
        "lenVar": 40
      },
      "kernels_config": {
        "python": {
          "library": "var_list.py",
          "delete_cmd_prefix": "del ",
          "delete_cmd_postfix": "",
          "varRefreshCmd": "print(var_dic_list())"
        },
        "r": {
          "library": "var_list.r",
          "delete_cmd_prefix": "rm(",
          "delete_cmd_postfix": ") ",
          "varRefreshCmd": "cat(var_dic_list()) "
        }
      },
      "types_to_exclude": [
        "module",
        "function",
        "builtin_function_or_method",
        "instance",
        "_Feature"
      ]
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}