transformer-experiments-knockout.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:38.619584Z",
     "start_time": "2019-08-19T17:09:37.539914Z"
    }
   },
   "outputs": [],
   "source": [
    "#%load_ext autoreload\n",
    "#%autoreload 2\n",
    "import os\n",
    "#os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\"\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from pathlib import Path\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.parallel\n",
    "import torch.utils.data\n",
    "from torch.autograd import Variable\n",
    "import numpy as np\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:38.764093Z",
     "start_time": "2019-08-19T17:09:38.715052Z"
    }
   },
   "outputs": [],
   "source": [
    "from torch.nn import Parameter\n",
    "\n",
    "def _load_from_state_dict(self, state_dict, prefix, local_metadata, strict,\n",
    "                          missing_keys, unexpected_keys, error_msgs):\n",
    "    r\"\"\"Copies parameters and buffers from :attr:`state_dict` into only\n",
    "    this module, but not its descendants. This is called on every submodule\n",
    "    in :meth:`~torch.nn.Module.load_state_dict`. Metadata saved for this\n",
    "    module in input :attr:`state_dict` is provided as :attr:`local_metadata`.\n",
    "    For state dicts without metadata, :attr:`local_metadata` is empty.\n",
    "    Subclasses can achieve class-specific backward compatible loading using\n",
    "    the version number at `local_metadata.get(\"version\", None)`.\n",
    "\n",
    "    .. note::\n",
    "        :attr:`state_dict` is not the same object as the input\n",
    "        :attr:`state_dict` to :meth:`~torch.nn.Module.load_state_dict`. So\n",
    "        it can be modified.\n",
    "\n",
    "    Arguments:\n",
    "        state_dict (dict): a dict containing parameters and\n",
    "            persistent buffers.\n",
    "        prefix (str): the prefix for parameters and buffers used in this\n",
    "            module\n",
    "        local_metadata (dict): a dict containing the metadata for this module.\n",
    "            See\n",
    "        strict (bool): whether to strictly enforce that the keys in\n",
    "            :attr:`state_dict` with :attr:`prefix` match the names of\n",
    "            parameters and buffers in this module\n",
    "        missing_keys (list of str): if ``strict=True``, add missing keys to\n",
    "            this list\n",
    "        unexpected_keys (list of str): if ``strict=True``, add unexpected\n",
    "            keys to this list\n",
    "        error_msgs (list of str): error messages should be added to this\n",
    "            list, and will be reported together in\n",
    "            :meth:`~torch.nn.Module.load_state_dict`\n",
    "    \"\"\"\n",
    "    for hook in self._load_state_dict_pre_hooks.values():\n",
    "        hook(state_dict, prefix, local_metadata, strict, missing_keys,\n",
    "             unexpected_keys, error_msgs)\n",
    "\n",
    "    local_name_params = itertools.chain(self._parameters.items(),\n",
    "                                        self._buffers.items())\n",
    "    local_state = {k: v.data for k, v in local_name_params if v is not None}\n",
    "\n",
    "    for name, param in local_state.items():\n",
    "        key = prefix + name\n",
    "        if key in state_dict:\n",
    "            input_param = state_dict[key]\n",
    "\n",
    "            # Backward compatibility: loading 1-dim tensor from 0.3.* to version 0.4+\n",
    "            if len(param.shape) == 0 and len(input_param.shape) == 1:\n",
    "                input_param = input_param[0]\n",
    "\n",
    "            if input_param.shape != param.shape:\n",
    "                # local shape should match the one in checkpoint\n",
    "                error_msgs.append(\n",
    "                    'size mismatch for {}: copying a param with shape {} from checkpoint, '\n",
    "                    'the shape in current model is {}.'.format(\n",
    "                        key, input_param.shape, param.shape))\n",
    "                #if not strict:\n",
    "                #    continue\n",
    "\n",
    "            if isinstance(input_param, Parameter):\n",
    "                # backwards compatibility for serialized parameters\n",
    "                input_param = input_param.data\n",
    "\n",
    "            try:\n",
    "                if False:\n",
    "                    input_param = input_param[torch.randperm(input_param.size()[0])]\n",
    "                param.copy_(input_param)\n",
    "            except Exception:\n",
    "                error_msgs.append(\n",
    "                    'While copying the parameter named \"{}\", '\n",
    "                    'whose dimensions in the model are {} and '\n",
    "                    'whose dimensions in the checkpoint are {}.'.format(\n",
    "                        key, param.size(), input_param.size()))\n",
    "                # PG load partially\n",
    "\n",
    "                if len(input_param.size()) == 3:\n",
    "                    error_msgs.append(\n",
    "                        'Partially copying the parameter named \"{}\", '\n",
    "                        'whose dimensions in the model are {} and '\n",
    "                        'whose dimensions in the checkpoint are {}. - trying {}'\n",
    "                        .format(\n",
    "                            key, param.size(), input_param.size(),\n",
    "                            param[:input_param.size()[0], :input_param.size(\n",
    "                            )[1], :input_param.size()[2]].shape))\n",
    "                else:\n",
    "                    error_msgs.append(\n",
    "                        'Partially copying the parameter named \"{}\", '\n",
    "                        'whose dimensions in the model are {} and '\n",
    "                        'whose dimensions in the checkpoint are {}. - trying {}'\n",
    "                        .format(key, param.size(), input_param.size(),\n",
    "                                param[:input_param.size()[0]].shape))\n",
    "\n",
    "                try:\n",
    "                    new_input_param = torch.empty_like(param)\n",
    "                    new_input_param = torch.nn.init.normal_(new_input_param,\n",
    "                                                            mean=input_param.mean(),\n",
    "                                                            std=input_param.std())\n",
    "\n",
    "                    if len(input_param.size()) == 3:\n",
    "                        new_input_param[:input_param.size()[0], :input_param.\n",
    "                                        size()[1], :input_param.size(\n",
    "                                        )[2]] = input_param\n",
    "                    else:\n",
    "                        new_input_param[:input_param.size()[0]] = input_param\n",
    "                    param.copy_(new_input_param)\n",
    "                except Exception as e:\n",
    "                    assert e\n",
    "                    error_msgs.append(\n",
    "                        'Failed to load weights partially {}'.format(e))\n",
    "        elif strict:\n",
    "            missing_keys.append(key)\n",
    "\n",
    "    if strict:\n",
    "        for key in state_dict.keys():\n",
    "            if key.startswith(prefix):\n",
    "                input_name = key[len(prefix):]\n",
    "                input_name = input_name.split(\n",
    "                    '.', 1)[0]  # get the name of param/buffer/child\n",
    "                if input_name not in self._modules and input_name not in local_state:\n",
    "                    unexpected_keys.append(key)\n",
    "\n",
    "def load_state_dict(self, state_dict, strict=True):\n",
    "    r\"\"\"Copies parameters and buffers from :attr:`state_dict` into\n",
    "    this module and its descendants. If :attr:`strict` is ``True``, then\n",
    "    the keys of :attr:`state_dict` must exactly match the keys returned\n",
    "    by this module's :meth:`~torch.nn.Module.state_dict` function.\n",
    "\n",
    "    Arguments:\n",
    "        state_dict (dict): a dict containing parameters and\n",
    "            persistent buffers.\n",
    "        strict (bool, optional): whether to strictly enforce that the keys\n",
    "            in :attr:`state_dict` match the keys returned by this module's\n",
    "            :meth:`~torch.nn.Module.state_dict` function. Default: ``True``\n",
    "\n",
    "    Returns:\n",
    "        ``NamedTuple`` with ``missing_keys`` and ``unexpected_keys`` fields:\n",
    "            * **missing_keys** is a list of str containing the missing keys\n",
    "            * **unexpected_keys** is a list of str containing the unexpected keys\n",
    "    \"\"\"\n",
    "    missing_keys = []\n",
    "    unexpected_keys = []\n",
    "    error_msgs = []\n",
    "\n",
    "    # copy state_dict so _load_from_state_dict can modify it\n",
    "    metadata = getattr(state_dict, '_metadata', None)\n",
    "    state_dict = state_dict.copy()\n",
    "    if metadata is not None:\n",
    "        state_dict._metadata = metadata\n",
    "\n",
    "    def load(module, prefix=''):\n",
    "        local_metadata = {} if metadata is None else metadata.get(\n",
    "            prefix[:-1], {})\n",
    "        module._load_from_state_dict(state_dict, prefix, local_metadata, True,\n",
    "                                     missing_keys, unexpected_keys, error_msgs)\n",
    "        for name, child in module._modules.items():\n",
    "            if child is not None:\n",
    "                load(child, prefix + name + '.')\n",
    "                \n",
    "    load(self)\n",
    "\n",
    "    if strict:\n",
    "        if len(unexpected_keys) > 0:\n",
    "            error_msgs.insert(\n",
    "                0, 'Unexpected key(s) in state_dict: {}. '.format(', '.join(\n",
    "                    '\"{}\"'.format(k) for k in unexpected_keys)))\n",
    "        if len(missing_keys) > 0:\n",
    "            error_msgs.insert(\n",
    "                0, 'Missing key(s) in state_dict: {}. '.format(', '.join(\n",
    "                    '\"{}\"'.format(k) for k in missing_keys)))\n",
    "\n",
    "    if strict and len(error_msgs) > 0:\n",
    "        raise RuntimeError(\n",
    "            'Error(s) in loading state_dict for {}:\\n\\t{}'.format(\n",
    "                self.__class__.__name__, \"\\n\\t\".join(error_msgs)))\n",
    "    return None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:39.233169Z",
     "start_time": "2019-08-19T17:09:39.227819Z"
    }
   },
   "outputs": [],
   "source": [
    "nn.Module._load_from_state_dict = _load_from_state_dict\n",
    "nn.Module.load_state_dict = load_state_dict\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:41.060744Z",
     "start_time": "2019-08-19T17:09:39.851184Z"
    }
   },
   "outputs": [],
   "source": [
    "from fastai import *\n",
    "from fastai.basic_train import *\n",
    "from fastai.basic_data import *\n",
    "from fastai.data_block import *\n",
    "from fastai.torch_core import *\n",
    "from fastai.train import *\n",
    "from fastai.callback import *\n",
    "from fastai.callbacks import *\n",
    "from fastai.distributed import *\n",
    "from fastai.layers import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:41.101982Z",
     "start_time": "2019-08-19T17:09:41.063665Z"
    }
   },
   "outputs": [],
   "source": [
    "fname_ext = lambda fname,ext: f'{str(fname)[:-4]}{ext}{str(fname)[-4:]}'\n",
    "\n",
    "def preprocess(fname, type_index=None, ext=''):\n",
    "    t  = pd.read_csv(fname_ext(fname,ext))\n",
    "    s  = pd.read_csv('structures.csv')\n",
    "    \n",
    "    has_y = 'scalar_coupling_constant' in t.columns\n",
    "\n",
    "    if has_y:\n",
    "        # atom-atom level\n",
    "        # molecule_name,atom_index_0,atom_index_1,type,fc,sd,pso,dso\n",
    "        scalar_couplings = pd.read_csv(f'scalar_coupling_contributions{ext}.csv') # fc,sd,pso,dso\n",
    "\n",
    "        # atom level\n",
    "        # molecule_name,atom_index,XX,YX,ZX,XY,YY,ZY,XZ,YZ,ZZ\n",
    "        magnetic_shielding = pd.read_csv('magnetic_shielding_tensors.csv')\n",
    "        # molecule_name,atom_index,mulliken_charge\n",
    "        mulliken_charges = pd.read_csv('mulliken_charges.csv')\n",
    "\n",
    "        # molecule level\n",
    "        # molecule_name,X,Y,Z\n",
    "        dipole_moments = pd.read_csv('dipole_moments.csv')\n",
    "        # molecule_name,potential_energy\n",
    "        potential_energy = pd.read_csv('potential_energy.csv')\n",
    "\n",
    "    t['molecule_index'] = pd.factorize(t['molecule_name'])[0] + t['id'].min()\n",
    "    # make sure we use the same indexes in train/test (test needs to provide type_index)\n",
    "    if type_index is not None:\n",
    "        t['type_idx'] = t['type'].apply(lambda x: type_index.index(x) ) # pd.factorize(pd.concat([pd.Series(type_index),t['type']]))[0][len(type_index):]\n",
    "    else:\n",
    "        t['type_idx'] = pd.factorize(t['type'])[0]\n",
    "\n",
    "    s['atom_idx'] = s['atom'].apply(lambda x: atoms.index(x) )\n",
    "\n",
    "    max_items = len(t.groupby(['molecule_name', 'atom_index_0']))# if has_y else 422550\n",
    "    max_atoms = int(s.atom_index.max() + 1)\n",
    "\n",
    "    if has_y:\n",
    "        contributions = ['fc','sd','pso','dso']\n",
    "        magnetic_tensors = ['XX','YX','ZX','XY','YY','ZY','XZ','YZ','ZZ']\n",
    "        XYZ = ['X','Y','Z']\n",
    "    xyz = ['x', 'y', 'z']\n",
    "    \n",
    "    x_xyz   = np.zeros((max_items,len(xyz),  max_atoms), dtype=np.float32)\n",
    "    x_type  = np.zeros((max_items,1,         max_atoms), dtype=np.int8)\n",
    "    x_ext   = np.zeros((max_items,1,         max_atoms), dtype=np.bool_)\n",
    "    x_atom  = np.empty((max_items,1,         max_atoms), dtype=np.int8)\n",
    "    x_atom[:] = -1\n",
    "\n",
    "    if has_y:\n",
    "        y_scalar   = np.zeros((max_items,len(contributions)   ,max_atoms), dtype=np.float32)\n",
    "        y_magnetic = np.zeros((max_items,len(magnetic_tensors),max_atoms), dtype=np.float32)\n",
    "        y_mulliken = np.zeros((max_items,1                    ,max_atoms), dtype=np.float32)\n",
    "\n",
    "        y_dipole   = np.zeros((max_items,len(XYZ)), dtype=np.float32)\n",
    "        y_potential= np.zeros((max_items,1       ), dtype=np.float32)\n",
    "\n",
    "        y_magnetic[...] = np.nan\n",
    "        y_mulliken[...] = np.nan\n",
    "    else:\n",
    "        xt_ids = np.zeros((max_items, max_atoms), dtype=np.int32)\n",
    "\n",
    "    m = np.zeros((max_items,), dtype=np.int32)\n",
    "    i = j = 0\n",
    "    \n",
    "    for (m_name, m_index) ,m_group in tqdm(t.groupby(['molecule_name', 'molecule_index'])):\n",
    "        ss = s[s.molecule_name==m_name]\n",
    "        n_atoms = len(ss)\n",
    "        if has_y:\n",
    "            magnetic = magnetic_shielding[\n",
    "                    (magnetic_shielding['molecule_name']==m_name)][magnetic_tensors].values.T\n",
    "\n",
    "            mulliken = mulliken_charges[\n",
    "                    (mulliken_charges['molecule_name']==m_name)]['mulliken_charge'].values.T\n",
    "\n",
    "            scs = scalar_couplings[scalar_couplings['molecule_name']==m_name]\n",
    "            \n",
    "            y_dipole[j,:]= dipole_moments[dipole_moments['molecule_name']==m_name][XYZ].values\n",
    "            y_potential[j,:]=potential_energy[\n",
    "                potential_energy['molecule_name']==m_name]['potential_energy'].values\n",
    "        \n",
    "        for a_name,a_group in m_group.groupby('atom_index_0'):\n",
    "            \n",
    "            ref_a = ss[ss['atom_index']==a_name]\n",
    "            \n",
    "            x_xyz[i] = 0.\n",
    "            x_type[i] = -1\n",
    "            x_ext[i] =  True\n",
    "            \n",
    "            x_xyz[i,:,:n_atoms] = (ss[xyz].values-ref_a[xyz].values).T  # xyz \n",
    "            x_type[i,0,a_group['atom_index_1']] = a_group['type_idx']  # type \n",
    "            x_ext[i,0,a_group['atom_index_1']] = a_group['ext']  # ext \n",
    "            x_atom[i,:,:n_atoms] = ss['atom_idx'].T                \n",
    "\n",
    "            if has_y:\n",
    "                y_scalar[i,:,a_group['atom_index_1']] = scs[scs['atom_index_0']==a_name][contributions]\n",
    "                y_magnetic[i,:,:n_atoms] = magnetic\n",
    "                y_mulliken[i,:,:n_atoms] = mulliken\n",
    "            else:\n",
    "                xt_ids[i,a_group['atom_index_1']] = a_group['id']  \n",
    "\n",
    "            m[i] = m_index\n",
    "            i+=1\n",
    "        j += 1\n",
    "    assert i == max_items\n",
    "    print(i,max_items)\n",
    "    if has_y:\n",
    "        return x_xyz,x_type,x_ext,x_atom, m, y_scalar, y_magnetic, y_mulliken, y_dipole, y_potential\n",
    "    else:\n",
    "        return x_xyz,x_type,x_ext,x_atom, m, xt_ids"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define where you want to use original training set '' or extended ''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:41.692190Z",
     "start_time": "2019-08-19T17:09:41.688409Z"
    }
   },
   "outputs": [],
   "source": [
    "ext = '_ext' # or ''"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load preprocessed or preprocess and save for later"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:48.324228Z",
     "start_time": "2019-08-19T17:09:42.786721Z"
    }
   },
   "outputs": [],
   "source": [
    "train_fname = Path('train.npz')\n",
    "types = ['1JHC', '2JHH', '1JHN', '2JHN', '2JHC', '3JHH', '3JHC', '3JHN'] \n",
    "atoms = 'CFHNO'\n",
    "\n",
    "try:\n",
    "    npzfile = np.load(fname_ext(train_fname, ext))\n",
    "    x_xyz   = npzfile['x_xyz']\n",
    "    x_type  = npzfile['x_type']\n",
    "    x_ext   = npzfile['x_ext']\n",
    "    x_atom  = npzfile['x_atom']\n",
    "\n",
    "    y_scalar    = npzfile['y_scalar']\n",
    "    y_magnetic  = npzfile['y_magnetic']\n",
    "    y_mulliken  = npzfile['y_mulliken']\n",
    "    y_dipole    = npzfile['y_dipole']\n",
    "    y_potential = npzfile['y_potential']\n",
    "    m = npzfile['m']\n",
    "    max_items, max_atoms = x_xyz.shape[0], x_xyz.shape[-1]\n",
    "except:\n",
    "    x_xyz,x_type,x_ext,x_atom, m, y_scalar, y_magnetic, y_mulliken, y_dipole, y_potential = \\\n",
    "        preprocess(train_fname.with_suffix('.csv'), type_index=types, ext=ext)\n",
    "    np.savez(fname_ext(train_fname, ext), \n",
    "             x_xyz=x_xyz,\n",
    "             x_type=x_type,\n",
    "             x_ext=x_ext,\n",
    "             x_atom=x_atom,\n",
    "             y_scalar=y_scalar,\n",
    "             y_magnetic=y_magnetic,\n",
    "             y_mulliken=y_mulliken,\n",
    "             y_dipole=y_dipole,\n",
    "             y_potential=y_potential,\n",
    "             m=m)\n",
    "n_types = int(x_type[~np.isnan(x_type)].max() + 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:48.332322Z",
     "start_time": "2019-08-19T17:09:48.326293Z"
    }
   },
   "outputs": [],
   "source": [
    "use_memmap = True\n",
    "try:\n",
    "    load_fn = np.load if not use_memmap else partial(np.lib.format.open_memmap, mode='r')\n",
    "    x_coulombmat = load_fn(f'x_coulombmat32{ext}.npy')\n",
    "except:\n",
    "    x_coulombmat = np.load(f'x_coulombmat{ext}.npy', allow_pickle=True)\n",
    "    x_coulombmat = np.array(x_coulombmat.tolist()).astype(np.float32)\n",
    "    np.save(f'x_coulombmat32{ext}.npy', x_coulombmat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:48.338202Z",
     "start_time": "2019-08-19T17:09:48.334641Z"
    }
   },
   "outputs": [],
   "source": [
    "x_qm9_mulliken = load_fn(f'x_qm9_mulliken{ext}.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:48.346751Z",
     "start_time": "2019-08-19T17:09:48.340324Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(1405126, 3, 29),\n",
       " (1405126, 1, 29),\n",
       " (1405126, 1, 29),\n",
       " (1405126, 1, 29),\n",
       " (1405126, 1, 29),\n",
       " (1405126, 4, 29),\n",
       " (1405126, 9, 29),\n",
       " (1405126, 1, 29),\n",
       " (1405126, 3),\n",
       " (1405126, 1),\n",
       " (1405126,)]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[v.shape for v in [x_xyz,x_type,x_ext,x_atom,x_qm9_mulliken, \n",
    "                   y_scalar, y_magnetic, y_mulliken, y_dipole, y_potential, m]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.367867Z",
     "start_time": "2019-08-19T17:09:48.348694Z"
    }
   },
   "outputs": [],
   "source": [
    "x_xyz_mean, x_xyz_std = Tensor(x_xyz.mean(axis=(0,2),keepdims=True)), Tensor(x_xyz.std(axis=(0,2),keepdims=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.647878Z",
     "start_time": "2019-08-19T17:09:49.369847Z"
    }
   },
   "outputs": [],
   "source": [
    "x_qm9_mulliken_mean = Tensor(x_qm9_mulliken.mean(axis=(0,2),keepdims=True))\n",
    "x_qm9_mulliken_std  = Tensor(x_qm9_mulliken.std( axis=(0,2),keepdims=True))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.653815Z",
     "start_time": "2019-08-19T17:09:49.650096Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 3, 1])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_xyz_std.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Fastai classes (this should should be done into its own `application` but who has time?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.673551Z",
     "start_time": "2019-08-19T17:09:49.657553Z"
    }
   },
   "outputs": [],
   "source": [
    "class MoleculeItem(ItemBase):\n",
    "    def __init__(self,i,xyz,type,ext,atom,qm9_mulliken,coulomb): \n",
    "        self.i,self.xyz,self.type,self.ext, self.atom,self.qm9_mulliken,self.coulomb = \\\n",
    "            i,xyz,type,ext,atom,qm9_mulliken,coulomb\n",
    "        self.data = [Tensor(xyz), LongTensor((type)), \n",
    "                     Tensor(ext), LongTensor((atom)),Tensor(qm9_mulliken), Tensor(coulomb)]\n",
    "    def __str__(self):\n",
    "        # TODO: count n_atoms correctly. \n",
    "        n_atoms = np.count_nonzero(np.sum(np.absolute(self.xyz), axis=0))+1\n",
    "        n_couplings = np.sum((self.type!=-1))\n",
    "        return f'{self.i} {n_atoms} atoms {n_couplings} couplings'\n",
    "    \n",
    "    def apply_tfms(self, tfms:Collection, **kwargs):\n",
    "        x = self.clone()\n",
    "        for t in tfms:\n",
    "            if t: x.data = t(x.data)\n",
    "        return x\n",
    "    \n",
    "    def clone(self):\n",
    "        return self.__class__(self.i,self.xyz,self.type,self.ext,self.atom,self.qm9_mulliken,self.coulomb)\n",
    "    \n",
    "class ScalarCouplingItem(ItemBase):\n",
    "    def __init__(self,scalar,magnetic,mulliken,dipole,potential,**kwargs): \n",
    "        self.scalar,self.magnetic,self.mulliken,self.dipole,self.potential = \\\n",
    "            scalar,magnetic,mulliken,dipole,potential\n",
    "        self.data = (Tensor(scalar), Tensor(magnetic), Tensor(dipole), Tensor(potential))\n",
    "    def __str__(self):\n",
    "        res, spacer, n_couplings = '', '', 0\n",
    "        for s in self.data[0].sum(dim=0):\n",
    "            if s==0.: spacer = ' * '\n",
    "            else: \n",
    "                res += f'{spacer}{s:.4f}'\n",
    "                spacer = ' '\n",
    "                n_couplings +=1\n",
    "        return f'{n_couplings}: {res}'\n",
    "    def __hash__(self): return hash(str(self))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.691659Z",
     "start_time": "2019-08-19T17:09:49.675862Z"
    }
   },
   "outputs": [],
   "source": [
    "class LMAEMaskedLoss(Module):\n",
    "    def __init__(self,\n",
    "                 contrib_w=0., magnetic_w=0., dipole_w=0., potential_w=0., \n",
    "                 types_w = [1]*n_types, return_all=False, proxy_log=torch.log, exclude_ext=False):\n",
    "        self.contrib_w,self.magnetic_w,self.dipole_w,self.potential_w = contrib_w,magnetic_w,dipole_w,potential_w\n",
    "        self.types_w = types_w\n",
    "        self.return_all = return_all\n",
    "        self.proxy_log = proxy_log\n",
    "        self.exclude_ext = exclude_ext\n",
    "    \n",
    "    def forward(self, input_outputs, t_scalar, t_magnetic, t_dipole, t_potential):    \n",
    "        type, ext, p_scalar, p_magnetic, p_dipole, p_potential = input_outputs\n",
    "        loss = 0.\n",
    "        n = 0\n",
    "        j_loss = [0] * n_types\n",
    "        for t in range(n_types):\n",
    "            mask = (type == t).squeeze(1) if not self.exclude_ext else ((type == t) & (ext == 0)).squeeze(1)\n",
    "            if mask.sum() > 0:\n",
    "                _output,_target = p_scalar.transpose(1,2)[mask], t_scalar.transpose(1,2)[mask] # scalars at the end\n",
    "                # LMAE scalar\n",
    "                s_loss = self.proxy_log((_output.sum(dim=-1) - _target.sum(dim=-1)).abs().mean()+1e-9)\n",
    "                loss += self.types_w[t] * s_loss\n",
    "                j_loss[t] += s_loss\n",
    "                # LMAE scalar contributions\n",
    "                #for i_contrib in range(_output.shape[-1]):\n",
    "                #    loss += self.contrib_w * \\\n",
    "                #        self.proxy_log((_output[...,i_contrib] - _target[...,i_contrib]).abs().mean()+1e-9)\n",
    "                n+=1\n",
    "        loss /= n\n",
    "        \n",
    "        if self.magnetic_w > 0:\n",
    "            mask = ~torch.isnan(t_magnetic)\n",
    "            loss += self.magnetic_w * MSELossFlat()(p_magnetic[mask], t_magnetic[mask])\n",
    "            \n",
    "        if self.dipole_w    > 0: loss += self.dipole_w    * MSELossFlat()(p_dipole,    t_dipole)\n",
    "        if self.potential_w > 0: loss += self.potential_w * MSELossFlat()(p_potential, t_potential)\n",
    "\n",
    "        return loss if not self.return_all else (loss, *j_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.698750Z",
     "start_time": "2019-08-19T17:09:49.693805Z"
    }
   },
   "outputs": [],
   "source": [
    "class ScalarCouplingList(ItemList):\n",
    "    def __init__(self, items:Iterator, **kwargs):\n",
    "        super().__init__(items, **kwargs)\n",
    "        self.loss_func = LMAEMaskedLoss\n",
    "\n",
    "    def get(self, i):\n",
    "        o = super().get(i)\n",
    "        return ScalarCouplingItem(*o)\n",
    "\n",
    "    def reconstruct(self,t): return 0; # TODO for viz !!!! ScalarCouplingItem(t.numpy())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quaterions allow us to rotate 3d points randoming with a nice uniform distribution of 3 numbers hece we use them, however it's still to be seen if are useful here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:09:49.733919Z",
     "start_time": "2019-08-19T17:09:49.700715Z"
    }
   },
   "outputs": [],
   "source": [
    "#https://github.com/facebookresearch/QuaterNet/blob/master/common/quaternion.py\n",
    "def qrot(q, v):\n",
    "    \"\"\"\n",
    "    Rotate vector(s) v about the rotation described by quaternion(s) q.\n",
    "    Expects a tensor of shape (*, 4) for q and a tensor of shape (*, 3) for v,\n",
    "    where * denotes any number of dimensions.\n",
    "    Returns a tensor of shape (*, 3).\n",
    "    \"\"\"\n",
    "    assert q.shape[-1] == 4\n",
    "    assert v.shape[-1] == 3\n",
    "    assert q.shape[:-1] == v.shape[:-1]\n",
    "    \n",
    "    original_shape = list(v.shape)\n",
    "    q = q.view(-1, 4)\n",
    "    v = v.view(-1, 3)\n",
    "    \n",
    "    qvec = q[:, 1:]\n",
    "    uv = torch.cross(qvec, v, dim=1)\n",
    "    uuv = torch.cross(qvec, uv, dim=1)\n",
    "    return (v + 2 * (q[:, :1] * uv + uuv)).view(original_shape)\n",
    "\n",
    "def random_rotation(data):\n",
    "    x_xyz = data[0].transpose(0,1)\n",
    "    r = torch.rand(3)\n",
    "    sq1_v1,sqv1,v2_2pi,v3_2pi = torch.sqrt(1-r[:1]),torch.sqrt(r[:1]),2*math.pi*r[1:2],2*math.pi*r[2:3]\n",
    "    q = torch.cat([sq1_v1*torch.sin(v2_2pi), sq1_v1*torch.cos(v2_2pi), \n",
    "                   sqv1  *torch.sin(v3_2pi), sqv1  *torch.cos(v3_2pi)], dim=0).unsqueeze(0)\n",
    "    x_xyz = qrot(q.expand(x_xyz.shape[0],-1), x_xyz).squeeze(0).transpose(0,1)\n",
    "    return (x_xyz, *data[1:])\n",
    "\n",
    "def normalize(data):\n",
    "    sq = False\n",
    "    if data[0].ndim < 3:\n",
    "        data[0].unsqueeze_(0)\n",
    "        data[4].unsqueeze_(0)\n",
    "        sq = True\n",
    "    x_xyz      = (data[0] - x_xyz_mean)          / x_xyz_std\n",
    "    x_mulliken = (data[4] - x_qm9_mulliken_mean) / x_qm9_mulliken_std\n",
    "    if sq:\n",
    "        x_xyz.squeeze_(0)\n",
    "        x_mulliken.squeeze_(0)\n",
    "    return (x_xyz, data[1],data[2],data[3],x_mulliken,data[5])\n",
    "\n",
    "def canonize_(data):\n",
    "    xyz,type,ext,atom,mulliken,coulomb = data\n",
    "    mask = (atom == -1).squeeze(0)\n",
    "    mask_atoms = ~mask.unsqueeze(0)\n",
    "    n_atoms = mask_atoms.sum()\n",
    "    i = torch.nonzero(type.squeeze(0) == -1)[0] # pick first one w/o j-coupling\n",
    "    xyz[:,mask], type[:,mask],ext[:,mask],atom[:,mask],mulliken[:,mask] = 0,-1,1,-1,0\n",
    "    return (xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms)\n",
    "\n",
    "def canonize(data):\n",
    "    xyz,type,ext,atom,mulliken,coulomb = data\n",
    "    mask = (atom == -1).squeeze(0)\n",
    "    i_max_atom = torch.nonzero(atom != -1).max() + 1\n",
    "    mask_atoms = torch.ones ((max_atoms,  max_atoms), dtype=torch.uint8)\n",
    "    zeros      = torch.zeros((i_max_atom,i_max_atom), dtype=torch.uint8)\n",
    "    mask_atoms[:zeros.shape[0],:zeros.shape[1]] = zeros\n",
    "    n_atoms = i_max_atom\n",
    "    xyz[:,mask], type[:,mask],ext[:,mask],atom[:,mask],mulliken[:,mask] = 0,-1,1,-1,0\n",
    "    return (xyz,type,ext,atom,mulliken, coulomb, mask_atoms, n_atoms)\n",
    "\n",
    "def knockout(data):\n",
    "    xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms = data\n",
    "    atom_to_knockout = random.randint(0, n_atoms-1)\n",
    "    mask_atoms[atom_to_knockout,:] = 1\n",
    "    return xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms\n",
    "\n",
    "def canonize2(data):\n",
    "    xyz,type,ext,atom,mulliken = data\n",
    "    mask = (atom == -1).squeeze(0)\n",
    "    i_max_atom = torch.nonzero(atom != -1).max() + 1\n",
    "    #mask_atoms = torch.ones((max_atoms,   max_atoms), dtype=torch.uint8)\n",
    "    #zeros      = torch.zeros((i_max_atom, i_max_atom), dtype=torch.uint8)\n",
    "    #mask_atoms[:zeros.shape[0],:zeros.shape[1]] = zeros\n",
    "\n",
    "    mask_atoms = torch.ones ((max_atoms, ), dtype=torch.uint8)\n",
    "    zeros      = torch.zeros((i_max_atom,), dtype=torch.uint8)\n",
    "    mask_atoms[:zeros.shape[0],] = zeros\n",
    "    #mask_atoms[(type==-1).squeeze(0),:] = 1\n",
    "    n_atoms = i_max_atom\n",
    "    xyz[:,mask], type[:,mask],ext[:,mask],atom[:,mask],mulliken[:,mask] = 0,-1,1,-1,0\n",
    "    return (xyz,type,ext,atom,mulliken, mask_atoms, n_atoms)\n",
    "\n",
    "def canonize_and_shuffle(data):\n",
    "    xyz,type,ext,atom,mulliken,coulomb = data\n",
    "    mask = (atom == -1).squeeze(0)\n",
    "    mask_atoms = ~mask.unsqueeze(0)\n",
    "    n_atoms = mask_atoms.sum()\n",
    "    i = torch.nonzero(type.squeeze(0) == -1)[0] # pick first one w/o j-coupling\n",
    "    xyz[:,mask], type[:,mask],ext[:,mask],atom[:,mask],mulliken[:,mask] = \\\n",
    "        0,-1,1,-1,0\n",
    "    \n",
    "    perm = torch.randperm(n_atoms)\n",
    "    xyz[:,~mask] = xyz[:,~mask][:,perm]\n",
    "    type[:,~mask] = type[:,~mask][:,perm]\n",
    "    ext[:,~mask] = ext[:,~mask][:,perm]\n",
    "    atom[:,~mask] = atom[:,~mask][:,perm]\n",
    "    mulliken[:,~mask] = mulliken[:,~mask][:,perm]\n",
    "    \n",
    "    return (xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "random.randint(0, 29-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Build `data` bunch etc. for fastai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:18.654986Z",
     "start_time": "2019-08-19T17:09:49.735995Z"
    }
   },
   "outputs": [],
   "source": [
    "data = ItemList(items=(MoleculeItem(i,*v) for i,v in \n",
    "                       enumerate(zip(x_xyz,x_type,x_ext,x_atom,x_qm9_mulliken,x_coulombmat))),\n",
    "                label_cls=ScalarCouplingItem)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:19.035949Z",
     "start_time": "2019-08-19T17:11:18.657069Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "_, idx_valid_split = train_test_split(range(m.max()+1), test_size=0.1, random_state=13)\n",
    "idx_valid_split = np.argwhere(np.isin(m, idx_valid_split)).flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:19.162729Z",
     "start_time": "2019-08-19T17:11:19.038706Z"
    },
    "pixiedust": {
     "displayParams": {}
    }
   },
   "outputs": [],
   "source": [
    "data = data.split_by_idx(idx_valid_split)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:22.994242Z",
     "start_time": "2019-08-19T17:11:19.164755Z"
    }
   },
   "outputs": [],
   "source": [
    "data = data.label_from_func(\n",
    "    func=lambda o: (y_scalar[o.i], y_magnetic[o.i], y_mulliken[o.i], y_dipole[o.i], y_potential[o.i]),\n",
    "    label_cls=ScalarCouplingList)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.000636Z",
     "start_time": "2019-08-19T17:11:22.996375Z"
    }
   },
   "outputs": [],
   "source": [
    "if False:\n",
    "    try:\n",
    "        xt_coulombmat = load_fn(f'xt_coulombmat32{ext}.npy')\n",
    "    except:\n",
    "        xt_coulombmat = np.load(f'xt_coulombmat{ext}.npy', allow_pickle=True)\n",
    "        xt_coulombmat = np.array(xt_coulombmat.tolist()).astype(np.float32)\n",
    "        np.save(f'xt_coulombmat32{ext}.npy', xt_coulombmat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.010259Z",
     "start_time": "2019-08-19T17:11:23.002572Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[<function normalize at 0x7f2ff958c7b8>, <function canonize at 0x7f2ff958c950>, <function knockout at 0x7f2ff958c9d8>] [<function normalize at 0x7f2ff958c7b8>, <function canonize at 0x7f2ff958c950>]\n"
     ]
    }
   ],
   "source": [
    "tfms = [normalize, canonize]\n",
    "tta_tfms = list(tfms)\n",
    "#tta_tfms.insert(0,random_rotation)\n",
    "tta_tfms.append(knockout)\n",
    "#tta_tfms[-1] = canonize_and_shuffle\n",
    "print(tta_tfms, tfms)\n",
    "data = data.transform((tta_tfms, tfms))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.143873Z",
     "start_time": "2019-08-19T17:11:23.012470Z"
    }
   },
   "outputs": [],
   "source": [
    "data=data.databunch()#num_workers=4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LabelList (78217 items)\n",
       "x: ItemList\n",
       "16 8 atoms 7 couplings,17 8 atoms 7 couplings,18 8 atoms 7 couplings,19 8 atoms 7 couplings,20 8 atoms 7 couplings\n",
       "y: ScalarCouplingList\n",
       "7: 83.5430 -2.3783 * -11.7004 -11.6979 3.2528 13.6913 3.2521,7: 83.5417 -2.3786 -11.7004 * -11.6996 13.6924 3.2525 3.2527,7: 83.5484 -2.3772 -11.6979 -11.6996 * 3.2524 3.2524 13.6921,7: -2.3788 83.5418 3.2528 13.6924 3.2524 * -11.7004 -11.6993,7: -2.3785 83.5430 13.6913 3.2525 3.2524 -11.7004 * -11.6976\n",
       "Path: ."
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.train_ds.filter_by_func(lambda item, _: len(item.data[2].cpu().numpy()[item.data[2].cpu().numpy()==0]) == 0)\n",
    "data.valid_ds.filter_by_func(lambda item, _: len(item.data[2].cpu().numpy()[item.data[2].cpu().numpy()==0]) == 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model\n",
    "Whole model here, self-contained (needs some cleanup)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.162815Z",
     "start_time": "2019-08-19T17:11:23.146818Z"
    }
   },
   "outputs": [],
   "source": [
    "class LMAEMetric(LearnerCallback):\n",
    "    _order=-20 # Needs to run before the recorder\n",
    "    def __init__(self, learn, val_only=True):\n",
    "        super().__init__(learn)\n",
    "        self.val_only=val_only\n",
    "        self.metric = LMAEMaskedLoss(return_all=True, exclude_ext=True)\n",
    "\n",
    "    def on_train_begin(self, **kwargs):\n",
    "        if not self.val_only: self.learn.recorder.add_metric_names(['tLMAE'])\n",
    "        self.learn.recorder.add_metric_names(['👉🏻LMAE👈🏻'] + [f'lmae{i}' for i in range(n_types)])\n",
    "            \n",
    "    def on_batch_end(self, train, last_output, last_target, **kwargs):\n",
    "        if self.val_only and train: return \n",
    "        preds,targs = self.preds[int(train)], self.targs[int(train)] # 0 val 1 train\n",
    "        if preds is None:\n",
    "            targs, preds = listify(last_target), listify(last_output)\n",
    "            targs,preds = [t.detach() for t in targs],[t.detach() for t in preds]\n",
    "        else:\n",
    "            for i,(o,t) in enumerate(zip(last_output, last_target)):\n",
    "                preds[i] = torch.cat([preds[i], o.detach()], dim=0)\n",
    "                targs[i] = torch.cat([targs[i], t.detach()], dim=0)\n",
    "        self.preds[int(train)], self.targs[int(train)] = preds,targs\n",
    "        \n",
    "    def on_epoch_begin(self, **kwargs):\n",
    "        self.targs, self.preds = [None, None], [None, None]\n",
    "\n",
    "    def on_epoch_end(self, last_metrics, **kwargs):\n",
    "        mets = []\n",
    "        if self.preds[1]: mets.append(self.metric.forward(self.preds[1], *self.targs[1])[0]) # just tLMAE\n",
    "        if self.preds[0]: mets.extend(self.metric.forward(self.preds[0], *self.targs[0]))\n",
    "        return add_metrics(last_metrics, mets) if mets else None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.240391Z",
     "start_time": "2019-08-19T17:11:23.166832Z"
    }
   },
   "outputs": [],
   "source": [
    "Activation = Enum('Activation', 'ReLU Swish GeLU')\n",
    "\n",
    "class PositionalEncoding(Module):\n",
    "    \"Implement the PE function.\"\n",
    "    def __init__(self, d_model, dropout, max_len=5000):\n",
    "        super(PositionalEncoding, self).__init__()\n",
    "        self.dropout = nn.Dropout(p=dropout)\n",
    "        \n",
    "        # Compute the positional encodings once in log space.\n",
    "        pe = torch.zeros(max_len, d_model)\n",
    "        \n",
    "        position = torch.arange(0., max_len).unsqueeze(1)\n",
    "        div_term = torch.exp(torch.arange(0., d_model, 2) * -(math.log(10000.0) / d_model))\n",
    "\n",
    "        pe[:, 0::2] = torch.sin(position * div_term)\n",
    "        pe[:, 1::2] = torch.cos(position * div_term)[:, :-1]\n",
    "        pe = pe.unsqueeze(0)\n",
    "        self.register_buffer('pe', pe)\n",
    "        print(pe.size())\n",
    "        \n",
    "    def forward(self, x):\n",
    "        x = x.transpose(2, 1)\n",
    "        x = x + Variable(self.pe[:, :x.size(1)], \n",
    "                         requires_grad=False)\n",
    "        return self.dropout(x).transpose(2, 1)\n",
    "    \n",
    "class GeLU(Module):\n",
    "    def forward(self, x): return 0.5 * x * (1 + torch.tanh(math.sqrt(2 / math.pi) * (x + 0.044715 * torch.pow(x, 3))))\n",
    "\n",
    "class Swish(Module):\n",
    "    def forward(self, x): return x * torch.sigmoid(x)\n",
    "\n",
    "_activ_func = {Activation.ReLU:nn.ReLU(inplace=True), Activation.GeLU:GeLU(), Activation.Swish: Swish()}\n",
    "\n",
    "def feed_forward(d_model:int, d_ff:int, ff_p:float=0., act:Activation=Activation.ReLU, double_drop:bool=True):\n",
    "    layers = [nn.Linear(d_model, d_ff), _activ_func[act]]\n",
    "    if double_drop: layers.append(nn.Dropout(ff_p))\n",
    "    return SequentialEx(*layers, nn.Linear(d_ff, d_model), nn.Dropout(ff_p), MergeLayer(), nn.LayerNorm(d_model))\n",
    "\n",
    "class MultiHeadAttention(Module):\n",
    "    \"MutiHeadAttention.\"\n",
    "    def __init__(self, n_heads:int, d_model:int, d_head:int=None, resid_p:float=0., attn_p:float=0., bias:bool=False,\n",
    "                 scale:bool=True):\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.n_heads,self.d_head,self.scale = n_heads,d_head,scale\n",
    "\n",
    "        self.attention1 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention2 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention3 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention4 = nn.Linear(d_model, d_model, bias=bias)\n",
    "        self.out = nn.Linear(n_heads * d_head, d_model, bias=bias)\n",
    "        self.drop_att,self.drop_res = nn.Dropout(attn_p),nn.Dropout(resid_p)\n",
    "        self.ln = nn.LayerNorm(d_model)\n",
    "\n",
    "    def forward(self, x:Tensor, mask:Tensor=None, **kwargs):\n",
    "        return self.ln(x + self.drop_res(self.out(self._apply_attention(x, mask=mask, **kwargs))))\n",
    "\n",
    "    def _apply_attention(self, x:Tensor, mask:Tensor=None):\n",
    "        if False:\n",
    "            if mask is not None:\n",
    "                x= x.masked_fill(mask[:,:,0].squeeze(1).unsqueeze(-1).repeat(1, 1, x.size(2)),  0.)\n",
    "        bs,x_len = x.size(0),x.size(1)\n",
    "        wq,wk,wv = self.attention1(x), self.attention2(x), self.attention3(x)\n",
    "        wq,wk,wv = map(lambda x:x.view(bs, x.size(1), self.n_heads, self.d_head), (wq,wk,wv))\n",
    "        wq,wk,wv = wq.permute(0, 2, 1, 3),wk.permute(0, 2, 3, 1),wv.permute(0, 2, 1, 3)\n",
    "        attn_score = torch.matmul(wq, wk)\n",
    "        if self.scale: attn_score.div_(self.d_head ** 0.5)\n",
    "        if mask is not None:\n",
    "            attn_score = attn_score.float().masked_fill(mask,  -1.0E9).type_as(attn_score)\n",
    "        attn_prob = self.drop_att(F.softmax(attn_score, dim=-1))\n",
    "        attn_vec = torch.matmul(attn_prob, wv)\n",
    "        return self.attention4(attn_vec.permute(0, 2, 1, 3).contiguous().contiguous().view(bs, x_len, -1))\n",
    "\n",
    "    def _attention_einsum(self, x, mask=None):\n",
    "        # Permute and matmul is a little bit faster but this implementation is more readable\n",
    "        bs,x_len = x.size(0),x.size(1)\n",
    "        wq,wk,wv = torch.chunk(self.attention(x), 3, dim=-1)\n",
    "        wq,wk,wv = map(lambda x:x.view(bs, x.size(1), self.n_heads, self.d_head), (wq,wk,wv))\n",
    "        attn_score = torch.einsum('bind,bjnd->bijn', (wq, wk))\n",
    "        if self.scale: attn_score.mul_(1/(self.d_head ** 0.5))\n",
    "        if mask is not None:\n",
    "            attn_score = attn_score.float().masked_fill(mask, -float('-inf')).type_as(attn_score)\n",
    "        attn_prob = self.drop_att(F.softmax(attn_score, dim=2))\n",
    "        attn_vec = torch.einsum('bijn,bjnd->bind', (attn_prob, wv))\n",
    "        return attn_vec.contiguous().view(bs, x_len, -1)\n",
    "\n",
    "\n",
    "class DecoderLayer(Module):\n",
    "    \"Basic block of a Transformer model.\"\n",
    "    #Can't use Sequential directly cause more than one input...\n",
    "    def __init__(self, n_heads:int, d_model:int, d_head:int, d_inner:int, resid_p:float=0., attn_p:float=0., ff_p:float=0.,\n",
    "                 bias:bool=True, scale:bool=True, act:Activation=Activation.ReLU, double_drop:bool=True,\n",
    "                 attn_cls:Callable=MultiHeadAttention):\n",
    "        self.mhra = attn_cls(n_heads, d_model, d_head, resid_p=resid_p, attn_p=attn_p, bias=bias, scale=scale)\n",
    "        self.ff   = feed_forward(d_model, d_inner, ff_p=ff_p, act=act, double_drop=double_drop)\n",
    "\n",
    "    def forward(self, x:Tensor, mask:Tensor=None, **kwargs): return self.ff(self.mhra(x, mask=mask, **kwargs))\n",
    "\n",
    "class Transformer(Module):\n",
    "    \"Transformer model: https://arxiv.org/abs/1706.03762.\"\n",
    "    def __init__(self, n_layers:int, n_heads:int, d_model:int, d_head:int, d_inner:int,\n",
    "                 resid_p:float=0., attn_p:float=0., ff_p:float=0., embed_p:float=0., bias:bool=True, scale:bool=True,\n",
    "                 act:Activation=Activation.ReLU, double_drop:bool=True, attn_cls:Callable=MultiHeadAttention,\n",
    "                 learned_pos_enc:bool=True, mask:bool=True):\n",
    "        self.mask = mask\n",
    "        #self.encoder = nn.Embedding(vocab_sz, d_model)\n",
    "        #self.pos_enc = nn.Embedding(ctx_len, d_model) if learned_pos_enc else PositionalEncoding(d_model)\n",
    "        #self.pos_enc = PositionalEncoding(29, 0.1)\n",
    "        self.drop_emb = nn.Dropout(embed_p)\n",
    "        self.layers = nn.ModuleList([DecoderLayer(n_heads, d_model, d_head, d_inner, resid_p=resid_p, attn_p=attn_p,\n",
    "                      ff_p=ff_p, bias=bias, scale=scale, act=act, double_drop=double_drop,\n",
    "                      attn_cls=attn_cls) for k in range(n_layers)])\n",
    "\n",
    "    def reset(self): pass\n",
    "\n",
    "    def forward(self, x, mask):\n",
    "        #bs, x_len = x.size()\n",
    "        #pos = torch.arange(0, x_len, device=x.device, dtype=x.dtype)\n",
    "        inp = self.drop_emb(x)# + self.pos_enc(pos)[None]) #.mul_(self.d_model ** 0.5)\n",
    "        #if False:\n",
    "        #    inp += self.pos_enc(x)[None]\n",
    "        #mask = None #torch.triu(x.new_ones(x_len, x_len), diagonal=1).byte()[None,None] if self.mask else None\n",
    "        #[None,:,:None] for einsum implementation of attention\n",
    "        for layer in self.layers: inp = layer(inp, mask=mask)\n",
    "        return inp #For the LinearDecoder\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.280402Z",
     "start_time": "2019-08-19T17:11:23.242589Z"
    }
   },
   "outputs": [],
   "source": [
    "class MultiHeadAttentionOld(Module):\n",
    "    \"MutiHeadAttention.\"\n",
    "    def __init__(self, n_heads:int, d_model:int, d_head:int=None, resid_p:float=0., attn_p:float=0., bias:bool=True,\n",
    "                 scale:bool=True):\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.n_heads,self.d_head,self.scale = n_heads,d_head,scale\n",
    "        self.attention = nn.Linear(d_model, 3 * n_heads * d_head, bias=bias)\n",
    "        self.out = nn.Linear(n_heads * d_head, d_model, bias=bias)\n",
    "        self.drop_att,self.drop_res = nn.Dropout(attn_p),nn.Dropout(resid_p)\n",
    "        self.ln = nn.LayerNorm(d_model)\n",
    "\n",
    "    def forward(self, x:Tensor, mask:Tensor=None, **kwargs):\n",
    "        return self.ln(x + self.drop_res(self.out(self._apply_attention(x, mask=mask, **kwargs))))\n",
    "\n",
    "    def _apply_attention(self, x:Tensor, mask:Tensor=None):\n",
    "        bs,x_len = x.size(0),x.size(1)\n",
    "        wq,wk,wv = torch.chunk(self.attention(x), 3, dim=-1)\n",
    "        wq,wk,wv = map(lambda x:x.view(bs, x.size(1), self.n_heads, self.d_head), (wq,wk,wv))\n",
    "        wq,wk,wv = wq.permute(0, 2, 1, 3),wk.permute(0, 2, 3, 1),wv.permute(0, 2, 1, 3)\n",
    "        attn_score = torch.matmul(wq, wk)\n",
    "        if self.scale: attn_score.div_(self.d_head ** 0.5)\n",
    "        if mask is not None:\n",
    "            attn_score = attn_score.float().masked_fill(mask,  -1.0E9).type_as(attn_score)\n",
    "        attn_prob = self.drop_att(F.softmax(attn_score, dim=-1))\n",
    "        attn_vec = torch.matmul(attn_prob, wv)\n",
    "        return attn_vec.permute(0, 2, 1, 3).contiguous().contiguous().view(bs, x_len, -1)\n",
    "    \n",
    "class AtomTransformerOld(Module):\n",
    "    def __init__(self,  n_heads,d_model, d_head=None, **kwargs):\n",
    "        \n",
    "        self.d_model = d_model\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.transformer = Transformer(n_heads=n_heads,d_model=d_model, d_head=d_head, **kwargs)\n",
    "        \n",
    "        #self.scalar    = nn.Conv1d(d_model+ n_types + 1, 4, 1)\n",
    "        self.scalar    = nn.Conv1d(d_model, 4, 1)\n",
    "        self.magnetic  = nn.Conv1d(d_model, 9, 1)\n",
    "        self.dipole    = nn.Linear(d_model*max_atoms, 3)\n",
    "        self.potential = nn.Linear(d_model*max_atoms, 1)\n",
    "        \n",
    "        self.n_atom_embedding = d_model//2\n",
    "        self.n_type_embedding = d_model - self.n_atom_embedding - 6\n",
    "        self.type_embedding = nn.Embedding(len(types)+1,self.n_type_embedding)\n",
    "        self.atom_embedding = nn.Embedding(len(atoms)+1,self.n_atom_embedding)\n",
    "    \n",
    "        #n_pos_encoder = d_model - n_type_embedding - n_atom_embedding\n",
    "        #self.pos_encoder = nn.Sequential(\n",
    "        #    nn.Conv1d(3+1+1,n_pos_encoder, 1), nn.ReLU(), nn.BatchNorm1d(n_pos_encoder),\n",
    "        #)\n",
    "        \n",
    "    def forward(self,xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms):\n",
    "        bs, _, n_pts = xyz.shape        \n",
    "        t = self.type_embedding((type+1).squeeze(1)) #* math.sqrt(self.n_atom_embedding) #.transpose(1,2)\n",
    "        a = self.atom_embedding((atom+1).squeeze(1)) #* math.sqrt(self.n_type_embedding) #.transpose(1,2)\n",
    "        \n",
    "        x = torch.cat([xyz, mulliken, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)        \n",
    "        #x = self.pos_encoder(x).transpose(1,2)\n",
    "\n",
    "        x = torch.cat([x.transpose(1,2), t, a], dim=-1) \n",
    "\n",
    "        mask = (coulomb == 0).unsqueeze(1)\n",
    "        #mask = torch.triu(x.new_ones(max_atoms, max_atoms), diagonal=1).byte()[None,:,:,None]#[None,None] \n",
    "        #print(mask.shape, mask)\n",
    "        x = self.transformer(x, mask).transpose(1,2).contiguous()\n",
    "        \n",
    "        #t_one_hot = torch.zeros(bs,n_types+1,n_pts,device=type.device,dtype=x.dtype).scatter_(1,type+1, 1.)\n",
    "        #scalar    = self.scalar(torch.cat([x, t_one_hot], dim=1))\n",
    "        scalar    = self.scalar(x)\n",
    "        \n",
    "        magnetic  = self.magnetic(x) \n",
    "        dipole    = self.dipole(x.view(bs,-1))\n",
    "        potential = self.potential(x.view(bs,-1))\n",
    "                \n",
    "        return type,ext,scalar,magnetic,dipole,potential\n",
    "    \n",
    "    def reset(self): pass\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.352590Z",
     "start_time": "2019-08-19T17:11:23.282680Z"
    }
   },
   "outputs": [],
   "source": [
    "class AtomTransformer(Module):\n",
    "    def __init__(self,  n_heads,d_model, d_head=None,dropout=0, **kwargs):\n",
    "        \n",
    "        self.d_model = d_model\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.transformer = Transformer(n_heads=n_heads,d_model=d_model, d_head=d_head, **kwargs)\n",
    "        \n",
    "        self.scalar    = nn.Conv1d(d_model+ n_types + 1, 1, 1)\n",
    "        self.magnetic  = nn.Conv1d(d_model, 9, 1)\n",
    "        self.dipole    = nn.Linear(d_model*max_atoms, 3)\n",
    "        self.potential = nn.Linear(d_model*max_atoms, 1)\n",
    "        \n",
    "        self.n_atom_embedding = d_model//2\n",
    "        self.n_type_embedding = d_model - self.n_atom_embedding - 3 # -3 best model\n",
    "        self.type_embedding = nn.Embedding(len(types)+1,self.n_type_embedding)\n",
    "        self.atom_embedding = nn.Embedding(len(atoms)+1,self.n_atom_embedding)\n",
    "        \n",
    "        self.dropout = nn.Dropout(dropout)\n",
    "    \n",
    "        #n_pos_encoder = d_model - n_type_embedding - n_atom_embedding\n",
    "        #self.pos_encoder = nn.Sequential(\n",
    "        #    nn.Conv1d(3+1+1,n_pos_encoder, 1), nn.ReLU(), nn.BatchNorm1d(n_pos_encoder),\n",
    "        #)\n",
    "        \n",
    "    def forward(self,xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms):\n",
    "        bs, _, n_pts = xyz.shape        \n",
    "        t = self.type_embedding((type+1).squeeze(1)) #* math.sqrt(self.n_atom_embedding) #.transpose(1,2)\n",
    "        a = self.atom_embedding((atom+1).squeeze(1)) #* math.sqrt(self.n_type_embedding) #.transpose(1,2)\n",
    "        \n",
    "        #x = torch.cat([xyz, mulliken, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        #x = torch.cat([xyz, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        x = xyz \n",
    "        #x = self.pos_encoder(x).transpose(1,2)\n",
    "\n",
    "        x = torch.cat([x.transpose(1,2), t, a], dim=-1) \n",
    "\n",
    "        mask = mask_atoms.to(dtype=torch.bool).unsqueeze(1)\n",
    "        #mask = (coulomb == 0).unsqueeze(1)\n",
    "        #mask = torch.triu(x.new_ones(max_atoms, max_atoms), diagonal=1).byte()[None,:,:,None]#[None,None] \n",
    "        #print(mask.shape, mask)\n",
    "        x = self.transformer(x, mask).transpose(1,2).contiguous()\n",
    "        x = self.dropout(x)\n",
    "        t_one_hot = torch.zeros(bs,n_types+1,n_pts,device=type.device,dtype=x.dtype).scatter_(1,type+1, 1.)\n",
    "        \n",
    "        scalar    = self.scalar(torch.cat([x, t_one_hot], dim=1))\n",
    "        magnetic  = self.magnetic(x) \n",
    "        dipole    = self.dipole(x.view(bs,-1))\n",
    "        potential = self.potential(x.view(bs,-1))\n",
    "                \n",
    "        return type,ext,scalar,magnetic,dipole,potential\n",
    "    \n",
    "    def reset(self): pass\n",
    "    \n",
    "    \n",
    "class AtomTransformer2(Module):\n",
    "    def __init__(self,  n_heads,d_model, d_head=None, **kwargs):\n",
    "        \n",
    "        self.d_model = d_model\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.transformer = Transformer(n_heads=n_heads,d_model=d_model, d_head=d_head, **kwargs)\n",
    "        \n",
    "        self.scalar    = nn.Conv1d(d_model+ n_types + 1, 4, 1)\n",
    "        self.magnetic  = nn.Conv1d(d_model, 9, 1)\n",
    "        self.dipole    = nn.Linear(d_model*max_atoms, 3)\n",
    "        self.potential = nn.Linear(d_model*max_atoms, 1)\n",
    "        \n",
    "        self.n_atom_embedding = d_model//2\n",
    "        self.n_type_embedding = d_model - self.n_atom_embedding - 6 # -3 best model\n",
    "        self.type_embedding = nn.Embedding(len(types)+1,self.n_type_embedding)\n",
    "        self.atom_embedding = nn.Embedding(len(atoms)+1,self.n_atom_embedding)\n",
    "    \n",
    "        #n_pos_encoder = d_model - n_type_embedding - n_atom_embedding\n",
    "        #self.pos_encoder = nn.Sequential(\n",
    "        #    nn.Conv1d(3+1+1,n_pos_encoder, 1), nn.ReLU(), nn.BatchNorm1d(n_pos_encoder),\n",
    "        #)\n",
    "        \n",
    "    def forward(self,xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms):\n",
    "        bs, _, n_pts = xyz.shape        \n",
    "        t = self.type_embedding((type+1).squeeze(1)) #* math.sqrt(self.n_atom_embedding) #.transpose(1,2)\n",
    "        a = self.atom_embedding((atom+1).squeeze(1)) #* math.sqrt(self.n_type_embedding) #.transpose(1,2)\n",
    "        \n",
    "        x = torch.cat([xyz, mulliken, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        #x = torch.cat([xyz, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        #x = xyz \n",
    "        #x = self.pos_encoder(x).transpose(1,2)\n",
    "\n",
    "        x = torch.cat([x.transpose(1,2), t, a], dim=-1) \n",
    "\n",
    "        mask = (coulomb == 0).unsqueeze(1)\n",
    "        #mask = torch.triu(x.new_ones(max_atoms, max_atoms), diagonal=1).byte()[None,:,:,None]#[None,None] \n",
    "        #print(mask.shape, mask)\n",
    "        x = self.transformer(x, mask).transpose(1,2).contiguous()\n",
    "        t_one_hot = torch.zeros(bs,n_types+1,n_pts,device=type.device,dtype=x.dtype).scatter_(1,type+1, 1.)\n",
    "        \n",
    "        scalar    = self.scalar(torch.cat([x, t_one_hot], dim=1))\n",
    "        magnetic  = self.magnetic(x) \n",
    "        dipole    = self.dipole(x.view(bs,-1))\n",
    "        potential = self.potential(x.view(bs,-1))\n",
    "                \n",
    "        return type,ext,scalar,magnetic,dipole,potential\n",
    "    \n",
    "    def reset(self): pass\n",
    "    \n",
    "\n",
    "class MultiHeadAttention3(Module):\n",
    "    \"MutiHeadAttention.\"\n",
    "    def __init__(self, n_heads:int, d_model:int, d_head:int=None, resid_p:float=0., attn_p:float=0., bias:bool=False,\n",
    "                 scale:bool=False):\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.n_heads,self.d_head,self.scale = n_heads,d_head,scale\n",
    "\n",
    "        self.attention1 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention2 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention3 = nn.Linear(d_model, n_heads * d_head, bias=bias)\n",
    "        self.attention4 = nn.Linear(d_model, d_model, bias=bias)\n",
    "        self.out = nn.Linear(n_heads * d_head, d_model, bias=bias)\n",
    "        self.drop_att,self.drop_res = nn.Dropout(attn_p),nn.Dropout(resid_p)\n",
    "        self.ln = nn.LayerNorm(d_model)\n",
    "        \n",
    "        self.scale=False\n",
    "\n",
    "    def forward(self, x:Tensor, mask:Tensor=None, **kwargs):\n",
    "        return self.ln(x + self.drop_res(self.out(self._apply_attention(x, mask=mask, **kwargs))))\n",
    "\n",
    "    def _apply_attention(self, x:Tensor, mask:Tensor=None):\n",
    "        if False:\n",
    "            if mask is not None:\n",
    "                x= x.masked_fill(mask[:,:,0].squeeze(1).unsqueeze(-1).repeat(1, 1, x.size(2)),  0.)\n",
    "        bs,x_len = x.size(0),x.size(1)\n",
    "        wq,wk,wv = self.attention1(x), self.attention2(x), self.attention3(x)\n",
    "        wq,wk,wv = map(lambda x:x.view(bs, x.size(1), self.n_heads, self.d_head), (wq,wk,wv))\n",
    "        wq,wk,wv = wq.permute(0, 2, 1, 3),wk.permute(0, 2, 3, 1),wv.permute(0, 2, 1, 3)\n",
    "        attn_score = torch.matmul(wq, wk)\n",
    "        if self.scale: attn_score.div_(self.d_head ** 0.5)\n",
    "        if mask is not None:\n",
    "            attn_score = attn_score.float().masked_fill(mask,  -1.0E9).type_as(attn_score)\n",
    "        attn_prob = self.drop_att(F.softmax(attn_score, dim=-1))\n",
    "        attn_vec = torch.matmul(attn_prob, wv)\n",
    "        return self.attention4(attn_vec.permute(0, 2, 1, 3).contiguous().contiguous().view(bs, x_len, -1))\n",
    "\n",
    "    def _attention_einsum(self, x, mask=None):\n",
    "        # Permute and matmul is a little bit faster but this implementation is more readable\n",
    "        bs,x_len = x.size(0),x.size(1)\n",
    "        wq,wk,wv = torch.chunk(self.attention(x), 3, dim=-1)\n",
    "        wq,wk,wv = map(lambda x:x.view(bs, x.size(1), self.n_heads, self.d_head), (wq,wk,wv))\n",
    "        attn_score = torch.einsum('bind,bjnd->bijn', (wq, wk))\n",
    "        if self.scale: attn_score.mul_(1/(self.d_head ** 0.5))\n",
    "        if mask is not None:\n",
    "            attn_score = attn_score.float().masked_fill(mask, -float('-inf')).type_as(attn_score)\n",
    "        attn_prob = self.drop_att(F.softmax(attn_score, dim=2))\n",
    "        attn_vec = torch.einsum('bijn,bjnd->bind', (attn_prob, wv))\n",
    "        return attn_vec.contiguous().view(bs, x_len, -1)\n",
    "\n",
    "    \n",
    "class AtomTransformer3(Module):\n",
    "    def __init__(self,  n_heads,d_model, d_head=None, **kwargs):\n",
    "        \n",
    "        self.d_model = d_model\n",
    "        d_head = ifnone(d_head, d_model//n_heads)\n",
    "        self.transformer = Transformer(n_heads=n_heads,d_model=d_model, d_head=d_head, **kwargs)\n",
    "        \n",
    "        self.scalar    = nn.Conv1d(d_model+ n_types + 1, 4, 1)\n",
    "        self.magnetic  = nn.Conv1d(d_model, 9, 1)\n",
    "        self.dipole    = nn.Linear(d_model*max_atoms, 3)\n",
    "        self.potential = nn.Linear(d_model*max_atoms, 1)\n",
    "        \n",
    "        self.n_atom_embedding = d_model//2\n",
    "        self.n_type_embedding = d_model - self.n_atom_embedding - 3 # -3 best model\n",
    "        self.type_embedding = nn.Embedding(len(types)+1,self.n_type_embedding)\n",
    "        self.atom_embedding = nn.Embedding(len(atoms)+1,self.n_atom_embedding)\n",
    "        \n",
    "        self.dropout = nn.Dropout(0.05)\n",
    "    \n",
    "        #n_pos_encoder = d_model - n_type_embedding - n_atom_embedding\n",
    "        #self.pos_encoder = nn.Sequential(\n",
    "        #    nn.Conv1d(3+1+1,n_pos_encoder, 1), nn.ReLU(), nn.BatchNorm1d(n_pos_encoder),\n",
    "        #)\n",
    "        \n",
    "    def forward(self,xyz,type,ext,atom,mulliken,coulomb, mask_atoms, n_atoms):\n",
    "        bs, _, n_pts = xyz.shape        \n",
    "        t = self.type_embedding((type+1).squeeze(1)) #* math.sqrt(self.n_atom_embedding) #.transpose(1,2)\n",
    "        a = self.atom_embedding((atom+1).squeeze(1)) #* math.sqrt(self.n_type_embedding) #.transpose(1,2)\n",
    "        \n",
    "        #x = torch.cat([xyz, mulliken, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        #x = torch.cat([xyz, ext, mask_atoms.float()], dim=1) #* math.sqrt(self.d_model)               \n",
    "        x = xyz \n",
    "        #x = self.pos_encoder(x).transpose(1,2)\n",
    "\n",
    "        x = torch.cat([x.transpose(1,2), t, a], dim=-1) \n",
    "\n",
    "        mask = (coulomb == 0).unsqueeze(1)\n",
    "        #mask = torch.triu(x.new_ones(max_atoms, max_atoms), diagonal=1).byte()[None,:,:,None]#[None,None] \n",
    "        #print(mask.shape, mask)\n",
    "        x = self.transformer(x, mask).transpose(1,2).contiguous()\n",
    "        x = self.dropout(x)\n",
    "        t_one_hot = torch.zeros(bs,n_types+1,n_pts,device=type.device,dtype=x.dtype).scatter_(1,type+1, 1.)\n",
    "        \n",
    "        scalar    = self.scalar(torch.cat([x, t_one_hot], dim=1))\n",
    "        magnetic  = self.magnetic(x) \n",
    "        dipole    = self.dipole(x.view(bs,-1))\n",
    "        potential = self.potential(x.view(bs,-1))\n",
    "                \n",
    "        return type,ext,scalar,magnetic,dipole,potential\n",
    "    \n",
    "    def reset(self): pass\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This callback allows to insert multiple stateful (not averaged) metrics in one pass. Addditionally we could add metrics for train if we want to."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model instantiation: where's all your TPUs/GPUs when you need a decent hyperparam sweep?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:23.367362Z",
     "start_time": "2019-08-19T17:11:23.354688Z"
    }
   },
   "outputs": [],
   "source": [
    "from fastai.callbacks import *\n",
    "#from fastai.callbacks import SaveModelCallback\n",
    "class SaveModelCustomCallback(TrackerCallback):\n",
    "    \"A `TrackerCallback` that saves the model when monitored quantity is best.\"\n",
    "    def __init__(self, learn:Learner, monitor:str='valid_loss', mode:str='auto', every:str='improvement', name:str='bestmodel'):\n",
    "        super().__init__(learn, monitor=monitor, mode=mode)\n",
    "        self.every,self.name = every,name\n",
    "        if self.every not in ['improvement', 'epoch']:\n",
    "            warn(f'SaveModel every {self.every} is invalid, falling back to \"improvement\".')\n",
    "            self.every = 'improvement'\n",
    "                 \n",
    "    def jump_to_epoch(self, epoch:int)->None:\n",
    "        try: \n",
    "            self.learn.load(f'{self.name}_{epoch-1}', purge=False)\n",
    "            print(f\"Loaded {self.name}_{epoch-1}\")\n",
    "        except: print(f'Model {self.name}_{epoch-1} not found.')\n",
    "\n",
    "    def on_epoch_end(self, epoch:int, **kwargs:Any)->None:\n",
    "        \"Compare the value monitored to its best score and maybe save the model.\"\n",
    "        if self.every==\"epoch\": self.learn.save(f'{self.name}_{epoch}')\n",
    "        else: #every=\"improvement\"\n",
    "            current = self.get_monitor_value()\n",
    "            if current is not None and self.operator(current, self.best):\n",
    "                print(f'Better model found at epoch {epoch} with {self.monitor} value: {current}.')\n",
    "                self.best = current\n",
    "                self.learn.save(f'{self.name}_{epoch}_{self.best}')\n",
    "\n",
    "    def on_train_end(self, **kwargs):\n",
    "        \"Load the best model.\"\n",
    "        if False and self.every==\"improvement\" and (self.learn.path/f'{self.learn.model_dir}/{self.name}.pth').is_file():\n",
    "            self.learn.load(f'{self.name}', purge=False)\n",
    "            \n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:31.712227Z",
     "start_time": "2019-08-19T17:11:23.369116Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/antor/miniconda3/lib/python3.6/site-packages/ipykernel_launcher.py:25: UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.\n"
     ]
    }
   ],
   "source": [
    "net, learner = None,None\n",
    "gc.collect()\n",
    "torch.cuda.empty_cache()\n",
    "\n",
    "net = AtomTransformer(n_layers=6, n_heads=16,d_model=1024,d_inner=4096, act=Activation.Swish, dropout=0)\n",
    "#net = AtomTransformer3(n_layers=6, n_heads=16,d_model=1024,d_inner=4096, attn_cls=MultiHeadAttention3)\n",
    "#net = AtomTransformer(n_layers=6, n_heads=16,d_model=2048,d_inner=2048)\n",
    "#net = AtomTransformer(n_layers=6, n_heads=8, d_model=512,d_inner=2048)\n",
    "\n",
    "#net = AtomTransformerOld(n_layers=6, n_heads=8,d_model=512,d_inner=2048, attn_cls=MultiHeadAttentionOld)\n",
    "\n",
    "#net = AtomTransformer2(n_layers=6, n_heads=12,d_model=768,d_inner=3072)\n",
    "\n",
    "\n",
    "#net = AtomTransformer(n_layers=6, n_heads=12,d_model=768,d_inner=3072)\n",
    "\n",
    "#net = AtomTransformer(n_layers=6, n_heads=16,d_model=1024,d_inner=4096, attn_cls=MultiHeadAttentionOld)\n",
    "\n",
    "learner = Learner(data,net, loss_func=LMAEMaskedLoss(),)\n",
    "\n",
    "learner.callbacks.append(LMAEMetric(learner))\n",
    "\n",
    "for p in learner.model.parameters():\n",
    "    if p.dim() > 1:\n",
    "        nn.init.xavier_uniform(p)\n",
    "        \n",
    "\n",
    "learner.callbacks.append(SaveModelCustomCallback(learner, monitor='👉🏻LMAE👈🏻', mode='min'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T13:53:26.675699Z",
     "start_time": "2019-08-19T13:53:24.807346Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "AtomTransformer\n",
       "======================================================================\n",
       "Layer (type)         Output Shape         Param #    Trainable \n",
       "======================================================================\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           1,049,600  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [16, 29, 29]         0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Linear               [29, 4096]           4,198,400  True      \n",
       "______________________________________________________________________\n",
       "Swish                [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 4096]           0          False     \n",
       "______________________________________________________________________\n",
       "Linear               [29, 1024]           4,195,328  True      \n",
       "______________________________________________________________________\n",
       "Dropout              [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "MergeLayer           [29, 1024]           0          False     \n",
       "______________________________________________________________________\n",
       "LayerNorm            [29, 1024]           2,048      True      \n",
       "______________________________________________________________________\n",
       "Conv1d               [1, 29]              1,034      True      \n",
       "______________________________________________________________________\n",
       "Conv1d               [9, 29]              9,225      True      \n",
       "______________________________________________________________________\n",
       "Linear               [3]                  89,091     True      \n",
       "______________________________________________________________________\n",
       "Linear               [1]                  29,697     True      \n",
       "______________________________________________________________________\n",
       "Embedding            [29, 509]            4,581      True      \n",
       "______________________________________________________________________\n",
       "Embedding            [29, 512]            3,072      True      \n",
       "______________________________________________________________________\n",
       "Dropout              [1024, 29]           0          False     \n",
       "______________________________________________________________________\n",
       "\n",
       "Total params: 82,011,644\n",
       "Total trainable params: 82,011,644\n",
       "Total non-trainable params: 0\n",
       "Optimized with 'torch.optim.adam.Adam', betas=(0.9, 0.99)\n",
       "Using true weight decay as discussed in https://www.fast.ai/2018/07/02/adam-weight-decay/ \n",
       "Loss function : LMAEMaskedLoss\n",
       "======================================================================\n",
       "Callbacks functions applied \n",
       "    LMAEMetric\n",
       "    SaveModelCustomCallback"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learner.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:31.717606Z",
     "start_time": "2019-08-19T17:11:31.714392Z"
    }
   },
   "outputs": [],
   "source": [
    "#sub_fname = \"loss-5.9943val-2.7766\" # uncomment or set None to skip loading trained net\n",
    "#sub_fname = \"loss-5.7552val-2.9950\"\n",
    "#sub_fname = \"loss-5.9943val-2.7766\"\n",
    "#sub_fname = \"loss-4.9516val-3.0042\"\n",
    "#sub_fname = \"loss-4.0414val-2.7287\"\n",
    "#sub_fname = \"loss-4.9044val-2.5880\"\n",
    "\n",
    "#sub_fname = \"loss-5.0862val-3.0047\"\n",
    "\n",
    "#sub_fname = 'loss-4.9516val-2.8229'\n",
    "#sub_fname = 'bestmodel_62_-2.2235493659973145'\n",
    "#sub_fname = 'loss-5.6540val-3.0131'\n",
    "#sub_fname = 'bestmodel_6_-0.6834144592285156'\n",
    "sub_fname = 'loss-5.1336val-3.0759'\n",
    "sub_fname = 'loss-4.6605val-2.9311'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:41.657025Z",
     "start_time": "2019-08-19T17:11:31.721110Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Attempting to load: loss-4.6605val-2.9311... Loaded\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    print(f\"Attempting to load: {sub_fname}... \", end=\"\")\n",
    "    learner.load(sub_fname, strict=False,with_opt=False)\n",
    "    print(\"Loaded\")\n",
    "except Exception as e:\n",
    "    print(\"NOT loaded! \", e)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T06:40:14.614754Z",
     "start_time": "2019-08-19T06:40:14.609905Z"
    }
   },
   "outputs": [],
   "source": [
    "if False:\n",
    "    for name, param in learner.model.named_parameters():\n",
    "\n",
    "        if 'scalar' not in name:\n",
    "            param.requires_grad = False\n",
    "            pass\n",
    "        else:\n",
    "            print(name)\n",
    "            try:\n",
    "                nn.init.xavier_uniform_(param)\n",
    "            except:\n",
    "                pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:41.662778Z",
     "start_time": "2019-08-19T17:11:41.658966Z"
    }
   },
   "outputs": [],
   "source": [
    "learner = learner.to_parallel() #b/c it NaNs loss (probably would need to change fp16 settings)\n",
    "data.batch_size = int(4096//2)\n",
    "data.batch_size = data.batch_size // 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Real loss func. Need to test different auxiliary tasks weights: `magnetic_w`, `dipole_w`, `potential`, weights of indivial `lmae`s: `types_w` and maybe `input_transform_w` and `feature_transform_w`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:11:41.668633Z",
     "start_time": "2019-08-19T17:11:41.664619Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.loss_func = LMAEMaskedLoss(magnetic_w=0, dipole_w=0, potential_w=0, \n",
    "                                  types_w = [1] + [1] * (n_types-1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T13:55:31.203434Z",
     "start_time": "2019-08-19T13:53:36.687673Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
      "Failed to compute the gradients, there might not be enough points.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2d6e50b908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#learner.data.batch_size = 512\n",
    "#learner.opt_func = RAdam\n",
    "learner.lr_find()\n",
    "learner.recorder.plot(suggestion=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "class ShowGraph(LearnerCallback):\n",
    "    \"Update a graph of learner stats and metrics after each epoch.\"\n",
    "    def on_epoch_end(self, n_epochs:int, last_metrics:MetricsList, **kwargs)->bool:\n",
    "        \"If we have `last_metrics` plot them in our pbar graph\"\n",
    "        if last_metrics is not None and last_metrics[0] is not None:\n",
    "            rec = self.learn.recorder\n",
    "            iters = range_of(rec.losses)\n",
    "            val_iter = np.array(rec.nb_batches).cumsum()\n",
    "            x_bounds = (0, (n_epochs - len(rec.nb_batches)) * rec.nb_batches[-1] + len(rec.losses))\n",
    "            y_bounds = (min((min(Tensor(rec.losses)), min(Tensor(rec.val_losses)))), \n",
    "                        max((max(Tensor(rec.losses)), max(Tensor(rec.val_losses)))))\n",
    "            rec.pbar.update_graph([(iters, rec.losses), (val_iter, rec.val_losses)], x_bounds, y_bounds)\n",
    "        return {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-19T17:12:05.721975Z",
     "start_time": "2019-08-19T17:12:05.714116Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[functools.partial(<class 'fastai.basic_train.Recorder'>, add_time=True, silent=False),\n",
       " __main__.ShowGraph]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "learner.callback_fns.append(ShowGraph)\n",
    "learner.callback_fns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "gc.collect()\n",
    "torch.cuda.empty_cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:21:04.766711Z",
     "start_time": "2019-08-19T17:12:17.245500Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 17:04:04 <p><table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: left;\">\n",
       "      <th>epoch</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>valid_loss</th>\n",
       "      <th>👉🏻LMAE👈🏻</th>\n",
       "      <th>lmae0</th>\n",
       "      <th>lmae1</th>\n",
       "      <th>lmae2</th>\n",
       "      <th>lmae3</th>\n",
       "      <th>lmae4</th>\n",
       "      <th>lmae5</th>\n",
       "      <th>lmae6</th>\n",
       "      <th>lmae7</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>-1.345332</td>\n",
       "      <td>-2.748473</td>\n",
       "      <td>-2.679040</td>\n",
       "      <td>-0.782845</td>\n",
       "      <td>-3.247969</td>\n",
       "      <td>-1.906672</td>\n",
       "      <td>-3.289868</td>\n",
       "      <td>-2.753965</td>\n",
       "      <td>-3.297657</td>\n",
       "      <td>-2.652932</td>\n",
       "      <td>-3.500413</td>\n",
       "      <td>13:40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>-1.543765</td>\n",
       "      <td>-2.776158</td>\n",
       "      <td>-2.703669</td>\n",
       "      <td>-0.826833</td>\n",
       "      <td>-3.247736</td>\n",
       "      <td>-1.965850</td>\n",
       "      <td>-3.314932</td>\n",
       "      <td>-2.768981</td>\n",
       "      <td>-3.311702</td>\n",
       "      <td>-2.676978</td>\n",
       "      <td>-3.516342</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>-1.707104</td>\n",
       "      <td>-2.781679</td>\n",
       "      <td>-2.706285</td>\n",
       "      <td>-0.845728</td>\n",
       "      <td>-3.228463</td>\n",
       "      <td>-1.970269</td>\n",
       "      <td>-3.311865</td>\n",
       "      <td>-2.791696</td>\n",
       "      <td>-3.319906</td>\n",
       "      <td>-2.664233</td>\n",
       "      <td>-3.518121</td>\n",
       "      <td>14:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>-1.857322</td>\n",
       "      <td>-2.774038</td>\n",
       "      <td>-2.702001</td>\n",
       "      <td>-0.866203</td>\n",
       "      <td>-3.174851</td>\n",
       "      <td>-1.994856</td>\n",
       "      <td>-3.323824</td>\n",
       "      <td>-2.788782</td>\n",
       "      <td>-3.302148</td>\n",
       "      <td>-2.653497</td>\n",
       "      <td>-3.511844</td>\n",
       "      <td>14:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>-1.964978</td>\n",
       "      <td>-2.787066</td>\n",
       "      <td>-2.714763</td>\n",
       "      <td>-0.891682</td>\n",
       "      <td>-3.231081</td>\n",
       "      <td>-1.985064</td>\n",
       "      <td>-3.310062</td>\n",
       "      <td>-2.800970</td>\n",
       "      <td>-3.315485</td>\n",
       "      <td>-2.667864</td>\n",
       "      <td>-3.515898</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>-2.058631</td>\n",
       "      <td>-2.792852</td>\n",
       "      <td>-2.721427</td>\n",
       "      <td>-0.936375</td>\n",
       "      <td>-3.245239</td>\n",
       "      <td>-2.018569</td>\n",
       "      <td>-3.319779</td>\n",
       "      <td>-2.796023</td>\n",
       "      <td>-3.295660</td>\n",
       "      <td>-2.649035</td>\n",
       "      <td>-3.510734</td>\n",
       "      <td>14:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>-2.161210</td>\n",
       "      <td>-2.771845</td>\n",
       "      <td>-2.705368</td>\n",
       "      <td>-0.954646</td>\n",
       "      <td>-3.231034</td>\n",
       "      <td>-2.002103</td>\n",
       "      <td>-3.279071</td>\n",
       "      <td>-2.796009</td>\n",
       "      <td>-3.297114</td>\n",
       "      <td>-2.655784</td>\n",
       "      <td>-3.427183</td>\n",
       "      <td>14:03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>-2.212181</td>\n",
       "      <td>-2.787713</td>\n",
       "      <td>-2.721007</td>\n",
       "      <td>-0.977791</td>\n",
       "      <td>-3.239849</td>\n",
       "      <td>-1.972915</td>\n",
       "      <td>-3.274491</td>\n",
       "      <td>-2.801153</td>\n",
       "      <td>-3.335361</td>\n",
       "      <td>-2.679051</td>\n",
       "      <td>-3.487444</td>\n",
       "      <td>14:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>-2.283310</td>\n",
       "      <td>-2.755374</td>\n",
       "      <td>-2.694543</td>\n",
       "      <td>-0.948804</td>\n",
       "      <td>-3.244930</td>\n",
       "      <td>-1.952572</td>\n",
       "      <td>-3.269340</td>\n",
       "      <td>-2.770906</td>\n",
       "      <td>-3.316216</td>\n",
       "      <td>-2.588831</td>\n",
       "      <td>-3.464745</td>\n",
       "      <td>14:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>-2.347851</td>\n",
       "      <td>-2.761206</td>\n",
       "      <td>-2.698659</td>\n",
       "      <td>-1.096253</td>\n",
       "      <td>-3.186007</td>\n",
       "      <td>-1.931477</td>\n",
       "      <td>-3.258833</td>\n",
       "      <td>-2.788242</td>\n",
       "      <td>-3.317108</td>\n",
       "      <td>-2.638488</td>\n",
       "      <td>-3.372868</td>\n",
       "      <td>14:07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>-2.377731</td>\n",
       "      <td>-2.734848</td>\n",
       "      <td>-2.675909</td>\n",
       "      <td>-1.120467</td>\n",
       "      <td>-3.238616</td>\n",
       "      <td>-1.878613</td>\n",
       "      <td>-3.170450</td>\n",
       "      <td>-2.705938</td>\n",
       "      <td>-3.188369</td>\n",
       "      <td>-2.654520</td>\n",
       "      <td>-3.450300</td>\n",
       "      <td>14:17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>-2.418602</td>\n",
       "      <td>-2.769939</td>\n",
       "      <td>-2.710850</td>\n",
       "      <td>-1.243265</td>\n",
       "      <td>-3.230296</td>\n",
       "      <td>-1.937306</td>\n",
       "      <td>-3.174768</td>\n",
       "      <td>-2.759432</td>\n",
       "      <td>-3.252047</td>\n",
       "      <td>-2.655418</td>\n",
       "      <td>-3.434271</td>\n",
       "      <td>14:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>-2.459236</td>\n",
       "      <td>-2.775788</td>\n",
       "      <td>-2.717164</td>\n",
       "      <td>-1.365606</td>\n",
       "      <td>-3.218607</td>\n",
       "      <td>-1.878281</td>\n",
       "      <td>-3.215246</td>\n",
       "      <td>-2.762856</td>\n",
       "      <td>-3.260286</td>\n",
       "      <td>-2.619052</td>\n",
       "      <td>-3.417375</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>-2.482899</td>\n",
       "      <td>-2.755309</td>\n",
       "      <td>-2.706052</td>\n",
       "      <td>-1.334042</td>\n",
       "      <td>-3.227862</td>\n",
       "      <td>-1.842550</td>\n",
       "      <td>-3.186350</td>\n",
       "      <td>-2.760401</td>\n",
       "      <td>-3.266184</td>\n",
       "      <td>-2.618684</td>\n",
       "      <td>-3.412338</td>\n",
       "      <td>14:22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>-2.503970</td>\n",
       "      <td>-2.739795</td>\n",
       "      <td>-2.681145</td>\n",
       "      <td>-1.506413</td>\n",
       "      <td>-3.097660</td>\n",
       "      <td>-1.718909</td>\n",
       "      <td>-3.153223</td>\n",
       "      <td>-2.749688</td>\n",
       "      <td>-3.243898</td>\n",
       "      <td>-2.597367</td>\n",
       "      <td>-3.382003</td>\n",
       "      <td>14:18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>-2.523919</td>\n",
       "      <td>-2.701810</td>\n",
       "      <td>-2.656072</td>\n",
       "      <td>-1.601915</td>\n",
       "      <td>-3.014671</td>\n",
       "      <td>-1.717525</td>\n",
       "      <td>-3.147288</td>\n",
       "      <td>-2.732723</td>\n",
       "      <td>-3.142624</td>\n",
       "      <td>-2.562435</td>\n",
       "      <td>-3.329393</td>\n",
       "      <td>14:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>-2.524151</td>\n",
       "      <td>-2.645114</td>\n",
       "      <td>-2.602362</td>\n",
       "      <td>-1.626355</td>\n",
       "      <td>-2.853826</td>\n",
       "      <td>-1.836561</td>\n",
       "      <td>-2.985532</td>\n",
       "      <td>-2.636134</td>\n",
       "      <td>-3.157125</td>\n",
       "      <td>-2.619428</td>\n",
       "      <td>-3.103937</td>\n",
       "      <td>14:13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>-2.531968</td>\n",
       "      <td>-2.744654</td>\n",
       "      <td>-2.690172</td>\n",
       "      <td>-1.681999</td>\n",
       "      <td>-3.192026</td>\n",
       "      <td>-1.675984</td>\n",
       "      <td>-3.118546</td>\n",
       "      <td>-2.737398</td>\n",
       "      <td>-3.175500</td>\n",
       "      <td>-2.599208</td>\n",
       "      <td>-3.340718</td>\n",
       "      <td>14:30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>-2.532512</td>\n",
       "      <td>-2.667063</td>\n",
       "      <td>-2.625432</td>\n",
       "      <td>-1.630134</td>\n",
       "      <td>-3.058774</td>\n",
       "      <td>-1.762710</td>\n",
       "      <td>-3.059277</td>\n",
       "      <td>-2.521735</td>\n",
       "      <td>-3.148603</td>\n",
       "      <td>-2.599665</td>\n",
       "      <td>-3.222559</td>\n",
       "      <td>14:20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>-2.559874</td>\n",
       "      <td>-2.682943</td>\n",
       "      <td>-2.638726</td>\n",
       "      <td>-1.577330</td>\n",
       "      <td>-3.102856</td>\n",
       "      <td>-1.808946</td>\n",
       "      <td>-3.032879</td>\n",
       "      <td>-2.692217</td>\n",
       "      <td>-3.057810</td>\n",
       "      <td>-2.604698</td>\n",
       "      <td>-3.233072</td>\n",
       "      <td>14:20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>-2.557418</td>\n",
       "      <td>-2.725682</td>\n",
       "      <td>-2.672161</td>\n",
       "      <td>-1.628770</td>\n",
       "      <td>-3.162508</td>\n",
       "      <td>-1.732425</td>\n",
       "      <td>-3.099150</td>\n",
       "      <td>-2.730033</td>\n",
       "      <td>-3.096067</td>\n",
       "      <td>-2.599581</td>\n",
       "      <td>-3.328756</td>\n",
       "      <td>14:28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>21</td>\n",
       "      <td>-2.584506</td>\n",
       "      <td>-2.718478</td>\n",
       "      <td>-2.672325</td>\n",
       "      <td>-1.550816</td>\n",
       "      <td>-3.080406</td>\n",
       "      <td>-1.701431</td>\n",
       "      <td>-3.106407</td>\n",
       "      <td>-2.736108</td>\n",
       "      <td>-3.230044</td>\n",
       "      <td>-2.630723</td>\n",
       "      <td>-3.342662</td>\n",
       "      <td>14:23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>22</td>\n",
       "      <td>-2.588574</td>\n",
       "      <td>-2.724339</td>\n",
       "      <td>-2.671386</td>\n",
       "      <td>-1.644729</td>\n",
       "      <td>-3.178526</td>\n",
       "      <td>-1.601259</td>\n",
       "      <td>-3.118309</td>\n",
       "      <td>-2.696237</td>\n",
       "      <td>-3.202448</td>\n",
       "      <td>-2.617023</td>\n",
       "      <td>-3.312555</td>\n",
       "      <td>14:16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>23</td>\n",
       "      <td>-2.619690</td>\n",
       "      <td>-2.729383</td>\n",
       "      <td>-2.683976</td>\n",
       "      <td>-1.657032</td>\n",
       "      <td>-3.231825</td>\n",
       "      <td>-1.760203</td>\n",
       "      <td>-3.050991</td>\n",
       "      <td>-2.709481</td>\n",
       "      <td>-3.105836</td>\n",
       "      <td>-2.573191</td>\n",
       "      <td>-3.383252</td>\n",
       "      <td>14:17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>24</td>\n",
       "      <td>-2.622829</td>\n",
       "      <td>-2.765367</td>\n",
       "      <td>-2.715625</td>\n",
       "      <td>-1.677616</td>\n",
       "      <td>-3.157955</td>\n",
       "      <td>-1.801989</td>\n",
       "      <td>-3.138006</td>\n",
       "      <td>-2.756558</td>\n",
       "      <td>-3.242727</td>\n",
       "      <td>-2.632722</td>\n",
       "      <td>-3.317426</td>\n",
       "      <td>14:19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>25</td>\n",
       "      <td>-2.649309</td>\n",
       "      <td>-2.755706</td>\n",
       "      <td>-2.708229</td>\n",
       "      <td>-1.692776</td>\n",
       "      <td>-3.199370</td>\n",
       "      <td>-1.768581</td>\n",
       "      <td>-3.066416</td>\n",
       "      <td>-2.740252</td>\n",
       "      <td>-3.208895</td>\n",
       "      <td>-2.632264</td>\n",
       "      <td>-3.357281</td>\n",
       "      <td>14:14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>26</td>\n",
       "      <td>-2.663086</td>\n",
       "      <td>-2.747771</td>\n",
       "      <td>-2.700248</td>\n",
       "      <td>-1.671167</td>\n",
       "      <td>-3.161822</td>\n",
       "      <td>-1.799876</td>\n",
       "      <td>-3.111472</td>\n",
       "      <td>-2.738188</td>\n",
       "      <td>-3.179581</td>\n",
       "      <td>-2.602674</td>\n",
       "      <td>-3.337208</td>\n",
       "      <td>14:18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>27</td>\n",
       "      <td>-2.676698</td>\n",
       "      <td>-2.785297</td>\n",
       "      <td>-2.730273</td>\n",
       "      <td>-1.681941</td>\n",
       "      <td>-3.205556</td>\n",
       "      <td>-1.792645</td>\n",
       "      <td>-3.131470</td>\n",
       "      <td>-2.740541</td>\n",
       "      <td>-3.270936</td>\n",
       "      <td>-2.638236</td>\n",
       "      <td>-3.380858</td>\n",
       "      <td>14:18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>28</td>\n",
       "      <td>-2.710459</td>\n",
       "      <td>-2.712149</td>\n",
       "      <td>-2.665599</td>\n",
       "      <td>-1.611307</td>\n",
       "      <td>-3.090433</td>\n",
       "      <td>-1.804396</td>\n",
       "      <td>-3.087729</td>\n",
       "      <td>-2.593960</td>\n",
       "      <td>-3.219311</td>\n",
       "      <td>-2.664354</td>\n",
       "      <td>-3.253303</td>\n",
       "      <td>14:11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>29</td>\n",
       "      <td>-2.700149</td>\n",
       "      <td>-2.753410</td>\n",
       "      <td>-2.699852</td>\n",
       "      <td>-1.617173</td>\n",
       "      <td>-3.229374</td>\n",
       "      <td>-1.726103</td>\n",
       "      <td>-3.162882</td>\n",
       "      <td>-2.764471</td>\n",
       "      <td>-3.097374</td>\n",
       "      <td>-2.618982</td>\n",
       "      <td>-3.382461</td>\n",
       "      <td>14:13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>30</td>\n",
       "      <td>-2.731170</td>\n",
       "      <td>-2.795199</td>\n",
       "      <td>-2.742439</td>\n",
       "      <td>-1.643909</td>\n",
       "      <td>-3.196108</td>\n",
       "      <td>-1.851191</td>\n",
       "      <td>-3.166698</td>\n",
       "      <td>-2.768557</td>\n",
       "      <td>-3.258217</td>\n",
       "      <td>-2.666529</td>\n",
       "      <td>-3.388303</td>\n",
       "      <td>14:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>31</td>\n",
       "      <td>-2.745506</td>\n",
       "      <td>-2.805357</td>\n",
       "      <td>-2.750351</td>\n",
       "      <td>-1.692843</td>\n",
       "      <td>-3.263676</td>\n",
       "      <td>-1.856091</td>\n",
       "      <td>-3.091559</td>\n",
       "      <td>-2.781124</td>\n",
       "      <td>-3.243724</td>\n",
       "      <td>-2.687706</td>\n",
       "      <td>-3.386084</td>\n",
       "      <td>14:08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>32</td>\n",
       "      <td>-2.759549</td>\n",
       "      <td>-2.833153</td>\n",
       "      <td>-2.774859</td>\n",
       "      <td>-1.725931</td>\n",
       "      <td>-3.255618</td>\n",
       "      <td>-1.862137</td>\n",
       "      <td>-3.163716</td>\n",
       "      <td>-2.798874</td>\n",
       "      <td>-3.300572</td>\n",
       "      <td>-2.663475</td>\n",
       "      <td>-3.428555</td>\n",
       "      <td>14:14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>33</td>\n",
       "      <td>-2.777858</td>\n",
       "      <td>-2.812116</td>\n",
       "      <td>-2.760837</td>\n",
       "      <td>-1.681780</td>\n",
       "      <td>-3.255600</td>\n",
       "      <td>-1.837522</td>\n",
       "      <td>-3.128767</td>\n",
       "      <td>-2.783957</td>\n",
       "      <td>-3.278590</td>\n",
       "      <td>-2.710536</td>\n",
       "      <td>-3.409946</td>\n",
       "      <td>14:09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>34</td>\n",
       "      <td>-2.793449</td>\n",
       "      <td>-2.832613</td>\n",
       "      <td>-2.775856</td>\n",
       "      <td>-1.764779</td>\n",
       "      <td>-3.254056</td>\n",
       "      <td>-1.895666</td>\n",
       "      <td>-3.211641</td>\n",
       "      <td>-2.750959</td>\n",
       "      <td>-3.254413</td>\n",
       "      <td>-2.662795</td>\n",
       "      <td>-3.412540</td>\n",
       "      <td>14:09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>35</td>\n",
       "      <td>-2.820778</td>\n",
       "      <td>-2.844601</td>\n",
       "      <td>-2.790989</td>\n",
       "      <td>-1.772499</td>\n",
       "      <td>-3.281681</td>\n",
       "      <td>-1.827678</td>\n",
       "      <td>-3.183285</td>\n",
       "      <td>-2.844349</td>\n",
       "      <td>-3.237296</td>\n",
       "      <td>-2.728351</td>\n",
       "      <td>-3.452771</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>36</td>\n",
       "      <td>-2.860034</td>\n",
       "      <td>-2.853294</td>\n",
       "      <td>-2.791064</td>\n",
       "      <td>-1.740022</td>\n",
       "      <td>-3.280120</td>\n",
       "      <td>-1.802512</td>\n",
       "      <td>-3.199559</td>\n",
       "      <td>-2.835940</td>\n",
       "      <td>-3.285317</td>\n",
       "      <td>-2.734996</td>\n",
       "      <td>-3.450042</td>\n",
       "      <td>14:16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>37</td>\n",
       "      <td>-2.833204</td>\n",
       "      <td>-2.875198</td>\n",
       "      <td>-2.817603</td>\n",
       "      <td>-1.778728</td>\n",
       "      <td>-3.308102</td>\n",
       "      <td>-1.862974</td>\n",
       "      <td>-3.211293</td>\n",
       "      <td>-2.858036</td>\n",
       "      <td>-3.323196</td>\n",
       "      <td>-2.743190</td>\n",
       "      <td>-3.455301</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>38</td>\n",
       "      <td>-2.900687</td>\n",
       "      <td>-2.838309</td>\n",
       "      <td>-2.779783</td>\n",
       "      <td>-1.787092</td>\n",
       "      <td>-3.176095</td>\n",
       "      <td>-1.915884</td>\n",
       "      <td>-3.108539</td>\n",
       "      <td>-2.821755</td>\n",
       "      <td>-3.268572</td>\n",
       "      <td>-2.716471</td>\n",
       "      <td>-3.443859</td>\n",
       "      <td>14:20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>39</td>\n",
       "      <td>-2.913214</td>\n",
       "      <td>-2.870275</td>\n",
       "      <td>-2.811803</td>\n",
       "      <td>-1.775534</td>\n",
       "      <td>-3.306304</td>\n",
       "      <td>-1.836201</td>\n",
       "      <td>-3.242153</td>\n",
       "      <td>-2.840878</td>\n",
       "      <td>-3.298588</td>\n",
       "      <td>-2.762555</td>\n",
       "      <td>-3.432211</td>\n",
       "      <td>14:12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>40</td>\n",
       "      <td>-2.945816</td>\n",
       "      <td>-2.907721</td>\n",
       "      <td>-2.847614</td>\n",
       "      <td>-1.808444</td>\n",
       "      <td>-3.320759</td>\n",
       "      <td>-1.919047</td>\n",
       "      <td>-3.243853</td>\n",
       "      <td>-2.867482</td>\n",
       "      <td>-3.364052</td>\n",
       "      <td>-2.771333</td>\n",
       "      <td>-3.485946</td>\n",
       "      <td>14:18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>41</td>\n",
       "      <td>-2.961558</td>\n",
       "      <td>-2.903800</td>\n",
       "      <td>-2.846324</td>\n",
       "      <td>-1.788797</td>\n",
       "      <td>-3.259542</td>\n",
       "      <td>-1.949312</td>\n",
       "      <td>-3.251453</td>\n",
       "      <td>-2.888451</td>\n",
       "      <td>-3.381554</td>\n",
       "      <td>-2.764787</td>\n",
       "      <td>-3.486693</td>\n",
       "      <td>14:01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>42</td>\n",
       "      <td>-2.973121</td>\n",
       "      <td>-2.925465</td>\n",
       "      <td>-2.866164</td>\n",
       "      <td>-1.811704</td>\n",
       "      <td>-3.327106</td>\n",
       "      <td>-1.863530</td>\n",
       "      <td>-3.307913</td>\n",
       "      <td>-2.904037</td>\n",
       "      <td>-3.416960</td>\n",
       "      <td>-2.781783</td>\n",
       "      <td>-3.516277</td>\n",
       "      <td>14:03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>43</td>\n",
       "      <td>-3.012027</td>\n",
       "      <td>-2.945095</td>\n",
       "      <td>-2.884117</td>\n",
       "      <td>-1.842726</td>\n",
       "      <td>-3.329088</td>\n",
       "      <td>-1.947831</td>\n",
       "      <td>-3.287923</td>\n",
       "      <td>-2.917989</td>\n",
       "      <td>-3.412166</td>\n",
       "      <td>-2.802860</td>\n",
       "      <td>-3.532350</td>\n",
       "      <td>13:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>44</td>\n",
       "      <td>-3.031290</td>\n",
       "      <td>-2.953729</td>\n",
       "      <td>-2.888817</td>\n",
       "      <td>-1.829209</td>\n",
       "      <td>-3.361046</td>\n",
       "      <td>-1.970207</td>\n",
       "      <td>-3.312055</td>\n",
       "      <td>-2.922200</td>\n",
       "      <td>-3.394711</td>\n",
       "      <td>-2.802621</td>\n",
       "      <td>-3.518486</td>\n",
       "      <td>14:01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>45</td>\n",
       "      <td>-3.043535</td>\n",
       "      <td>-2.938565</td>\n",
       "      <td>-2.879377</td>\n",
       "      <td>-1.856260</td>\n",
       "      <td>-3.387436</td>\n",
       "      <td>-1.919571</td>\n",
       "      <td>-3.303322</td>\n",
       "      <td>-2.898764</td>\n",
       "      <td>-3.391622</td>\n",
       "      <td>-2.784399</td>\n",
       "      <td>-3.493640</td>\n",
       "      <td>14:03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>46</td>\n",
       "      <td>-3.094906</td>\n",
       "      <td>-2.958269</td>\n",
       "      <td>-2.897243</td>\n",
       "      <td>-1.836454</td>\n",
       "      <td>-3.403807</td>\n",
       "      <td>-1.962216</td>\n",
       "      <td>-3.323985</td>\n",
       "      <td>-2.901929</td>\n",
       "      <td>-3.415480</td>\n",
       "      <td>-2.788979</td>\n",
       "      <td>-3.545097</td>\n",
       "      <td>14:09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>47</td>\n",
       "      <td>-3.103240</td>\n",
       "      <td>-2.996716</td>\n",
       "      <td>-2.927547</td>\n",
       "      <td>-1.891122</td>\n",
       "      <td>-3.400423</td>\n",
       "      <td>-2.022477</td>\n",
       "      <td>-3.334812</td>\n",
       "      <td>-2.943041</td>\n",
       "      <td>-3.463151</td>\n",
       "      <td>-2.820658</td>\n",
       "      <td>-3.544696</td>\n",
       "      <td>14:04</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>48</td>\n",
       "      <td>-3.143216</td>\n",
       "      <td>-2.990183</td>\n",
       "      <td>-2.923592</td>\n",
       "      <td>-1.844270</td>\n",
       "      <td>-3.403337</td>\n",
       "      <td>-2.019234</td>\n",
       "      <td>-3.320647</td>\n",
       "      <td>-2.965506</td>\n",
       "      <td>-3.428466</td>\n",
       "      <td>-2.840561</td>\n",
       "      <td>-3.566710</td>\n",
       "      <td>14:08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>49</td>\n",
       "      <td>-3.160671</td>\n",
       "      <td>-2.979655</td>\n",
       "      <td>-2.920542</td>\n",
       "      <td>-1.898843</td>\n",
       "      <td>-3.423032</td>\n",
       "      <td>-2.017552</td>\n",
       "      <td>-3.289376</td>\n",
       "      <td>-2.972967</td>\n",
       "      <td>-3.429029</td>\n",
       "      <td>-2.834635</td>\n",
       "      <td>-3.498897</td>\n",
       "      <td>14:14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>50</td>\n",
       "      <td>-3.186113</td>\n",
       "      <td>-3.032015</td>\n",
       "      <td>-2.962419</td>\n",
       "      <td>-1.879456</td>\n",
       "      <td>-3.435697</td>\n",
       "      <td>-2.058877</td>\n",
       "      <td>-3.398696</td>\n",
       "      <td>-2.977523</td>\n",
       "      <td>-3.496806</td>\n",
       "      <td>-2.856029</td>\n",
       "      <td>-3.596267</td>\n",
       "      <td>14:05</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>51</td>\n",
       "      <td>-3.223648</td>\n",
       "      <td>-3.022255</td>\n",
       "      <td>-2.955349</td>\n",
       "      <td>-1.905449</td>\n",
       "      <td>-3.442013</td>\n",
       "      <td>-2.024325</td>\n",
       "      <td>-3.372566</td>\n",
       "      <td>-2.995387</td>\n",
       "      <td>-3.481207</td>\n",
       "      <td>-2.824750</td>\n",
       "      <td>-3.597096</td>\n",
       "      <td>13:58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>52</td>\n",
       "      <td>-3.244465</td>\n",
       "      <td>-3.035395</td>\n",
       "      <td>-2.963674</td>\n",
       "      <td>-1.902289</td>\n",
       "      <td>-3.427646</td>\n",
       "      <td>-2.019521</td>\n",
       "      <td>-3.397113</td>\n",
       "      <td>-2.991748</td>\n",
       "      <td>-3.495748</td>\n",
       "      <td>-2.866159</td>\n",
       "      <td>-3.609169</td>\n",
       "      <td>13:56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>53</td>\n",
       "      <td>-3.271755</td>\n",
       "      <td>-3.027669</td>\n",
       "      <td>-2.955694</td>\n",
       "      <td>-1.904088</td>\n",
       "      <td>-3.433967</td>\n",
       "      <td>-2.029869</td>\n",
       "      <td>-3.397466</td>\n",
       "      <td>-2.997291</td>\n",
       "      <td>-3.474319</td>\n",
       "      <td>-2.871814</td>\n",
       "      <td>-3.536740</td>\n",
       "      <td>14:01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>54</td>\n",
       "      <td>-3.303994</td>\n",
       "      <td>-3.051290</td>\n",
       "      <td>-2.978745</td>\n",
       "      <td>-1.909121</td>\n",
       "      <td>-3.466590</td>\n",
       "      <td>-2.119110</td>\n",
       "      <td>-3.374722</td>\n",
       "      <td>-3.001902</td>\n",
       "      <td>-3.450465</td>\n",
       "      <td>-2.883003</td>\n",
       "      <td>-3.625045</td>\n",
       "      <td>14:02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>55</td>\n",
       "      <td>-3.328321</td>\n",
       "      <td>-3.067236</td>\n",
       "      <td>-2.992588</td>\n",
       "      <td>-1.921254</td>\n",
       "      <td>-3.431655</td>\n",
       "      <td>-2.119983</td>\n",
       "      <td>-3.422694</td>\n",
       "      <td>-3.014394</td>\n",
       "      <td>-3.504194</td>\n",
       "      <td>-2.887795</td>\n",
       "      <td>-3.638732</td>\n",
       "      <td>14:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>56</td>\n",
       "      <td>-3.358905</td>\n",
       "      <td>-3.080061</td>\n",
       "      <td>-3.006556</td>\n",
       "      <td>-1.963917</td>\n",
       "      <td>-3.457951</td>\n",
       "      <td>-2.123321</td>\n",
       "      <td>-3.432697</td>\n",
       "      <td>-3.018756</td>\n",
       "      <td>-3.542839</td>\n",
       "      <td>-2.884884</td>\n",
       "      <td>-3.628086</td>\n",
       "      <td>13:58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>57</td>\n",
       "      <td>-3.382239</td>\n",
       "      <td>-3.083541</td>\n",
       "      <td>-3.007826</td>\n",
       "      <td>-1.957760</td>\n",
       "      <td>-3.468108</td>\n",
       "      <td>-2.119706</td>\n",
       "      <td>-3.441438</td>\n",
       "      <td>-3.012138</td>\n",
       "      <td>-3.549907</td>\n",
       "      <td>-2.894952</td>\n",
       "      <td>-3.618603</td>\n",
       "      <td>13:56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>58</td>\n",
       "      <td>-3.403547</td>\n",
       "      <td>-3.104383</td>\n",
       "      <td>-3.025363</td>\n",
       "      <td>-1.953253</td>\n",
       "      <td>-3.492403</td>\n",
       "      <td>-2.133121</td>\n",
       "      <td>-3.457292</td>\n",
       "      <td>-3.042641</td>\n",
       "      <td>-3.564639</td>\n",
       "      <td>-2.908413</td>\n",
       "      <td>-3.651142</td>\n",
       "      <td>14:01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>59</td>\n",
       "      <td>-3.426418</td>\n",
       "      <td>-3.108679</td>\n",
       "      <td>-3.031108</td>\n",
       "      <td>-1.954579</td>\n",
       "      <td>-3.490631</td>\n",
       "      <td>-2.146619</td>\n",
       "      <td>-3.464525</td>\n",
       "      <td>-3.044297</td>\n",
       "      <td>-3.570198</td>\n",
       "      <td>-2.914170</td>\n",
       "      <td>-3.663845</td>\n",
       "      <td>13:58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>60</td>\n",
       "      <td>-3.456873</td>\n",
       "      <td>-3.114371</td>\n",
       "      <td>-3.036221</td>\n",
       "      <td>-1.976712</td>\n",
       "      <td>-3.489186</td>\n",
       "      <td>-2.155019</td>\n",
       "      <td>-3.459773</td>\n",
       "      <td>-3.048234</td>\n",
       "      <td>-3.572049</td>\n",
       "      <td>-2.921060</td>\n",
       "      <td>-3.667734</td>\n",
       "      <td>13:50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>61</td>\n",
       "      <td>-3.470994</td>\n",
       "      <td>-3.123409</td>\n",
       "      <td>-3.041936</td>\n",
       "      <td>-1.963358</td>\n",
       "      <td>-3.502327</td>\n",
       "      <td>-2.161102</td>\n",
       "      <td>-3.475718</td>\n",
       "      <td>-3.052109</td>\n",
       "      <td>-3.580457</td>\n",
       "      <td>-2.927252</td>\n",
       "      <td>-3.673167</td>\n",
       "      <td>13:58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>62</td>\n",
       "      <td>-3.493585</td>\n",
       "      <td>-3.123539</td>\n",
       "      <td>-3.043234</td>\n",
       "      <td>-1.987992</td>\n",
       "      <td>-3.494529</td>\n",
       "      <td>-2.156209</td>\n",
       "      <td>-3.476925</td>\n",
       "      <td>-3.052073</td>\n",
       "      <td>-3.583078</td>\n",
       "      <td>-2.924615</td>\n",
       "      <td>-3.670447</td>\n",
       "      <td>13:55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>63</td>\n",
       "      <td>-3.510224</td>\n",
       "      <td>-3.129211</td>\n",
       "      <td>-3.047004</td>\n",
       "      <td>-1.989746</td>\n",
       "      <td>-3.506343</td>\n",
       "      <td>-2.171654</td>\n",
       "      <td>-3.478615</td>\n",
       "      <td>-3.057465</td>\n",
       "      <td>-3.574298</td>\n",
       "      <td>-2.921814</td>\n",
       "      <td>-3.676102</td>\n",
       "      <td>13:52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>64</td>\n",
       "      <td>-3.529662</td>\n",
       "      <td>-3.138843</td>\n",
       "      <td>-3.055849</td>\n",
       "      <td>-2.006945</td>\n",
       "      <td>-3.517643</td>\n",
       "      <td>-2.168443</td>\n",
       "      <td>-3.491108</td>\n",
       "      <td>-3.063338</td>\n",
       "      <td>-3.593415</td>\n",
       "      <td>-2.933640</td>\n",
       "      <td>-3.672263</td>\n",
       "      <td>13:57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>65</td>\n",
       "      <td>-3.549557</td>\n",
       "      <td>-3.139750</td>\n",
       "      <td>-3.056791</td>\n",
       "      <td>-1.993790</td>\n",
       "      <td>-3.519794</td>\n",
       "      <td>-2.178052</td>\n",
       "      <td>-3.486221</td>\n",
       "      <td>-3.060828</td>\n",
       "      <td>-3.595798</td>\n",
       "      <td>-2.933738</td>\n",
       "      <td>-3.686103</td>\n",
       "      <td>13:57</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>66</td>\n",
       "      <td>-3.564250</td>\n",
       "      <td>-3.144718</td>\n",
       "      <td>-3.060513</td>\n",
       "      <td>-1.993234</td>\n",
       "      <td>-3.523807</td>\n",
       "      <td>-2.186024</td>\n",
       "      <td>-3.496622</td>\n",
       "      <td>-3.067790</td>\n",
       "      <td>-3.598065</td>\n",
       "      <td>-2.934985</td>\n",
       "      <td>-3.683574</td>\n",
       "      <td>14:02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>67</td>\n",
       "      <td>-3.563585</td>\n",
       "      <td>-3.142576</td>\n",
       "      <td>-3.059216</td>\n",
       "      <td>-1.997634</td>\n",
       "      <td>-3.515396</td>\n",
       "      <td>-2.182414</td>\n",
       "      <td>-3.494532</td>\n",
       "      <td>-3.065658</td>\n",
       "      <td>-3.597117</td>\n",
       "      <td>-2.935441</td>\n",
       "      <td>-3.685538</td>\n",
       "      <td>14:10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>68</td>\n",
       "      <td>-3.584130</td>\n",
       "      <td>-3.147837</td>\n",
       "      <td>-3.063499</td>\n",
       "      <td>-2.005533</td>\n",
       "      <td>-3.524775</td>\n",
       "      <td>-2.191517</td>\n",
       "      <td>-3.496710</td>\n",
       "      <td>-3.069022</td>\n",
       "      <td>-3.598145</td>\n",
       "      <td>-2.935103</td>\n",
       "      <td>-3.687190</td>\n",
       "      <td>14:06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>69</td>\n",
       "      <td>-3.582775</td>\n",
       "      <td>-3.147274</td>\n",
       "      <td>-3.062885</td>\n",
       "      <td>-2.001316</td>\n",
       "      <td>-3.522973</td>\n",
       "      <td>-2.188678</td>\n",
       "      <td>-3.497720</td>\n",
       "      <td>-3.067370</td>\n",
       "      <td>-3.601610</td>\n",
       "      <td>-2.935834</td>\n",
       "      <td>-3.687580</td>\n",
       "      <td>14:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>70</td>\n",
       "      <td>-3.584350</td>\n",
       "      <td>-3.147742</td>\n",
       "      <td>-3.063228</td>\n",
       "      <td>-2.000999</td>\n",
       "      <td>-3.521950</td>\n",
       "      <td>-2.188676</td>\n",
       "      <td>-3.498109</td>\n",
       "      <td>-3.068861</td>\n",
       "      <td>-3.602209</td>\n",
       "      <td>-2.937190</td>\n",
       "      <td>-3.687835</td>\n",
       "      <td>13:53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>71</td>\n",
       "      <td>-3.593262</td>\n",
       "      <td>-3.148260</td>\n",
       "      <td>-3.063699</td>\n",
       "      <td>-2.001170</td>\n",
       "      <td>-3.522323</td>\n",
       "      <td>-2.189451</td>\n",
       "      <td>-3.498910</td>\n",
       "      <td>-3.069365</td>\n",
       "      <td>-3.602329</td>\n",
       "      <td>-2.937218</td>\n",
       "      <td>-3.688823</td>\n",
       "      <td>14:00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with 👉🏻LMAE👈🏻 value: -2.6790401935577393.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f3003952898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 1 with 👉🏻LMAE👈🏻 value: -2.703669309616089.\n",
      "Better model found at epoch 2 with 👉🏻LMAE👈🏻 value: -2.706284999847412.\n",
      "Better model found at epoch 4 with 👉🏻LMAE👈🏻 value: -2.7147634029388428.\n",
      "Better model found at epoch 5 with 👉🏻LMAE👈🏻 value: -2.7214267253875732.\n",
      "Better model found at epoch 27 with 👉🏻LMAE👈🏻 value: -2.7302730083465576.\n",
      "Better model found at epoch 30 with 👉🏻LMAE👈🏻 value: -2.742439031600952.\n",
      "Better model found at epoch 31 with 👉🏻LMAE👈🏻 value: -2.7503509521484375.\n",
      "Better model found at epoch 32 with 👉🏻LMAE👈🏻 value: -2.7748594284057617.\n",
      "Better model found at epoch 34 with 👉🏻LMAE👈🏻 value: -2.7758564949035645.\n",
      "Better model found at epoch 35 with 👉🏻LMAE👈🏻 value: -2.7909886837005615.\n",
      "Better model found at epoch 36 with 👉🏻LMAE👈🏻 value: -2.7910637855529785.\n",
      "Better model found at epoch 37 with 👉🏻LMAE👈🏻 value: -2.8176026344299316.\n",
      "Better model found at epoch 40 with 👉🏻LMAE👈🏻 value: -2.847614288330078.\n",
      "Better model found at epoch 42 with 👉🏻LMAE👈🏻 value: -2.866163730621338.\n",
      "Better model found at epoch 43 with 👉🏻LMAE👈🏻 value: -2.8841166496276855.\n",
      "Better model found at epoch 44 with 👉🏻LMAE👈🏻 value: -2.8888168334960938.\n",
      "Better model found at epoch 46 with 👉🏻LMAE👈🏻 value: -2.8972432613372803.\n",
      "Better model found at epoch 47 with 👉🏻LMAE👈🏻 value: -2.9275474548339844.\n",
      "Better model found at epoch 50 with 👉🏻LMAE👈🏻 value: -2.962419033050537.\n",
      "Better model found at epoch 52 with 👉🏻LMAE👈🏻 value: -2.963674306869507.\n",
      "Better model found at epoch 54 with 👉🏻LMAE👈🏻 value: -2.9787449836730957.\n",
      "Better model found at epoch 55 with 👉🏻LMAE👈🏻 value: -2.9925878047943115.\n",
      "Better model found at epoch 56 with 👉🏻LMAE👈🏻 value: -3.006556272506714.\n",
      "Better model found at epoch 57 with 👉🏻LMAE👈🏻 value: -3.007826328277588.\n",
      "Better model found at epoch 58 with 👉🏻LMAE👈🏻 value: -3.0253632068634033.\n",
      "Better model found at epoch 59 with 👉🏻LMAE👈🏻 value: -3.0311076641082764.\n",
      "Better model found at epoch 60 with 👉🏻LMAE👈🏻 value: -3.0362207889556885.\n",
      "Better model found at epoch 61 with 👉🏻LMAE👈🏻 value: -3.0419363975524902.\n",
      "Better model found at epoch 62 with 👉🏻LMAE👈🏻 value: -3.043233871459961.\n",
      "Better model found at epoch 63 with 👉🏻LMAE👈🏻 value: -3.047004461288452.\n",
      "Better model found at epoch 64 with 👉🏻LMAE👈🏻 value: -3.055849075317383.\n",
      "Better model found at epoch 65 with 👉🏻LMAE👈🏻 value: -3.056790590286255.\n",
      "Better model found at epoch 66 with 👉🏻LMAE👈🏻 value: -3.0605130195617676.\n",
      "Better model found at epoch 68 with 👉🏻LMAE👈🏻 value: -3.0634992122650146.\n",
      "Better model found at epoch 71 with 👉🏻LMAE👈🏻 value: -3.0636985301971436.\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#learner.data.batch_size = 512*2\n",
    "#learner.data.batch_size = 512\n",
    "#learner.opt_func = RAdam\n",
    "#learner.opt_func = partial(torch.optim.Adam, betas=(0.9,0.99), eps=1e-2)\n",
    "learner.fit_one_cycle(72, 1e-4, start_epoch=0)#, moms=(0.75,0.70))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:21:05.284490Z",
     "start_time": "2019-08-21T02:21:04.781955Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAu0AAAEKCAYAAACmF2vAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3Xd8lfX5//HXlU0gIZDBCCMBEiBsCSgCCuJAHCiOuuqoo7+2trbWhdZdtbbaVlv1W0fraBURQVERByBDNgiElRDCSFhJCCGBJGRdvz/OTRtjIAFycp+TXM/H435wzn3u+5N3QE+u3OdzXx9RVYwxxhhjjDG+K8DtAMYYY4wxxpjjs6LdGGOMMcYYH2dFuzHGGGOMMT7OinZjjDHGGGN8nBXtxhhjjDHG+Dgr2o0xxhhjjPFxVrQbY4wxxhjj46xoN8YYY4wxxsdZ0W6MMcYYY4yPC3I7gC+KiYnRhIQEt2MYY8wJW7VqVb6qxrqdoynZe7Yxxp819H3bivY6JCQksHLlSrdjGGPMCRORHW5naGr2nm2M8WcNfd+26THGGGOMMcb4OCvajTHGGGOM8XFWtBtjjDHGGOPjrGg3xhhjjDHGx1nRbowxxhhjjI/zatEuIuNFJF1EMkXkgTpeDxWR953Xl4lIQo3XJjv700XkgvrGFJE7nX0qIjE19ouIvOi8tk5ETvPed2yMMcYYY0zj81rRLiKBwEvAhUAKcK2IpNQ67FbggKr2Av4CPOucmwJcA/QDxgMvi0hgPWN+C5wL1G6bcyGQ5Gx3AK805vdpjDHGGGOMt3mzT/twIFNVswBEZAowEdhY45iJwGPO42nA30VEnP1TVPUIsE1EMp3xONaYqvqds692jonA26qqwFIRiRKRTqq6p1G/W3PKsgtKWLG9gL1FZVRUKuEhgbQKCSQqPJj4qFbEt2tFbJvQuv6NjTHmhHy5YS8Z+4rpEBnG6YnRdIsOdzuSMcYPVFUre4vK2LH/MLlFRyguq6CorJIjFVUM7BLFuSkdvPa1vVm0xwPZNZ7nAKcf6xhVrRSRg0C0s39prXPjncf1jdmQHPHA94p2EbkDz5V4unXrVs+QpjEtyMjj73MzWb69oN5jI0KDSOkcyYD4tpzWvR0je8bQNjy4CVIaY5qTeem5vLf8fz8ahie051fjkhiVFHOcs4wxLU12QQlLsvazNruQtTmFZOw9RHlVdZ3H3nBGN78t2uu6HKoNPOZY++uazlN7zJPJgaq+CrwKkJqaWt+YphEcLK1g8vR1zErbS5d2rXjgwj6M7R1H9+hwggMDKCmvpLS8ioKScnYdKCXnQClbcotJ21XE20t38PqibQQIDOwSxbg+cVw6uDPdo1u7/W0ZY/zAM5MG8ugl/dhZUMK8zbm8vWQHN7yxjMuHxPP7y/rTOtQWDDemJVJVVu04wOz1e5mXnsvWvMMARIQFMahLFLeMTKB7dGu6tQ+nY9swIlsFERkWTGhQgNdnAnjzXSkH6FrjeRdg9zGOyRGRIKAtUFDPufWNeTI5TBPbmneIW99cQc6BUu45P5nbz+pBaFDg946JCAsmIiyYuMgw+nSM/N5rFVXVrMspZEFGPgu25PH8Vxk8/1UGg7pGccVp8Uw6rQtt7IeuMeY4woIDSe4QQXKHCG4emcDL87byt7lb2LSniDduHkZ8VCu3IxpjmkjOgRKmrcphxne72LG/hJCgAM7oEc31p3fnrOQYesS0ISDA3em54pnq7YWBPUV4BjAO2AWsAK5T1Q01jvkFMEBV/5+IXANMUtWrRaQf8C6eeeydgTl4biSVBoy5HUhV1Xzn+UXAncAEPFNpXlTVo/Pj65SamqorV6489b8EU6fM3GKueXUZoLxyw1CGJbQ/5TF3F5byydrdfLRmN5v2FBERGsSVqV24+cwEu/puWhQRWaWqqW7naEqN+Z69cEseP//PaqLCg3n/jhF0tsLdmGZt9c4DvLFwG5+v34MCZ/aMZtKQLlzQv2OTXfxr6Pu214p2J8QE4K9AIPBPVX1KRJ4AVqrqTBEJA94BhuC5wn5NjZtMHwJ+AlQCv1bVz481prP/V8B9QEcgF5ilqrc5N7b+HU8XmhLgFlU97ru7Fe3ek3OghMteWowIvHf7GfSKa9Oo46sq32UX8tbi7cxK20O1wmWD4/nVuF5WvJsWwYr2U7c2u5AbXl9GbEQoM34+0u6bMaYZWpa1n+e/zGD59gIiwoK47vRu3DgiwZVP2HyiaPdXVrR7x+EjlVzxymJ2FZYy/WdnktQhwqtfL7eojFcXZPHO0h1UVitXntaF316QTFxEmFe/rjFusqK9cazYXsB1ry3ljB7R/OvmYQQF2lqExjQHaTkH+dOX6SzIyCMuIpSfjenJ1aldXb2PpaHv2/YuZJqEqnLftHVk7CvmpetO83rBDhAXGcbvLk5h4X1j+fEZ3Zn+XQ7nPDeff8zfSnll3Xd+G2MMwLCE9vz+sv4s3JLPn7/KcDuOMeYUFRwu54EP13HJ3xeRllPIgxP6sOC+sdwyMtFvbjz3j5TG701blcNnaXu4f3wfzkqObdKvHRcZxmOX9uOmMxP4/acbeebzzUxZkc3Tlw9gRM/oJs1ijPEfPxrWjVU7DvB/87dyTp84Uhvh/htjTNOqrlamrMjmj19spriskttHJ/KrcUlEhPnftDe70m68LrughMc/2cjwxPbccVYP13IkxrTmjZuH8eYtw6hW5drXlvLwR+s5fKTStUzGGN/2yCX9iG/XirunrrX3CmP8THZBCde+tpQHZ6TRu0MEs341mocuSvHLgh2saDdepqo8OCMNgD9fPYhAl9slAYzpHcfsu87iJyMT+feyHVzw1wUszdrvdixjjA9qExrE81cNZmdBCX+bm+l2HGNMA6gqU5bvZPxfF7BhdxHPXjGAKXecQe+O3p+a601WtBuv+nz9XhZuyeee85Pp0s53lglvFRLII5ek8MFPRxAcGMB1ry3lr19nUFVtN2YbY75veGJ7rhrahdcXZpGZe8jtOMaY4zhYUsHtb6/igelpDOjSls/vGs2PhnXz+sJHTcGKduM1h49U8uSnG0npFMkNZ3R3O06dUhPa8+kvR3HZ4Hj++vUWrn99KfuKytyOZYzxMfdf2IfwkEAem7kB67pmjG9al1PIRX9byPyMXH53UV/eve0Murb3nQuGp8qKduM1r3yzlT0Hy3jysv4+3S6tdWgQf/7RYJ67ahBrsw8y4YWFrNhe4HYsY4wPiWkTym/OS2ZRZj7zM/LcjmOMqUFV+ffSHVz5yhKqq5WpPx3BbaN7uL6CaWPz3UrK+LXc4jLeWLSNSwZ1Zmj3dm7HaZArh3bhk1+OpG2rYK57bSlTlu90O5Ixxodcf3p3urRrxZ++SKfaptIZ4xPKK6uZPD2N3320njN7RfPZr0YzpJt/1B0nyop24xUvzc2kvKqau89LdjvKCekVF8GMn49kRM8YHpiexmMzN1BRZT3djTEQEhTAb85NZsPuIj5fv9ftOMa0eIUl5dz0z+VMWZHNnWN78c+bhtGudYjbsbzGinbT6LILSnh3+U6uTu1KYkxrt+OcsLbhwfzzplRuH53Im4u3c9tbK63VmzEGgMuGxJPcoQ3Pf5VuN64b46Jt+YeZ9PJiVu04wJ+vHsQ9F/RudtNharOi3TS6v83dQoAId41LcjvKSQsKDOChi1L4w6QBLNySx3WvLWX/oSNuxzLGb4jIeBFJF5FMEXmgjte7i8gcEVknIt+ISJdar0eKyC4R+XvTpa5fYIBw17hksvIO8+UGu9pujBu+23mAy1/+lsLSCv5z++lMOq1L/Sc1A1a0m0a152ApM77bxTXDutKxbZjbcU7ZNcO78eqPU0nfV8wVryxm5/4StyMZ4/NEJBB4CbgQSAGuFZGUWoc9B7ytqgOBJ4Bnar3+JDDf21lPxvj+HUmIDueV+Vutk4wxTWzRlnyuf30ZkWHBzPj5mQxrQSsVW9FuGtUbC7dRrXDbaPdWPm1s56Z04D+3nUFhaQWTXlnM5r1FbkcyxtcNBzJVNUtVy4EpwMRax6QAc5zH82q+LiJDgQ7Al02Q9YQFBgg/Pbsn63IOsnirLcxmTFP5PG0PP3lzBd3ahzPt/42ge7T/TcE9FVa0m0ZTWFLOe8t3csnATs2qLyrA0O7tmPb/ziQoQLj21aVs2H3Q7UjG+LJ4ILvG8xxnX01rgSucx5cDESISLSIBwPPAvV5PeQomnRZPXEQoL39jq6Qa0xSmrsjmF++uZkCXtrx/xwjiIv3/0/wTZUW7aTTvLNnB4fIq/t+Ynm5H8YpecW14/6dnEB4SxHWvLWNdTqHbkYzxVXXdDVZ7Hsk9wNki8h1wNrALqAR+DsxS1WyOQ0TuEJGVIrIyL6/p+6aHBgVyy8hEvs3cb5++GeNlU5bv5L4P1zEqKZZ3bh1O2/BgtyO5wop20yiOVFbx1pLtjOkdS5+OkW7H8Zru0a2ZcscZRIQFcf1ry1i984DbkYzxRTlA1xrPuwC7ax6gqrtVdZKqDgEecvYdBEYAd4rIdjzz3m8UkT/U/gKq+qqqpqpqamxsrJe+jeO7ZlhXQoMCeGfJDle+vjEtwdQV2TwwPY0xvWN59cdDCQ8JcjuSa6xoN41i9vq95B8q55aRiW5H8bqu7cOZ+tMRtG8Two1vLGdttl1xN6aWFUCSiCSKSAhwDTCz5gEiEuNMhQGYDPwTQFWvV9VuqpqA52r826r6g+4zvqBd6xAuHdSZ6at3cbC0wu04xjQ7U1dmc//0dZyVHMv/3TCUsOBAtyO5yop20yjeXrKDhOhwRveKcTtKk+gc1Yr37xhBu9bB3PSv5aTvLXY7kjE+Q1UrgTuBL4BNwFRV3SAiT4jIpc5hY4B0EcnAc9PpU66EPUU3nZlAaUUVH67KcTuKMc3Kh6tyuP/DdYzqFcOrP7aCHaxoN41g/a6DrNpxgBvO6N7sFzaoqWPbMP5z6xmEBgVw/evL2JZ/2O1IxvgMVZ2lqsmq2lNVn3L2PaKqM53H01Q1yTnmNlX9wUIIqvqmqt7Z1NlPRP/4tpzWLYp/L91BtS22ZEyj+GLDXu6dtpaRPWN47cZUK9gdVrSbU/bOkh2EBQdw1dCu9R/czHSLDufft55OtSo3vL6MXYWlbkcyxjSxG0ckkJV/2No/GtMIlmzdzy/f+46BXaJ49Ua7wl6TFe3mlBwsqeDjtbu4fEh8i72bO6lDBG//ZDhFpRXc8PoyWznVmBZmfP+ORIYF8cGq4za8McbUY/2ug9z+9kq6tw/nXzcPa9E3ndbFinZzSmZ8l0NZRTXXn97d7Siu6h/fln/dMozdhaXc9vZKSsur3I5kjGkiYcGBXDYkntnr99oNqcacpKy8Q9z0z+W0bRXM27cOp13rELcj+Rwr2s0p+WBVDv06R9I/vq3bUVyXmtCeF64ZzJrsQu6a8h1VNr/VmBbjqqFdOVJZzSdrd9d/sDHme3KLy/jxG8tR4J1bh9OpbSu3I/kkK9rNSdu4u4gNu4u4amgXt6P4jPH9O/HoxSl8uXEfj3+yAVUr3I1pCfrHR9KnYwQfrLQpMsaciNLyKm5/ayUFh8t585Zh9Iht43Ykn2VFuzlpH6zKJiQwgImDa69O3rLdPDKR20cn8vaSHby6IMvtOMaYJiAiXJ3albU5B60FrDENVF2t3D11Det2HeSFawYzsEuU25F8mhXt5qSUV1bz8ZrdnJfSwead1WHyhX25aGAnnvl8M5+n7XE7jjGmCVw2JJ7gQLGr7cY00LNfbObz9Xt5aEJfzu/X0e04Ps+KdnNS5m7eR8Hhcq5MtakxdQkIEJ6/ahCndYvi7qlrWb/roNuRjDFe1r51CGN6x/HJut12T4sx9Xhv+U7+MT+L60/vxq2jmv9q6o3BinZzUj5YmUOHyFDOSop1O4rPCgsO5P9+PJSo8GDueHslecXWCtKY5m7i4M7sKzrCsm3Ws92YY/k2M5/ffbSes5JjefzSfoi0nIUZT4UV7eaE7T90hG8y8rh8SBcCW9AKqCcjLiKM125M5UBJBT99ZyVHKq0VpDHN2bg+HWgdEmhdZIw5huyCEn7x7mp6xLTmpeuGEBRopWhD2d+UOWGz0vZQVa1cNqSz21H8Qv/4tjx/9SBW7yzkoRnrraOMMc1Yq5BAzkvpwKy0vZRXVrsdxxifUlJeye1vr6S6WnntxlQiwlrmoowny6tFu4iMF5F0EckUkQfqeD1URN53Xl8mIgk1Xpvs7E8XkQvqG1NEEp0xtjhjhjj7u4nIPBH5TkTWicgEb37PLcHMtbtJ7tCGPh0j3Y7iNyYM6MRd45KYtiqHNxZtczuOMcaLJg6O52BpBQsy8tyOYozPUFXum7aO9H3FvHjtEBJiWrsdye94rWgXkUDgJeBCIAW4VkRSah12K3BAVXsBfwGedc5NAa4B+gHjgZdFJLCeMZ8F/qKqScABZ2yA3wFTVXWIM+bL3vh+W4pdhaWs2H6ASwbaVfYTdde4JC7s35FnPt/Mkq0239WY5mpUUgztwoOZaVNkjPmvfyzI4tN1e7j3gt6M6R3ndhy/5M0r7cOBTFXNUtVyYAowsdYxE4G3nMfTgHHiuRthIjBFVY+o6jYg0xmvzjGdc85xxsAZ8zLnsQJHLwm3Bexd9BR8ts7z13fJICvaT1RAgPCnqwbRPTqcX763mr0Hy9yOZIzxguDAAC4c0ImvNu6jpLzS7TjGuG5BRh5/nL2ZiwZ04mdn93Q7jt/yZtEeD9RsVpvj7KvzGFWtBA4C0cc591j7o4FCZ4zaX+sx4AYRyQFmAb88lW+qpZu5djeDurS1j7VOUpvQIP5xw1BKyqv4xburbc6rMc3UpYM6U1pRxZxNuW5HMcZV2QUl/PK970juEMGfrhponWJOgTeL9rr+VWrfgXesYxprP8C1wJuq2gWYALwjIj/4vkXkDhFZKSIr8/JsHmJdsvIOsX5XkV1lP0VJHSJ49oqBrNpxgKdnbXI7jjHGC4YltCemTQizN+x1O4oxrjlS6blAVa3KP348lPCQILcj+TVvFu05QNcaz7vww6kp/z1GRILwTF8pOM65x9qfD0Q5Y9T+WrcCUwFUdQkQBsTUDquqr6pqqqqmxsZa7/G6zFy7GxGbGtMYLhnUmZ+MTOTNxdv5eM0ut+MYYxpZYIBwfr+OzNucS1mFtXo1LdPTn21iXc5BnrtqEN2j7RP6U+XNon0FkOR0dQnBcxPozFrHzARuch5fCcxVTz+8mcA1TneZRCAJWH6sMZ1z5jlj4Iz5sfN4JzAOQET64ina7VL6CVJVPlm7m9MT29MhMsztOM3C5Al9GJbQjgc+TCNjX7HbcYwxjWx8v46UlFexcEu+21GMaXKfrdvDW0t2cNuoRC7o19HtOM2C14p2Z375ncAXwCY8HVw2iMgTInKpc9gbQLSIZAJ3Aw84527Ac3V8IzAb+IWqVh1rTGes+4G7nbGinbEBfgvcLiJrgfeAm9UaZZ+wLbmH2Jp3mIusa0yjCQ4M4KXrTqN1aBC/+M9qSsvtapwxzcmIntFEhgXx+fo9bkcxpkltyz/M/R+uY0i3KO6/sI/bcZoNr04uUtVZeG7+rLnvkRqPy4CrjnHuU8BTDRnT2Z+Fp7tM7f0bgZEnmt183+z1exGBC1I6uB2lWYmLDOMvPxrEjf9czhOfbuCZSQPdjmSMaSTBgQGcm9KBrzfuo7yympAgW8/QNH9lFVX84j+rCQoU/n7daQTbiqeNxv4mTYPMXr+Xod3aEWdTYxrd6KRYfnZ2T95bnm1LnxvTzFzYvxNFZZUszbK1GUzL8PgnG9m4p4i/XD2Y+KhWbsdpVqxoN/Xaub+EjXuKGN/f5qR5y2/OS+a0blFMnp7Gzv0lbscxxjSS0UkxhIcE8vl66yJjmr+Za3fz3vKd/GxMT8b2sQWUGpsV7aZeszd45mPajSTeExwYwIvXDiFA4M73rH+7Mc1FWHAgY/vE8dXGvVRV2+1UpvnKLijhoRlpnNYtit+el+x2nGbJinZTr9nr99KvcyRd24e7HaVZ69IunD9eOZB1OQf50xeb3Y5jjGkkF/bvSP6hclbtOOB2FGO8orKqmt+8vwYUXrhmCEE2j90r7G/VHNe+ojJW7yzkQpsa0yTG9+/EjSO689rCbczbbCspGtMcnJ0cS3CgMGfTPrejGOMVL83bysodB/j95f3tAp8XWdFujutLZzU/m8/edB6c0Jc+HSO4d9pa8g8dcTuOMeYURYQFc3piNF9b0W6aoVU7CnhhTgaXD4ln4uB4t+M0a1a0m+OavWEvPWNb0ysuwu0oLUZYcCB/vWYwRWWVPPBhGrasgDH+b1zfOLbmHWZ7/mG3oxjTaIrKKrhryhri27XiiYn93I7T7FnRbo7pwOFylmYV2FV2F/TpGMl9F/Tm6037mLIi2+04xphTdG5fzxoXdrXdNCcPf7SePQfL+OuPhhARFux2nGbPinZzTF9v2kdVtVrXGJf8ZGQiI3tF8+SnG+3qnDF+rmv7cJI7tGHOJrtXxTQPM77L4eM1u7lrXBJDu7dzO06LYEW7OaY5m3LpGBnGgPi2bkdpkQIChOeuGkRQgPDr99dQWWVtII3xZ+P6dmDF9gIOlla4HcWYU5JzoISHP9rAsIR2/GJsL7fjtBhWtJs6HamsYuGWPM7pG4eIuB2nxerUthVPTxrAmuxC/j4v0+04xphTcG7fOCqrlfkZeW5HMeakVVcr936wDlXlz1cPJjDAaoSmYkW7qdOyrAIOl1cxzlY0c93FAztz+ZB4/jY3k+92Wp9n4x9EZLyIpItIpog8UMfr3UVkjoisE5FvRKSLs3+wiCwRkQ3Oaz9q+vTeMbhrO9q3DrHWj8avvb1kO0uy9vPwxSnW3rGJWdFu6jRn0z7CggMY2SvG7SgGeHxiPzpGhvGb99dQUl7pdhxjjktEAoGXgAuBFOBaEUmpddhzwNuqOhB4AnjG2V8C3Kiq/YDxwF9FJKppkntXYIAwtncc36Tn2XQ345ey8g7xh9mbGdM7lh8N6+p2nBbHinbzA6rK15tyGdUrhrDgQLfjGCAyLJjnrx7EjoIS/jg73e04xtRnOJCpqlmqWg5MASbWOiYFmOM8nnf0dVXNUNUtzuPdQC4Q2ySpm8C5feM4WFrBSlsd1fiZyqpq7p66ltCgQJ69YqBNnXWBFe3mBzL2HWJXYSnn9OngdhRTwxk9orn5zATeXLydJVv3ux3HmOOJB2r2Ks1x9tW0FrjCeXw5ECEi0TUPEJHhQAiw1Us5m9zo5FhCAgOYayseGz/zjwVZrMku5MnL+tMhMsztOC2SFe3mB472ER7X1+az+5r7LuhDQnQ4905by+EjNk3G+Ky6LsHVXiXsHuBsEfkOOBvYBfz3P2oR6QS8A9yiqj+YSyIid4jIShFZmZfnPzd2tgkNYlhiO+an+09mYzbuLuKvX2dw0cBOXDqos9txWiwr2s0PzNm0jwHxbe03aR/UKiSQ564axK7CUp75fJPbcYw5lhyg5oTXLsDumgeo6m5VnaSqQ4CHnH0HAUQkEvgM+J2qLq3rC6jqq6qaqqqpsbH+NXtmTHIc6fuK2V1Y6nYUY+p1pLKKu6euoW2rEJ6c2N/tOC2aFe3me/YfOsJ32YV2ld2HpSa059aRifx76U4Wbcl3O44xdVkBJIlIooiEANcAM2seICIxInL0Z9Bk4J/O/hBgBp6bVD9owsxN5uzenl8yFljrR+MHXvh6C5v3FvPsFQNo3zrE7TgtmhXt5nvmpeehCuNsPrtPu+eC3vSIac39H66juMwWajG+RVUrgTuBL4BNwFRV3SAiT4jIpc5hY4B0EckAOgBPOfuvBs4CbhaRNc42uGm/A+9KimtD57ZhfGNTZIyPW5NdyP/N38rVqV0Y19fqArdZ0W6+Z86mfXSIDKV/fKTbUcxxhAUH8tzVg9hzsJSnZ9k0GeN7VHWWqiarak9VfcrZ94iqznQeT1PVJOeY21T1iLP/36oarKqDa2xr3PxeGpuIcHbvWL7NzKfCWj8aH3Wksop7P1hLh8gwfndx7Y6txg0NKtpFZJSI3OI8jhWRRO/GMm4or6xmQUYe5/TpYK2c/MBp3dpx+1k9eG95tn3MboyfOTs5juIjlay21o/GR700N5MtuYd4+vIBRIYFux3H0ICiXUQeBe7HM+cQIBj4tzdDGXcs27bfVkH1M785N5lecW24/8N1FNk0GWP8xshe0QQFCN/YL9zGB23cXcTL32xl0pB4xlpN4DMacqX9cuBS4DD8d7GLCG+GMu6YuzmX0CBbBdWfhAUH8vxVg9hXVMbTn9k0GWP8RURYMEO7W+tH43sqqqq5d9paosJDeOQSmxbjSxpStJerquL02BWR1t6NZNwyPyOPM3pE0yrEVkH1J4O6RnH76B5MWZHN4q3WTcYYfzGmdxwb9xSRW1TmdhRj/uvVBVls2F3EkxP7ERVu3WJ8SUOK9qki8g8gSkRuB74GXvduLNPUsgtKyMo7zNnJ/tXv2Hj8+txkEqLDmTw9jdLyKrfjGGMa4Oj77XybImN8RGZuMS/M2cKEAR25cEAnt+OYWuot2lX1OWAa8CHQG3hEVV/0djDTtI7OqzzaP9j4l1YhgTwzaSA79pfwl68z3I5jjGmAvp0iiIsItXntxidUVSv3TVtHeEggj19qiyj5oobciPqsqn6lqveq6j2q+pWIPNsU4UzTmZ+eR9f2regRY7Of/NWIntFcO7wbry/MYm12odtxjDH1EBHOTo5lYUYeldb60bjsrcXbWb2zkEcvSSE2ItTtOKYODZkec14d+y5s7CDGPUcqq1i8NZ+zk2Ot1aOfmzyhD7ERodz/4TrKK60IMMbXjekdR1FZJWtz7Bdt456d+0v40xfpnNMnjssGx7sdxxzDMYt2EfmZiKQBvUVkXY1tG7Cu6SIab1u1/QAl5VWcnWxtnfxdZFgwv79sAJv3FvOP+VvdjmOMqceoXjEECLY6qnGNqnL/h+sIChCeury/XbzzYce70v4ucAkw0/nz6DZUVW9ogmymiczPyCM4UDizZ7TbUUwjOC+lAxcN7MTf5maSmVvsdhzj50QkVURmiMhq58JNmojYhZtYorbSAAAgAElEQVRG0jY8mMFdo1i4xTo/GXe8tzybJVn7efCivnRq28rtOOY4jlm0q+pBVd2uqteq6g6gFE/bxzYi0q3JEhqv+yY9j2EJ7WkdGuR2FNNIHrukH+Ghgdw3bR1V1ep2HOPf/gP8C7gCz4Wbi50/TSMZlRTLupxCDpbYAmmmaeUWlfHM55sY0SOaa4Z1dTuOqUdDbkS9RES2ANuA+cB24POGDC4i40UkXUQyReSBOl4PFZH3ndeXiUhCjdcmO/vTReSC+sYUkURnjC3OmCE1XrtaRDaKyAYRebch2VuKPQdLSd9XbK0em5nYiFAeuTiF1TsLeWfJdrfjGP+Wp6ozVXWbqu44urkdqjkZnRRDtWLrLJgm9/gnGzlSWc3TkwbYtBg/0JAbUX8PnAFkqGoiMA74tr6TRCQQeAnPTaspwLUiUntprVuBA6raC/gL8KxzbgpwDdAPGA+8LCKB9Yz5LPAXVU0CDjhjIyJJwGRgpKr2A37dgO+5xVjgtBob09vmszc3lw+J56zkWP74RTo5B0rcjmP816Mi8rqIXCsik45ubodqTgZ3jaJNaBALM61oN01n7uZ9fJa2h1+O7UWidY7zCw0p2itUdT8QICIBqjoPGNyA84YDmaqaparlwBRgYq1jJgJvOY+nAePE86veRGCKqh5R1W1ApjNenWM655zjjIEz5mXO49uBl1T1AICq5jYge4vxTXoeHSPDSO7Qxu0oppGJCE9f7um1++CM9XgWNjbmhN2C5z1/PP+7t+liVxM1M8GBAZzRI5qFW+xmVNM0Dh+p5OGPNpAU14afnt3T7TimgRpStBeKSBtgAfAfEXkBqGzAefFAdo3nOc6+Oo9R1UrgIBB9nHOPtT8aKHTGqP21koFkEflWRJaKyPi6worIHSKyUkRW5uW1jDfOiqpqFm2xVo/NWZd24dx7QW8WZOQxc+1ut+MY/zRIVVNV9SZVvcXZfuJ2qOZmdFIM2QWl7Nh/2O0opgX4y1cZ7Cos5elJAwgJakgpaHxBQ/6lJgIlwG+A2cBWGnYTUl1VYO1Lfcc6prH2AwQBScAY4FrgdRGJ+sHBqq86P5hSY2NbxvzuNdmFFB+pZIytgtqs3TgigUFd2vLkpxspLCl3O47xP0vrmNpoGtnopBgA6yJjvG79roP889ttXHd6N4YltHc7jjkB9RbtqnpYVatVtVJV38Izp7zOq9W15AA1b0XuAtS+1PffY0QkCGgLFBzn3GPtzweinDFqf60c4GNVrXCm2qTjKeJbvG/ScwkMEM7sFeN2FONFgQHC05MGcKCkgj98vtntOMb/jALWOA0ArOWjlyTGtCY+qpVNkTFeVVlVzeTpaUS3CeX+8X3cjmNO0PEWV4p0Orj8XUTOF487gSzg6gaMvQJIcrq6hOC5sXRmrWNmAjc5j68E5qpn4u1M4Bqnu0winiJ7+bHGdM6Z54yBM+bHzuOPgLHO9xSDZ7pMVgPyN3vzM/I4rVsUbVsFux3FeFm/zm25bVQiU1Zksyxrv9txjH8Zj+c9+Hys5aPXiAijk2JYvHU/lVW2mrHxjreW7CBt10EevSTFfvb7oeNdaX8H6A2kAbcBXwJXARNVtfYNpT/gzC+/E/gC2ARMVdUNIvKEiFzqHPYGEC0imcDdwAPOuRuAqcBGPFNyfqGqVcca0xnrfuBuZ6xoZ2ycY/eLyEY8hf29zo21LVpe8RHW7yqyrjEtyF3nJtGlXSsmz0jjSGWV23GM/9BjbKaRjUqKobisknW7DrodxTRDuwpLef7LdMb2juWiAZ3cjmNOwvFW0+mhqgMAROR1PFNQuqlqg5dYVNVZwKxa+x6p8bgMzy8CdZ37FPBUQ8Z09mfh6S5Te7/i+YXg7obmbgmOtnq0/uwtR3hIEL+/rD83/2sFr3yzlV+fm+x2JOMfPuN/9w6FAYl4phn2czNUczSyZwwisDAjn9O6tXM7jmlGVJVHP16PKjwxsb81n/BTx7vS/t+l2VS1Cth2IgW78W3zM/KIaRNCSqdIt6OYJjSmdxyXDOrMy/O2kpl7yO04xg+o6gBVHej8mYTn4sgit3M1R+1ahzAgvi2LMm1eu2lcs9fv5etNudx9XjJd24e7HcecpOMV7YNEpMjZioGBRx+LSFFTBTSNr6paWbglj7OSYwkIsN+2W5pHLk4hLDiAh2akWe92c8JUdTUwzO0czdWoXjGs3llIcVlF/Qcb0wBFZRU8OnMD/TpHcsvIBLfjmFNwzKJdVQNVNdLZIlQ1qMZjuzzrx9blFHKgpMKmxrRQsRGhTJ7Ql2XbCvhgZY7bcYyPE5G7a2z3iMi7gF0K9pJRSTFUVStLswrcjmKaiT/NTif/0BGemTSAoEDrye7P7F+vBZqfkYcIjE6yor2l+lFqV4YltOOpWZvIP3TE7TjGt0XU2ELxzHGvtxmBOTlDu7ejVXAgi6z1o2kEq3Yc4N/LdnDTmQkM7PKDJWqMn7GivQWan5HHoC5RtG8d4nYU45KAAOGZSQMoKa/k959udDuO8W0bVfVxZ3tKVf+DtXz0mtCgQE7v0d4WWTKnrKKqmgenp9ExMozfnt/b7TimEVjR3sIcOFzOmuxCmxpj6BUXwc/O7slHa3bbgi7meCY3cJ9pJKN6xZCVf5hdhaVuRzF+7NUFWaTvK+aJif1pE3q8ZoHGX1jR3sIszMxHFc7ubUW7gZ+P7UWPmNY8NGM9peXWu938j4hcKCJ/A+JF5MUa25tApcvxmrWznIsqNkXGnKwd+w/z4pwtjO/XkfNSOrgdxzSSeov2o91iam3ZIjJDRHo0RUjTeOan5xEVHswgm9tmgLDgQJ66fAA7C0p4ce4Wt+MY37IbWAmUAatqbDOBC1zM1ewlxbWhQ2QoC2yKjDkJqspDM9YTHBjAY5facgrNSUM+L/kznjfvd/EsrnEN0BHP4hr/BMZ4K5xpXNXVyvyMPEYnxRJorR6NY0TPaK4a2oXXFmQxcXBn+nS05lAGVHUtsFZE3lVV6z/YhESEUb1imbN5H1XVau/X5oR8tGYXizLzeWJiPzq2DXM7jmlEDZkeM15V/6GqxapapKqvAhNU9X3AlmzzIxv3FJF/6IjNZzc/8OCEvkS2Cmby9DSqqq13u/me4SLylYhkiEiWiGwTkSy3QzV3o5NiKCypYMPug25HMX7kwOFynvx0E4O7RnH96d3djmMaWUOK9moRuVpEApzt6hqv2U93PzI/wzM/8qzkGJeTGF/TrnUID1/cl+92FvLush1uxzG+5Q08n7iOwrOoUioNWFxJRMaLSLqIZIrIA3W83l1E5ojIOhH5RkS61HjtJhHZ4mw3NeL34jdG9vK8T1sXGXMinp61iaLSCp6ZNMA+oWmGGlK0Xw/8GMgF9jmPbxCRVsCdXsxmGtn8jDz6dY4kLsI+LjM/dNngeEYnxfDH2ensKypzO47xHQdV9XNVzVXV/Ue3450gIoHAS8CFQApwrYik1DrsOeBtVR0IPAE845zbHngUOB0YDjwqIi3uU93YiFD6dopkkRXtpoGWbN3PB6tyuG10D/p2smmOzVG9RbuqZqnqJaoao6qxzuNMVS1V1UVNEdKcuqKyClbvOGBTY8wxiQi/v6w/5VXVPDZzg9txjO+YJyJ/EpERInLa0a2ec4YDmc7Pj3JgCj9ckCkFmHP0a9R4/QLgK1UtUNUDwFfA+Mb5VvzL6KQYVu4ooKTcmvWY4yurqOKhGWl0bd+Ku8YluR3HeEm9N6KKSCxwO5BQ83hV/Yn3YpnGtjhzP5XVakW7Oa7u0a351bgk/vRFOl9v3Me51irMeK54g2dazFEKnHOcc+KB7BrPc2qMc9Ra4ArgBeByIEJEoo9xbvyJx/Z/o3rF8OqCLJZtK2Bs7zi34xgf9vI3W8nKP8zbPxlOq5BAt+MYL2lI95iPgYXA14A1cvZT8zPyaBMaxGndW9ynzOYE3XFWD2au2c0jH69nRM9oWtuiHC2aqo49idPqmkxb+x6oe4C/i8jNwAJgF57+7w05FxG5A7gDoFu3bicR0fcNT2xPSFAAi7bkW9Fujikzt5hXvslk4uDO/+3xb5qnhsxpD1fV+1V1qqp+eHTzejLTaFSVBRl5jOwVTXCgradlji84MICnJw1gT1EZz3+Z4XYc4zIR6SAib4jI587zFBG5tZ7TcoCuNZ53wdM6+L9UdbeqTlLVIcBDzr6DDTnXOfZVVU1V1dTY2OZZqIQFBzI8ob2tWGyOqbpaeXD6esJDgnj44tq3jZjmpiEV3KciMsHrSYzXbM07xK7CUvsN3DTY0O7tuP70bry5eBtpOdZyroV7E/gC6Ow8zwB+Xc85K4AkEUkUkRA863vMrHmAiMSIyNGfQZPxrPuB87XOF5F2zg2o5zv7WqRRSTFk7DtkN4ebOk1dmc3y7QU8OKEPMW1C3Y5jvKwhRftdeAr3Umc11GIRKfJ2MNN4vkl3Wj0mWdFuGu6+8Z4fAg9MX0dlVbXbcYx7YlR1KlANoKqV1DNV0jnmTjzF9iZgqqpuEJEnRORS57AxQLqIZAAdgKeccwuAJ/EU/iuAJ5x9LdLoJGv9aOqWV3yEp2dtYnhie65O7Vr/CcbvNaR7TISqBqhqK1WNdJ5bLyE/Mj8jj56xrenaPtztKMaPRIYF89il/diwu4h/fbvd7TjGPYedG0QVQETOAOr9+EVVZ6lqsqr2VNWjBfkjqjrTeTxNVZOcY25T1SM1zv2nqvZytn9559vyD307RhLdOoRFNkXG1PLkpxspq6jm6csHIGI92VuCY95hJiJ9VHXzsVp7qepq78UyjaWsoorl2wpsZTRzUi7s35FxfeL481cZjO/f0X7xa5nuxjO1paeIfAvEAle6G6nlCAgQRiXFsChzP9XVSoAtmGOAb9Jzmbl2N3eNS6JXXBu345gmcrwr7Xc7fz5fx/acl3OZRrI0az9HKqs5u7dNjTEnTkR44rL+iMAjH69H1RZBbmmcCzRnA2cCPwX6qeo6d1O1LKN6xZB/6Aib9xa7HcX4gJLySn730Xp6xrbm52N7uh3HNKFjXmlX1TucP0+m3ZfxEfMz8ggNCuD0xPZuRzF+Kj6qFXefl8zvP9vEZ2l7uHhg5/pPMs2Gs7rpBP63Vsf5IoKq/tnVYC3IaOd+pEWZeaR0ttmpLd0LX28h50Ap799xBqFB1pO9JWlQ/z8ROVNErhORG49u3g5mGsf8jDzO6BFNWLD9j21O3s1nJtA/PpLHP9nIwdIKt+OYpvUJcDMQDUTU2EwT6dg2jKS4NnYzqmHD7oO8vmgbP0rtyuk9ot2OY5pYQ1ZEfQfoCazhfx0DFHjbi7lMI8guKCEr7zA32Hx2c4qCAgP4w6SBXPr3RTw7ezNPXz7A7Uim6XRR1YFuh2jpRiXF8O6ynZRVVNlFmBaqqlp5cHoa7cKDmTyhj9txjAsacqU9FRipqj9X1V8626+8HcycuvkZnm4DNp/dNIb+8W25ZWQi7y7bycrtLbYDX0v0uYic73aIlm50UgxHKqtZuf2A21GMS95esp21OQd5+OIUosJD3I5jXNCQon090NHbQUzjm5+RR5d2regR09rtKKaZuPu8ZOKjWjF5ehrllda7vYVYCsywtTrcdXpiNMGBwsJMa/3YEu0uLOW5L9I5KzmWSwfZfUUtVUOK9hhgo4h8ISIzj27eDmZOTXllNYsz8zk7Odb6t5pG0zo0iCcm9mNL7iFeXbDV7TimaTwPjADCba0O97QODeK0bu1YmGHz2luiR2duoEqVpy7rbz/TW7B657QDj3k7hGl8q3Yc4HB5FWcn29QY07jG9e3AhAEdeXFuJhcN7EyifZLT3G0B1qv1+3Td6KQYnvsyg/xDR2zJ+hZk9vq9fLVxHw9c2MfWymjhjnul3Wn19bCqzq+9NVE+c5LmZ+QRFCCc2SvG7SimGXr0kn6EBgbw0Iw0693e/O0BvhGRySJy99HN7VAt0dHWj99m2tX2lqK4rILHZm6gT8cIbh2V6HYc47LjFu2qWgWUiEjbJspjGsn8jDxSE9rRJrQhH6YYc2I6RIZx34V9WLx1P9NX73I7jvGubcAcIARr+eiq/vFtadsq2Fo/tiDPfZHOvuIy/nDFQIIDG9Sl2zRjDfkvoAxIE5E3ROTFo1tDBheR8SKSLiKZIvJAHa+Hisj7zuvLRCShxmuTnf3pInJBfWOKSKIzxhZnzJBaX+tKEVERSW1Idn+2r6iMTXuKODs5zu0ophm7fng3TusWxe8/20jB4XK34xgvUdXHVfVx4M/A8zWemyYWGCCM6hXDoi359glXC7B65wHeXrqDm0YkMLhrlNtxjA9oSNH+GfAwsABYVWM7LmdqzUvAhUAKcK2IpNQ67FbggKr2Av4CPOucmwJcA/QDxgMvi0hgPWM+C/xFVZOAA87YR7NEAL8CljXg+/V7C462erT57MaLAgKEZyYNpLiskqc+2+R2HOMlItJfRL7D00lsg4isEpF+budqqUYlxbC3qIzM3ENuRzFeVFFVzYPT0+gQEcZvz092O47xEfUW7ar6Vl1bA8YeDmSqapaqlgNTgIm1jpkIHB1rGjBOPLdFTwSmqOoRVd0GZDrj1Tmmc845zhg4Y15W4+s8CfwRz6cGzd436XnERYTSt5N9gm28q3fHCO44qwcfrs5hsc2zba5eBe5W1e6q2h34LfCay5larFHOfUo2RaZ5e33hNjbvLebxif2ICAt2O47xEfUW7SKSJCLTRGSjiGQd3RowdjyQXeN5jrOvzmNUtRI4iGep7GOde6z90UChM8b3vpaIDAG6quqnDcjs9yqqqlmQkcfY3nHWFso0iV+NS6J7dDgPfbSesoqq+k8w/qa1qs47+kRVvwGsZZBLurYPJzGmNYvsl+Rma+f+El6Yk8H5KR24oJ8tk2P+pyHTY/4FvAJUAmOBt4F3GnBeXRVj7Ul4xzqmUfaLSACeaTe/PU5OTxCRO0RkpYiszMvz38UrVm4/QPGRSsb2sfnspmmEBQfy1GUD2JZ/mJfmZbodxzS+LBF5WEQSnO13eG5ONS4Z1SuGpVn7bYGzZkhVeeijNIICAnh8os1CM9/XkKK9larOAURVd6jqY3imotQnB+ha43kXYPexjhGRIKAtUHCcc4+1Px+IcsaouT8C6I+nXdl24AxgZl03o6rqq6qaqqqpsbH+Oxd8XnouwYHCqCRr9WiazqikGC4fEs//zd/Kln3FbscxjesnQCwwHZjhPL7F1UQt3KikGErKq1i984DbUUwj+3D1LhZuyee+8b3p1LaV23GMj2lQ9xjnivUWEblTRC4HGnIZdwWQ5HR1CcFzY2ntlVRnAjc5j68E5joLeMwErnG6yyQCScDyY43pnDPPGQNnzI9V9aCqxqhqgqom4FmO+1JVXdmA/H5p7uZcTk+MtlaPpsn97qK+tA4NYvL0NKqrrbNFc6GqB1T1V6p6mqoOUdW7VNWqRReN6BlNYICwyOa1Nyu5xWU8+elGUru344bTu7sdx/ighhTtvwbC8XRfGQrcwP8K7WNy5pffCXwBbAKmquoGEXlCRC51DnsDiBaRTOBu4AHn3A3AVGAjMBv4hapWHWtMZ6z7gbudsaKdsVuU7IISMnMP2dQY44roNqE8OKEvK3cc4N3lO92OY06RiMw83uZ2vpYsMiyYwV2jWLjFf6dymh96bOYGSiuqePbKgQQE2D1p5ofqvRyrqisARERV9YQ+ElXVWcCsWvseqfG4DLjqGOc+BTzVkDGd/Vl4usscL8+YhuT2V3M35wJwjhXtxiVXDe3CzDW7+cPnmxnbJ474KPt414+NwHPj/3t42uVaFeFDRifF8MKcLRSWlBMVHlL/CcanzV6/h1lpe7n3gt70jG3jdhzjoxrSPWaEiGzEc2UbERkkIi97PZk5YXM355IY05rEGGvsYNwhIjwzaQBV1cqD09NsARj/1hF4EM99QS8A5wH5qjpfVee7mswwOikGVVi8db/bUcwpOlhSwcMfbyClUyR3nNXD7TjGhzVkesxfgQuA/QCquhY4y5uhzIkrKa9kSdZ+xva2q+zGXV3bh3Pf+N7Mz8hj+updbscxJ8mZkjhbVW/CcxN/Jp6b+n/pcjQDDOoSRURokE2RaQaenrWJgsPl/PHKgQQHNqQsMy1Vg/7rUNXsWrusGbOPWZzpaf9lU2OML7hpRAKp3dvxxKcbyS1uEWuaNUtOM4BJwL+BXwAv4ukiY1wWFBjAiJ7RLMjIt0+0/Ni3mfm8vzKbO87qQf/4tm7HMT6uIUV7toiciafveYiI3IMzVcb4jrnpubQOCWR4Ynu3oxhDQIDw7JUDKa2o4pGPNtR/gvE5IvIWsBg4DXhcVYep6pOqah+f+IjRSTHsKixl+/4St6OYk1BSXskD09fRI6Y1d41LcjuO8QMNKdr/H54rLPF4+qQPBn7uzVDmxKgq8zbnMiophpAg+2jN+IaesW34zbnJzN6wl1lpe9yOY07cj4Fk4C5gsYgUOVuxiBS5nM0Ao5I8a4rYFBn/9PyXGWQXlPLMpAGEBQe6Hcf4gXorPFXNV9XrVbWDqsap6g3AjU2QzTTQ5r3F7DlYZlNjjM+5fXQiA+Lb8sjH6zlwuNztOOYEqGqAqkY4W2SNLUJVI93OZyAhOpxu7cOZn25Fu7/5bucB/vXtNm44oxun94h2O47xEyd7WfbuRk1hTsnRVo92E6rxNUGBATx7xUAKSyp44tONbscxplkREcb2juXbrfmUVditZv7iSGUVD3yYRofIMO4f38ftOMaPnGzRbv16fcjczbn0j48kLjLM7SjG/EBK50h+PrYXM77bxdzN+9yOY0yzMqZPHGUV1SzbVuB2FNNAf5uTSfq+Yp6+fAARYcFuxzF+5GSLdrtV3UfkFR9h9c4DnJ/S0e0oxhzTnWN7kdyhDQ9OX09RWYXbcYxpNkb0iCY0KIB5zieuxretzS7klflbuWpoF1u93JywYxbtR282qmMrBjo3YUZzHHM27UMVzkvp4HYUY44pJCiAP145iNziMp6ZZc2njGksYcGBnNkzmvkZNq/d15VVVHHPB2uJiwjldxenuB3H+KFjFu113HxU8yakoKYMaY7ty4376Nq+FX06RrgdxZjjGtw1ittG9+C95dlWYBjTiMb0jmNb/mG25R92O4o5jr9+vYUtuYf4wxUDadvKpsWYE2f9Af3Y4SOVLMrM57y+HRGx2wyM77v7vGR6xbXh/mnrOFhi02SaMxEZLyLpIpIpIg/U8Xo3EZknIt+JyDoRmeDsDxaRt0QkTUQ2icjkpk/vX442Ifgm3abI+KrVOw/w6oKtXDu8K2cnx7odx/gpK9r92IKMPMorqzm/n02NMf4hLDiQP189iLxDR3jsE1t0qbkSkUDgJeBCIAW4VkRqzwf4HTBVVYcA1wAvO/uvAkJVdQAwFPipiCQ0RW5/1S06nB6xrZlnrR990tFpMZ3atuLBCX3djmP8mBXtfuyrjfuICg8mtXs7t6MY02ADu0Rxp9NNZvZ6W3SpmRoOZKpqlqqWA1OAibWOUeBov/e2wO4a+1uLSBDQCigHbDGneoztHcfSrP2UlFe6HcXU8twX6WTlHebZKwZatxhzSqxo91OVVdXM2ZzLOX3iCAq0f0bjX+48pxf94yN5aMZ68g8dcTuOaXzxQHaN5znOvpoeA24QkRxgFvBLZ/804DCwB9gJPKeq1s+wHmN7x1FeWc2SrfvdjmJqWLG9gDecRZRGJcW4Hcf4Oav2/NTy7QUcLK2wVo/GLwUHBvDnqwdTXFbJg9PTULUuss1MXTfZ1P5HvhZ4U1W7ABOAd0QkAM9V+io8XcoSgd+KSI8ffAGRO0RkpYiszMuzaSHDEtsRHhLIPJvX7jNKyiu594O1xEe1YvKFNi3GnDor2v3UVxv3ERoUwFnJ9pu78U/JHSL47fnJfLlxHzO+2+V2HNO4coCuNZ534X/TX466FZgKoKpLgDAgBrgOmK2qFaqaC3wLpNb+Aqr6qqqmqmpqbKzd2BcaFMjIXjHM25xnvwT7iGdmbWb7/hL+eOVAWoda0z1z6qxo90Oqypcb9jE6KYbwEHsjMP7rttE9SO3ejkdnbmDPwVK345jGswJIEpFEEQnBc6PpzFrH7ATGAYhIXzxFe56z/xzxaA2cAWxusuR+bGzvOHYVlrI175DbUVq8eZtzeWfpDm4blciZPe3immkcVrT7oQ27i9hVWGoLKhm/FxggPHfVICqrlPumrbMrhM2EqlYCdwJfAJvwdInZICJPiMilzmG/BW4XkbXAe8DN6vkP4CWgDbAeT/H/L1Vd1+TfhB8a09vzicNcWx3VVfmHjnDvtLX06RjBPRf0djuOaUbsMq0f+ixtD4EBYvPZTbOQENOaBy/qy8Mfreetxdu5eWSi25FMI1DVWXhuMK2575EajzcCI+s47xCeto/mBHWO8iy09/XGXO44q6fbcVokVeWBD9MoKq3k37edTlhwoNuRTDNiV9r9jKoyK20PZ/aMpl3rELfjGNMobji9G2N7x/L055vZvNe6+xlzss5P6cDKHQUUHC53O0qL9P6KbL7etI/7xvemT8fI+k8w5gRY0e5nNuwuYsf+Ei4a0MntKMY0GhHhT1cNIjIsmLveW0NZRZXbkYzxS+eldKRaYc6mfW5HaXG25R/m8U82cmbPaH5inxgaL7Ci3c/8d2pMP5saY5qXmDahPHfVQNL3FfOHz+2+Q2NORv/4SDq1DeOrjVa0N6WKqmp+/f4aggOF568eREBAXV1PjTk1VrT7kZpTY9rb1BjTDI3pHcctIxN4c/F25m62osOYEyUinNu3Awu35NsnVk3o73MzWZtdyNOTBtCpbSu345hmyop2P3J0aswEmxpjmrH7x/ehT8cI7v1gHbnFZW7HMcbvnJfSgdKKKhZtyXc7SouwfFsBf5u7hcuHxHPxwM5uxzHNmBXtfmSWMzXmApsaY7AxjgAAAB5USURBVJqxsOBAXrx2CIeOVHLvB+uorrY2kMb8//buOz6qMu3/+OdKTygBQmihJEjvvetiBVGxr9hRdm3oFvVhcYu6+uyjq/52XV+6qzyKujZEcQUrNlZ9QOlJJIEgYoAQOoQWICS5f3/MwQ0xoWZyZibf9+s1r5y555x7rguPZ645c5/7HI8h7VNoEB+jITK1YMfeEn45bSltmyTx4EU9/A5HIpyK9jChoTFSl3Rq3oDfndeVz1du4fl5+X6HIxJW4mKi+EnnVD5dsYkyfekNGuccd7+RxbY9JTx5VT/q666nEmQq2sPEsvW7yNfQGKlDrh3SjrO6NuPhD5aTXVDkdzgiYeXsbs3ZuqeEzHU7/A4lYk2dm8+nKzZzz5gu9EhL9jscqQNUtIeJfy1dT1x0FGN6qGiXusHMePSy3qTWj2fiq0vYue+g3yGJhI2RnZsRE2V8pCEyQZFdUMTDHyzn7G7NGT8s3e9wpI5Q0R4GSsvKmZVVyBldmpGcFOt3OCK1pnG9OJ68uh8bivYz6c0sAne5F5GjSU6MZUj7FD7O2aT/b2rY7v0HueO1paTWj+fRy3phpukdpXaoaA8Dc7/bxtY9B7iob5rfoYjUun5tGzP53C7MztnE1Ln5focjEjZG9WjB6q17ydu02+9QIoZzjt/+axkFO/bxxJV9aZSka8yk9gS1aDez0WaWZ2arzGxyFa/Hm9nr3uvzzSy9wmv3eO15ZjbqaH2aWYbXx7den3Fe+51mlmtm2Wb2qZm1C2bOwfD20vU0TIjh9C6pfoci4osJIzI4u1tzHnp/OUvXaoyuyLEY3b0FUQbvZW/wO5SI8dLXa3gnq5Bfn9WRAelN/A5H6pigFe1mFg08BZwLdAOuNLNulVabAOxwznUA/gr82du2GzAO6A6MBv5uZtFH6fPPwF+dcx2BHV7fAEuBAc65XsCbwCPByDdY9h4o5cNlGzmvVyviY6L9DkfEF2bGY5f1pkVyAre/upSi4hK/QxIJeakN4hmckcJ732zQEJkasHjNDh58N5czujTjtpEd/A5H6qBgnmkfBKxyzq12zpUA04ALK61zIfCit/wmcKYFBoddCExzzh1wzn0PrPL6q7JPb5szvD7w+rwIwDk3xzlX7LV/DbQOQq5B83HuJvYdLONiDY2ROi45KZanrurH5t37uWt6luZvFzkG5/Vqyeote1mxUUNkTsbWPQeY+MoSWiYn8tef9iEqSuPYpfYFs2hPA9ZVeF7gtVW5jnOuFNgJpBxh2+raU4Air4/q3gsCZ98/qCpYM7vJzBaZ2aItW7YcNbnaMmNJAWmNEhnQrrHfoYj4rnebRvzh/G58umIzT3z2rd/hiIS80T0CQ2Te/0ZDZE5UaVk5d7y6lB3FJfzjmn6aEEJ8E8yivaqvoZVPjVW3Tk21/+eNzK4BBgCPVrEuzrkpzrkBzrkBqamhMXa8YEcx/7dqK5f2S9O3ehHPtUPacWm/1jz+ybd8lLPR73BEQlrT+vEMaZ/Ce9kaInOiHv0oj69Wb+N/Lu5J91aaj138E8yivQBoU+F5a6CwunXMLAZIBrYfYdvq2rcCjbw+fvReZnYW8DtgrHPuwEllVYveWFQAwE8HtjnKmiJ1h5nxp4t70Kt1MndOz2LV5j1+hyQS0s7r1ZLVW/eyfIOGyByvD5dt5JnPV3PNkLZc2j+sRtdKBApm0b4Q6OjN6hJH4MLSWZXWmQVc7y1fBnzmAqcCZgHjvNllMoCOwILq+vS2meP1gdfnTAAz6ws8Q6Bg3xykXGtcWbnjjUXrOLVjKq0bJ/kdjkhISYiN5ulr+pMQG8VN/1zErv268ZJIdUZ11xCZE7Fi4y7ump5JH29Ynojfgla0e+PLbwdmA8uB6c65HDN7wMzGeqs9B6SY2SrgTmCyt20OMB3IBT4EJjrnyqrr0+vrN8CdXl8pXt8QGA5TH3jDzDLNrPIXh5D0xbdbKNy5n3E6yy5SpVaNEnnqqn6s3V7Mr6dl6sJUkWo0rR/P0FNSeDe7UENkjtG2PQeY8MIi6ifE8My1/TV7m4SEmKOvcuKcc+8D71dqu7fC8n7g8mq2/RPwp2Pp02tfTWB2mcrtZx134CFg2oK1pNSL46yuzf0ORSRkDW6fwr0XdOPemTn85eOV3D2qs98hiYSkC3unMWlGNpnriujbVhMbHElJaTm3vLyYrXsO8MYtQ2neMMHvkEQA3RE1JG3evZ9Pl2/m0v6tiYvRfyKRI7l2SDuuGNCGJ+esYsbiAr/DEQlJ5/ZsQXxMFG8tWe93KCHNOcfv3/6Ghfk7ePTy3vRq3cjvkER+oIowBL02fx2l5U5DY0SOgZnx4EU9GNo+hclvZTN/9Ta/QxIJOQ0SYjmnewveyS6kpLTc73BC1tS5+UxfVMAdZ3RgbO9WfocjchgV7SGmpLScl+evYWTnVNqn1vc7HJGwEBcTxdPX9KdtkyRufnkx32/d63dIIiHnkr5pFBUfZE5e2MzJUKs+Xb6JP72Xy6juzfn1WZ38DkfkR1S0h5j3vilky+4D3DA8w+9QRMJKclIsU8cPJMqMG19YyI69JX6HJBJSTu3YlKb143hriYaRVZa5rojbX11K91bJ/EV3PJUQpaI9hDjneH5uPu1T63Fqh6Z+hyMSdtql1GPKtf1Zv2MfN7+8mP0Hy/wOSSRkxERHMbZ3Gp+t2ExRsb7UHpK/dS8TXlhI0wZxTB0/kHrxQZ2jQ+SEqWgPIUvWFpFdsJMbhqXrW77ICRqQ3oTHftqbBd9v51fTMinTVJAiP7ikXxoHyxzvZFW+12HdtG3PAcY/v4By53jxhkGkNoj3OySRaqloDyHPfP4dDRNiuKSf7romcjLG9m7FH87vxoc5G/n9299obmoRT/dWDenSogHTFq6r8/9f7CspY8KLi9iwcz/PXj9Q15FJyFPRHiLyNu7mo9xNjB+eoZ/mRGrAhBEZTDz9FF5bsI7HPsrzOxyRkGBmXD24LTmFu8gu2Ol3OL45UFrGLS8vJqugiCeu7Ev/dpq7XkKfivYQ8fd/ryIpLpobhqX7HYpIxLj7nM5cOagtT835juf+73u/wxEJCRf1TSMpLppX5q/xOxRflJaV88vXMvl85RYevqQno7q38DskkWOioj0E5G/dyztZhVw7pB2N68X5HY5IxDAz/vuiHpzbowUPvpvL9IXr/A5JxHcNEmIZ27sV72RtYNf+g36HU6vKyx2T3szmw5yN3Ht+N64Y2NbvkESOmYr2EPDknFXEREcx4VRN8yhS06KjjMfH9eG0Tqn85q1s3tRdU2uFmY02szwzW2Vmk6t4va2ZzTGzpWaWbWZjKrzWy8y+MrMcM/vGzHQf+Rp29eB27DtYxttL684dUp1z/GHmMt5aup67z+nEjSP0mSvhRUW7z1Zs3MWMJQWMH5ZOswb6XBIJhviYaKZc258RHZryX29maZ7qIDOzaOAp4FygG3ClmXWrtNrvgenOub7AOODv3rYxwMvALc657sBIoG6dDq4FPVsn0zMtmVe+XlsnLkh1zvHHd3J5Zf5abvnJKUw8vYPfIYkcNxXtPnvkwzzqx8dw28hT/A5FJKIlxEYz5doBDDslhbveyKpTZxh9MAhY5Zxb7ZwrAaYBF1ZaxwENveVk4NAchOcA2c65LADn3DbnnCbcD4Jrh7Qjb9Nu5n23ze9Qgqq83PH7t5fxwrx8JozI4DejO2OmaZUl/Kho99HXq7fx2YrN3DayA42SNJZdJNgS46J59rqBDMlI4c7pmRoqEzxpQMULCAq8toruB64xswLgfeAOr70T4MxstpktMbNJwQ62rhrbpxVN68fzv1+u9juUoCkrd0x+K5tX5q/l1pGn8Pvzuqpgl7Clot0nB8vKuX9WDq2SE7hheLrf4YjUGYlx0Tw3fgDDTmnK3W9kMVWzygRDVVVR5TEYVwIvOOdaA2OAl8wsCogBRgBXe38vNrMzf/QGZjeZ2SIzW7Rly5aajb6OSIiN5vqh7fh33hZWbtrtdzg1rrSsnP96I4vpiwr4xZkdmTRKZ9glvKlo98kLc/NZsXE3943tTkJstN/hiNQpSXExPDd+AKO7t+CBd3P5y0d5dWJcby0qANpUeN6a/wx/OWQCMB3AOfcVkAA09bb93Dm31TlXTOAsfL/Kb+Ccm+KcG+CcG5CamhqEFOqGa4a0IyE2imcj7Gz7vpLAPOyHLjq98+xOKtgl7Klo98H6on389ZOVnNmlGed0a+53OCJ1UnxMNE9e1ZcrBrThic9Wcd+sHMrKVbjXkIVARzPLMLM4Aheazqq0zlrgTAAz60qgaN8CzAZ6mVmSd1HqT4DcWou8jmlcL47L+7fh7aWFbN613+9wasT2vSVc9ezXfLpiMw9e2J3bz+jod0giNUJFey0rK3fcNT0TA+4f213f/EV8FBMdxcOX9uTm09rzz6/WcMvLiykuKfU7rLDnnCsFbidQgC8nMEtMjpk9YGZjvdXuAn5uZlnAa8B4F7AD+AuBwj8TWOKce6/2s6g7fnZqBmXO8fTn4X+2fd32Yi57eh45hbv4x9X9uHZout8hidSYGL8DqGv+98vVfL16O49c1os2TZL8DkekzjMz7hnTlZbJCTzwbi6XP/0Vz14/gJbJiX6HFtacc+8TGNpSse3eCsu5wPBqtn2ZwLSPUgvapdTjkr5pvDx/DTf/pD3NG4bn9MOL12zn5peWUFJaxssTBjMoo4nfIYnUKJ1pr0Vfr97GY7PzGNOzBZf3b+13OCJSwfjhGTx3/UDyt+7loqfm8k3BTr9DEqk1d5zRkfJyx9/nrPI7lBMyfeE6rpwyn3rx0cy4dZgKdolIKtprydptxdz68mLapSTx0CW9NCxGJASd3qUZM24bRkxUFJc9PY/pi9YdfSORCNA2JYnLB7TmtQXrWF+0z+9wjlmpNxPbpBnZDMpowsyJw+nYvIHfYYkEhYr2WrC+aB/XPDcfBzx3/UCSE2P9DklEqtGlRUNm3j6c/u0aM+nNbCa9mcX+g7q3j0S+28/oCAZ//mCF36Eck8KifYyb8jUvzMvnxuEZvHDDQN3zRCKaivYgW7V5D+OmfMWO4hJeuGEQ6U3r+R2SiBxF0/rxvDRhMHec0YHpiwq4+O/z+G7LHr/DEgmqtEaJ3Hxae2ZlFbIwf7vf4RzRx7mbOPdvX7J8wy7+Nq4P917QjZholTQS2bSHB4lzjneyCrn4qbkUHwhcFNOnTSO/wxKRYxQdZdx1Tmeev2EgG3fuY8zfvuT5ud9TrmkhJYLdOvIUWiYncH+IToFaXFLKfTOX8fN/LqJNk0Te/cWpXNin8s12RSKTivYaVlpWzhcrt3Dtcwu447WltE+tx6w7RtBbBbtIWDq9czNm/+o0hp2Swh/fyeXqZ+dTsKPY77BEgiIpLobJ53Yhp3AXL87L9zucw8z7biujH/+SF79aww3D05lx6zAy9Ou11CGa8rGG3DU9i5WbdvP91r3sOVBK0/rx3HdBN64bmk50lC46FQlnzRomMHX8QF5fuI4H383lnL9+wS/P7MiNIzKI1U/yEmHG9m7F20vX88jsFYzsnEr71Pq+xrOz+CCPzF7BK/PXkp6SxOs3DWFw+xRfYxLxg4r2GlJaXk5K/Th6t0lm2ClNObNrM+Jjov0OS0RqiJkxblBbhndoyh/fyeGhD1bw5uICHriwB0NPUQEhkcPMePjSXpz9l8+5+40sXr95qC9fTsvKHa8tWMv/+yiPnfsO8rMRGdx1TmcS4/TZKnWTORd6Y9b8NmDAALdo0SK/wxCREPZJ7ibufyeHgh37OLNLM+4e1ZmuLRv6HRZmttg5N8DvOGqTjtnBMTNzPb+clsmNwzO494Jutfa+zjk+X7mFP3+Yx/INuxiU0YT7LuhG91bJtRaDSG061uO2zrSLiJyAs7o1Z3iHpkyd+z3PfP4dY574kgt6teL2MzrQSfNESwS4sE8aS9cWMXXu93Rv1ZBLg3xTwEPF+uOffEvmuiLSGiXy1FX9GNOzhe5tIoKKdhGRE5YYF83E0ztwzeB2PPPFdzw/N59ZWYWc1imVn43I4NSOTVVsSFj73Xldydu4m0kzsqmfEMOo7i1q/D32HyxjVmYhL8zLJ3fDLtIaJfI/F/fksv6tiYvRNSMihwT1/wYzG21meWa2yswmV/F6vJm97r0+38zSK7x2j9eeZ2ajjtanmWV4fXzr9Rl3tPcQEakJyUmxTBrdhXmTz+DuczqxfMMurpu6gJGP/Zu/ffIt67ZrthkJT7HRUUy5rj8905K5/dUlzMxcXyP9OudYunYHf3wnhyEPfcqkGdmUlTsevqQnc+4eyVWD26pgF6kkaGPazSwaWAmcDRQAC4ErnXO5Fda5DejlnLvFzMYBFzvnrjCzbsBrwCCgFfAJ0MnbrMo+zWw68JZzbpqZPQ1kOef+Ud17HCl2jY8UkZNxoLSM97I38MaiAr5avQ2A3m0acXrnVE7v3IyeaclEBWlWKY1pl2DYue8gP39xEQvytzNhRAZ3n8AFoftKyliQv50vV27ho9xNrN1eTFxMFGd1bcZ1Q9MZnNFEv0xJnXSsx+1gFu1Dgfudc6O85/cAOOceqrDObG+dr8wsBtgIpAKTK657aD1vsx/1CTwMbAFaOOdKK753de/hjpC4PgBEpKYU7ChmZmYhnyzfROa6IpyDBvEx9GqTTJ82jejUvAHtUurRrkkSjZJiT7poUdEuwVJSWs6D7+by0tdrSGuUyE2ntee8Xi1pWj/+sPVKy8rZvPsA64v2sXrLHnIKd7Fs/U6Wrd9FSVk5cTFRDGmfwgW9WjKqRwsaJsT6lJFIaAiFC1HTgHUVnhcAg6tbxyu2dwIpXvvXlbY9dMuzqvpMAYqcc6VVrF/de2w94cxERI5R68ZJTDy9AxNP78D2vSV8sXILi9ZsJ3NdEc98vprSCnedjIkyGiTE0CAhltho47djunJm1+Y+Ri/yH3ExUTx4UQ/O69WShz5YwX2zcrhvVg7NGsRTLz6GfSVlFJeUsrek7LC7qdaPj6Fbq4aMH57O8A5NGZTeRNM2ipyAYBbtVZ0uqnx2u7p1qmuvaoDbkdY/1jgws5uAmwDatm1bxSYiIienSb04LuqbxkV9A+cU9h8sY+32YtZsK2bNtr3sKC5h175Sdu8/yMEyR8NEnYGU0DOkfQozJw5n2fqdfPntVr7fuod9B8tJjI0iMTaaBgmxtGqUSFrjRNo2SaJdk6SgDQcTqUuCWbQXAG0qPG8NFFazToE3dCUZ2H6Ubatq3wo0MrMY72x7xfWre4/DOOemAFMg8FPrcWUqInICEmKj6dS8gaaIlLDUIy2ZHmmaO12ktgTz0uyFQEdvVpc4YBwwq9I6s4DrveXLgM+8seazgHHezC8ZQEdgQXV9etvM8frA63PmUd5DRERERCQsBO1Muzd+/HZgNhANTHXO5ZjZA8Ai59ws4DngJTNbReDs9zhv2xxvNphcoBSY6JwrA6iqT+8tfwNMM7P/BpZ6fVPde4iIiIiIhIugzR4TzjQTgYiEK80eIyISXo71uK07F4iIiIiIhDgV7SIiIiIiIU5Fu4iIiIhIiFPRLiIiIiIS4lS0i4iIiIiEOM0eUwUz2wKsOYFNmxK40VOkiuT8Ijk3iOz8Ijk3OP782jnnUoMVTCjSMbtakZxfJOcGyi+cnUhux3TcVtFeg8xsUSRPtRbJ+UVybhDZ+UVybhD5+fkp0v9tIzm/SM4NlF84C2ZuGh4jIiIiIhLiVLSLiIiIiIQ4Fe01a4rfAQRZJOcXyblBZOcXyblB5Ofnp0j/t43k/CI5N1B+4SxouWlMu4iIiIhIiNOZdhERERGREKeivQaY2WgzyzOzVWY22e94jsTMpprZZjNbVqGtiZl9bGbfen8be+1mZk94eWWbWb8K21zvrf+tmV1fob2/mX3jbfOEmVkt5tbGzOaY2XIzyzGzX0ZYfglmtsDMsrz8/ui1Z5jZfC/W180szmuP956v8l5Pr9DXPV57npmNqtDu675sZtFmttTM3o3A3PK9fSfTzBZ5bRGxb4Yjv/eHY2U6Zodzfjpmh3duoXfMds7pcRIPIBr4DmgPxAFZQDe/4zpCvKcB/YBlFdoeASZ7y5OBP3vLY4APAAOGAPO99ibAau9vY2+5sffaAmCot80HwLm1mFtLoJ+33ABYCXSLoPwMqO8txwLzvbinA+O89qeBW73l24CnveVxwOvecjdvP40HMrz9NzoU9mXgTuBV4F3veSTllg80rdQWEftmuD1CYX84jlh1zA7f/HTMDu/c8gmxY3atJR+pD+8ffHaF5/cA9/gd11FiTufwD4A8oKW33BLI85afAa6svB5wJfBMhfZnvLaWwIoK7Yet50OeM4GzIzE/IAlYAgwmcBOHmMr7IzAbGOotx3jrWeV99NB6fu/LQGvgU+AM4F0v1ojIzXvPfH78ARBx+2Y4PEJhfzjOeNPRMTus80PH7LDKzXvPfELsmK3hMScvDVhX4XmB1xZOmjvnNgB4f5t57dXldqT2giraa53301tfAmc2IiY/76fITGAz8DGBMxFFzrnSKmL6IQ/v9Z1ACsefd215HJgElHvPU4ic3AAc8JGZLTazm7y2iNk3w0wo7A8nI+L2Gx2zgfA7rumYXcv7ZswJJCGHq2oMkqv1KIKjutyOt71WmVl9YAbwK+fcriMMEwu7/JxzZUAfM2sE/AvoeoSYjjePqr7E10p+ZnY+sNk5t9jMRh5qPkI8YZNbBcOdc4Vm1gz42MxWHGHdsNs3w0yk/nuF5X6jY3b4Hdd0zP6RWtk3dab95BUAbSo8bw0U+hTLidpkZi0BvL+bvfbqcjtSe+sq2muNmcUSOPi/4px7y2uOmPwOcc4VAf8mMHaukZkd+gJeMaYf8vBeTwa2c/x514bhwFgzywemEfi59XEiIzcAnHOF3t/NBD68BxGB+2aY8H1/OEkRs9/omB22xzUds/3YN2tzfFAkPgj8WrGawAUUhy6W6O53XEeJOZ3Dx0c+yuEXVjziLZ/H4RdWLPDamwDfE7ioorG33MR7baG37qELK8bUYl4G/BN4vFJ7pOSXCjTylhOBL4HzgTc4/MKf27zliRx+4c90b7k7h1/4s5rART8hsS8DI/nPRU0RkRtQD2hQYXkeMDpS9s1we/i9P5xAvOnomB2O+emYHaa5EaLH7Fr9DxupDwJXDa8kMFbtd37Hc5RYXwM2AAcJfNObQGBc2afAt97fQzuUAU95eX0DDKjQz43AKu9xQ4X2AcAyb5snIXADr1rKbQSBn5eygUzvMSaC8usFLPXyWwbc67W3J3AV+irvgBnvtSd4z1d5r7ev0NfvvBzyqHDFeijsyxz+ARARuXl5ZHmPnEPvHyn7Zjg+QmFfP8Y4dcwO3/x0zA7T3AjRY7buiCoiIiIiEuI0pl1EREREJMSpaBcRERERCXEq2kVEREREQpyKdhERERGREKeiXUREREQkxKloF6mGme3x/qab2VU13PdvKz2fV5P9i4jUNTpmS6RT0S5ydOnAcX0AmFn0UVY57APAOTfsOGMSEZGqpaNjtkQgFe0iR/cwcKqZZZrZr80s2sweNbOFZpZtZjcDmNlIM5tjZq8SuLkCZva2mS02sxwzu8lrexhI9Pp7xWs7dIbIvL6Xmdk3ZnZFhb7/bWZvmtkKM3vFzOxQf2aW68XyWK3/64iIhBYdsyUixfgdgEgYmAzc7Zw7H8A7kO90zg00s3hgrpl95K07COjhnPvee36jc267mSUCC81shnNuspnd7pzrU8V7XQL0AXoDTb1tvvBe60vgls+FwFxguJnlAhcDXZxzzswa1Xj2IiLhRcdsiUg60y5y/M4BrjOzTGA+gdsad/ReW1Dh4A/wCzPLAr4G2lRYrzojgNecc2XOuU3A58DACn0XOOfKCdzuOx3YBewHnjWzS4Dik85ORCSy6JgtEUFFu8jxM+AO51wf75HhnDt01mbvDyuZjQTOAoY653oDS4GEY+i7OgcqLJcBMc65UgJnimYAFwEfHlcmIiKRT8dsiQgq2kWObjfQoMLz2cCtZhYLYGadzKxeFdslAzucc8Vm1gUYUuG1g4e2r+QL4ApvDGYqcBqwoLrAzKw+kOycex/4FYGfaUVE6jIdsyUiaUy7yNFlA6XeT6YvAH8j8DPnEu/Coi0EzphU9iFwi5llA3kEfm49ZAqQbWZLnHNXV2j/FzAUyAIcMMk5t9H7AKlKA2CmmSUQOOPz6xNLUUQkYuiYLRHJnHN+xyAiIiIiIkeg4TEiIiIiIiFORbuIiIiISIhT0S4iIiIiEuJUtIuIiIiIhDgV7SIiIiIiIU5Fu4iIiIhIiFPRLiIiIiIS4lS0i4iIiIiEuP8PgGZu46KlD1EAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2d6b9fc400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learner.recorder.plot_lr(show_moms=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:21:06.711105Z",
     "start_time": "2019-08-21T02:21:05.287164Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2d6c403048>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learner.recorder.plot_losses()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:21:08.400771Z",
     "start_time": "2019-08-21T02:21:06.713776Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZQAAAfVCAYAAAC1EThTAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3Xd4lFXa+PHvnd57IKTQQjf0WEAFC7qIKNZd113Xui669n3f3fV1i/vTLa5u17X3ta9iAUSwgUqR0AMBpCckgVBCAiGk3b8/5gkGmJCQmWQmk/tzXXNl5syZ89wPJLlzznmec0RVMcYYYzwV5OsAjDHGBAZLKMYYY7zCEooxxhivsIRijDHGKyyhGGOM8QpLKMYYY7zCEooxxhivsIRijDHGKyyhGGOM8YoQXwfQkVJSUrR3796+DsMYYzqVJUuW7FLV1JbqdamE0rt3b/Ly8nwdhjHGdCoisrU19WzIyxhjjFdYQjHGGOMVfpdQRORhEVkrIitFZJqIJDRTL0FE/uvULRCRMR0dqzHGmG/5XUIB5gA5qjoMWA/c20y9fwCzVHUQMBwo6KD4jDHGuOF3CUVVZ6tqnfNyIZB5dB0RiQPGAc86n6lR1fKOi9IYY8zR/C6hHOUG4EM35X2BMuB5EVkmIs+ISLS7BkTkZhHJE5G8srKy9oy1S5mxsoTi8oO+DsMY40d8klBE5GMRyXfzmNKkzn1AHfCKmyZCgFHA46o6EjgA/NLdsVT1KVXNVdXc1NQWL6M2rbDnQA0/fXUpL8zf4utQjDF+xCf3oajqhOO9LyLXApOBc9X9HsVFQJGqLnJe/5dmEorxvuWFewHYsuuAjyMxxvgTvxvyEpGJwC+Ai1W1yl0dVS0FCkVkoFN0LrCmg0Ls8pZtc01Xbdvj9r/HGNNF+V1CAR4FYoE5IrJcRJ4AEJF0EZnZpN7twCsishIYAfyh40PtmhoTSuGeKtx3II0xXZHfLb2iqv2aKS8GJjV5vRzI7ai4jEt9g7K8sJzwkCAO1NSz+0ANKTHhvg7LGOMH/LGHYvzYhp372X+ojgmDuwM27GWM+ZYlFHNClm1zTchPGZEOuIa9jDEGLKGYE7RsWzkJUaGc2d91CfbW3e4TiqpSVVPn9j1jTGCyhGJOyLLCvYzMSiAyLJjuceHNDnm9t7yYkf9vDgs27u7gCI0xvmIJxbRaRXUt3+zcz8ieiQD0SopmWzM9lK+37OFQXQM3v5zHutLKjgzTGOMjllBMq60s3IcqjOzpWgA6Kymq2R7K2pIKBnSPISosmOue/5qSfbZMizGBzhKKabWl2/YiAsOzXAmlZ1IUpRXVVNfWH1GvoUFZW1rJ2OwUnr/uFCqr67j++cVUVNf6ImxjTAexhGJabdm2vfRLjSEuIhSAXslRABTtPbKXsm1PFVU19QzuEcuQ9DievGY0G8v285OXlnCorv6Ydo0xgcESimkVVWVZYfnh4S5wDXnBsfeirC2tAGBQWhwAp/dL4eErhrNg025e/7qwgyI2xnQ0SyimVbbsrqK8qvbwhDy4hryAYybm15RUEiQwMC32cNklIzMY3COOd5dv75iAjTEdzhKKaZXGGxpHNUkoKTFhRIUFs/XoHkpJBX1SookIDT6ifMqIdJZtK2/2yjBjTOdmCcW0ytJte4kJD6Fft5jDZSJCz6SoY+6WLyitYFCPuGPauGi46+7691dYL8WYQGQJxbTKsm3lDM+KJzhIjijvedSlw5XVtRTuOcgQNwklIyGSU/ok8e7yYlul2JgAZAnFtKiqpo61pZWMzEo85r3GhNKYIBpvYhzUZP6kqSkj0tmwcz9rSiraL2BjjE9YQjEtWrJ1L/UNyoishGPe65kcRXVtA2WVhwAocBLFYDc9FIBJOT0ICRLeX17cfgEbY3zCEopp0XNfbiYpOozT+6Uc817Poy4dLiitJD4ylB7xEW7bSowOY/yAVN5fUUxDgw17GRNILKGY4yooqeCzdWVcP7Y3kWHBx7zfmFAaVx0uKKlgUFosInJM3UYXj0inZF81i7fsaZ+gjTE+YQnFHNcTczcSHRbMj8b0dvt+RmIkIq4eSkODsq60stnhrkbnDelOZGgw7/rJsNfW3Qf4aHWpr8MwptOzhGKatW13FR+sKObqU3sSHxXqtk54SDDp8ZFs21N1xJIrxxMVFsL5J3Vn5qoSauoa2iP0E/LvzzZyy3+WsLOy2tehGNOpWUIxzXrqi42EBAVx05l9j1svK8mVUFqakG/qkhEZ7DtYywcriindV83Oymr2HKhhZ0U1a0srmL9xF9NXFjN9ZTG19e2bdNaUVNCg8OEq66UY44kQXwdg/FNZ5SHezCvislEZdI9zP8HeqGdSFJ+tK6Og1LXkyoDux++hAJzRP4Wk6DB+9taKFutOGZHO3747gqCg5udl2qquvoF1O1yXOr+/ophrx/b2+jGM6SosoQSYlxdupbqmnh+PO36voiXPf7WZ2voGbm5FO72SoymrLGLZtr1ul1xxJzQ4iJduOIU1xRXUq1LXoDQ0KEFBQlJUGEnRrsfs1aX8Zc56YiNCeGBKznEn+9tiy+4D1NQ10L9bDEu27qVobxWZiVFePYYxXYUllFZ4f0UxS7fu5f6LT/J1KMdVsu8gD0xfQ7AIPzitJ1Fhbfvvraiu5eUFW7kgJ42+qTEt1m9cdXjBxt1MzElr9XFyMuLJyYg/bp2BabEcqKnnibkbiYsI5ecTB7W6/dZYU+Lqnfzs/IFM/c8SZqws4Sfjs716DGO6CptDaYUtuw7wwvwth+cI/NW/Pt1ATV0DB2vr+XTtzja388rCbVQequOW8f1aVb+Xk1DqGrRV8ycn6hcTB3L1qT359+cbeWLuRq+2vbakgpAg4ZxB3RielcAHK/3jyjNjOiNLKK1w7ZjexISH8O/PvfvLzJu27a7izcWF/ODUnqTGhjNjZUmb2qmurefZLzdzZv8UhmYev/fQqPFeFKDFK7zaQkR4YEoOFw1P508frm3zublTUFJBv24xhIUEcdGwHuRvr2BT2X6vtW9MV2IJpRXio0L54Wm9mLGymM27Dvg6HLf+/sl6goOEO87tz6ScND5du5MDh+pOuJ23lhSxa/8hbjmr9cM+CVGhxIa7htfao4cCEBwk/PW7w+nXLYanv9jktXYLSr69b2bysHREYLoXE5YxXYkllFa68Yw+hAYH8YQf9lI27Kzk3WXbuXZsb7rHRTB5eDqH6hr4uGDHCbVTV9/Ak3M3MiIrgTF9k1v9ORGhZ3IU8ZGhpLVwRZgnQoOD+G5uJssLy9nohV7E3gM1lFZUH+5VpcVHcHLvJN5fYashG9MWllBaKTU2nKtOzuKdZUUUlx/0dThH+Nucb4gMDWaqM5k8umci3eNOfNjrg5XFFO09yK1nZZ/w1VQTT0rjslEZXr8K62iXjMggSODtJUUet1Vw1FbF4NqzZcPO/YcvJTbGtJ7fJRQReVhE1orIShGZJiLHLHErIgNFZHmTR4WI3NXesd08PhtVeGqe94ZcTsTywnJyH/yYW19Zwqz8Uqpr68nfvo8Zq0q48Yw+JEWHARAUJEwa2oPP15dRWV3bqrYbGpTHP9/IgO4xTBjc/YRju/3c/vz2ova/Cq5bXATjB6TyztLt1Hu4uGSBc4VX02G6C3LSCA4SPlhhk/PGnCi/SyjAHCBHVYcB64F7j66gqutUdYSqjgBGA1XAtPYOLCMhkktHZvD64m3s2n+ovQ93jMc+20B1bT1fb97D1P8s4eQHP+aWV5YQHxnKTUfdLzJ5WDo1JzDs9cnanazfsZ9bzspulxsIveny0ZmUVlQzf+Muj9pZW1JBSkw4qbHhh8tSYsIZm53MBytKbNjLmBPkdwlFVWerauNs8kIgs4WPnAtsVNWt7RuZy9SzsjlU18BzX24+br2ivVWU7vPe2lCbdx3g44IdXH96bxbeey4v3XAK38lJo7K6jrsn9Ccu4si1tkZmJZAeH9GqYS9V5d+fbyAzMZKLhqV7Leb2MmFwd+IiQlo17DVnzQ4m/+sLlmw9dmXjgtIKt1elXTQ8nW17qli1fZ9X4jWmq/D3GxtvAN5ooc5VwGvNvSkiNwM3A/Ts2dPjgLJTY5g0tAcvLdhKQlQoo3omkpMRT0RoMPuqapm+qphpS7eTt3UvSdFhvDV1DNmtuDkQXL/Ym5uDeP6rzYQGBXHNmF6EBAcxbkAq4wakNttW47DXSwu2su9gLfGR7hd3BFi4aQ/LtpXzwJSTCAn2u78xjhERGsxFw9N5e2kRldW1xEYce277Dtbyuw9W885S1/71T8zdxNM/Sjr8fl19A+t37Oc6N0utnDuoGyIwd10ZwzKP3VTMGOOeT357iMjHIpLv5jGlSZ37gDrgleO0EwZcDLzVXB1VfUpVc1U1NzW1+V/AJ+Ke8wbQLS6cP8xcyxVPLGDo/R8x8e/zOPn3H3PftHz2Hazlrgn9CRL40bNft2oS/8ChOib/60vueWP5MUMt5VU1vJVXxMUj0ukW2/qrqCYPT6emvoE5a5of9qqurecvs9eREhPGlblZrW7b1y4fnUl1bQMzVx3bA5u7vozv/G0e7y0v5o5z+nHjGX34dO1OdlZ822PcvMu15Iq7HkpyTDg56fHM+6asXc/BmEDjkx6Kqk443vsici0wGThXjz+QfQGwVFVP7PpYD2WnxvDpz86irPIQy7btZem2clYX72NsdgqXjcrgpPQ4RIQJg7vz/acWcs2zi3hr6tjDk+bu/Oa91awurmB1cQXDMuO57vQ+h9979ettHKyt56Yz+zT7eXeGZ8aTkRDJjJXFXDH62JHDA4fq+PFLeeRt3ctfrhzeqjW4/MXIrAT6pkTz9pLtfO9kV89z74Ea/jCzgLeWFNGvWwxPXjOa4VkJbN51gGe/3MxbS4r46dmuu/8b97RveoVXU+MGpPDE3E1UVNceM5xojHHP78Y3RGQi8AvgYlWtaqH69znOcFd7S40N5/yT0vjlBYN4+cZT+c1FQ8jJiD88bJWTEc8z1+ZStPcg1z//NfubudHwnaVFvL20iDvO7c+Ewd34/cwCVhSWA1BT18CL87dwZv+UZn/5NUdEmDysB/O+2cXDH62loskVX/uqavnhs4tYtHkPf/3ucC53k3D8mYhw+ehMvt6yh627DzBtWRHn/nUu05ZtZ+r4bKbffgbDs1zDVX1SojmtbxJvLC48vO3w2tJKQoOl2eHIcf1TqW9Q5m/Y3WHnZExn53cJBXgUiAXmOJcEPwEgIukiMrOxkohEAecB7/gmzNY5tW8yj109ivziCm56cTEl+44c/tpUtp9fvZvPKb2TuOOcfjxy5XC6xUbw01eXsq+qlhmritlRcYgbzzix3kmjW87KZvKwHjz22UbG/fkznp63ie3lB7nq6YWs3l7BY1eP4rJRnSuZNHLd9wKXP76Au99YQa/kKKbfcQa/vGDQMb2tq07uybY9VSzc5EoQBSUVZKe6llxxZ2TPRKLDgm3Yy5gT4HcJRVX7qWpW42XBqjrVKS9W1UlN6lWparKq+v2lOBOGdOcvVw5n6bZyzn7kc/46Zz1VNXUcqqvn9teWERYSxN+vGkFIcBAJUWE8evVISvdV8z//XcEzX2ymf7cYxh9nAv54EqLC+MdVI5l++xkMy0zg9zMLOOOhT9m8az/PXJt7QqsD+5se8ZGcM7Abh2rreeCSHN6eOrbZXtzEnDTiI0N5bXEhAGtLKhlynGViwkKCGJOdwrz1ZXb5sDGt5O9XeQWMS0ZmMLpXIg/NWss/P/mG17/exknpcawuruDpH+WSnhB5uO7Inon88oJBPDijAIA/XTbU4zvQczLieemGU5i/cRevLNzG9af3Jrd3Ussf9HOPXj2KelViwo//rRwRGsylIzN4ddE2NpXtd5ZcOf4Q4vgBKXxcsIMtu6vokxLtzbCNCUh+10MJZFlJUTx69SjevmUs6QmRfLaujOvG9ua8IcfemX7jGX24cGgPMhIiuWRkhtdiGJudwmM/GBUQyQQgMiy4xWTS6HsnZ1FT38AfZq4FYFALKyM3XpY9b70NexnTGtZD8YHRvRJ555axrNq+r9kNpkSER68eSXVtQ6e6+sqfDe4Rx/CshMOrB7TUQ+mVHE2v5CjmrS+zrYGNaQXrofhIUJAwPCuB4OMscyIiRIZZMvGm75/sutcmNTaclJjwFmq7rvZasGk3NXUN7R2aMZ2eJRTTpUwenk5UWDCD0lq3Edi4AalU1dST52bpFmPMkWzIy3QpMeEhPHnNaJKjW+6dAJzWN4mQIGHe+l2MzU5p5+iM6dysh2K6nDP7pzIkvXU3icZGhDKqVyJf2P0oxrTIEooxLRg/IJXVxRWUVXb8lgXGdCaWUIxpwbj+rsuHP1pd6uNIjPFvllCMacFJ6XEMSovl1+/l88D0NVTX1vs6JGP8kiUUY1oQFCS8fctYfnhqL579cjOT/vEFS7ft9XVYxvgdSyjGtEJ0eAgPXJLDKzedyqG6Bq54fD5PzN3o67CM8SuWUIw5Aaf3S+Gju8cxYXB3HvloHUV7W9phwZiuwxKKMScoJjyE+y8+CRF4at4mX4djjN+whGJMG6QnRHL5qExeX1zIzsrqlj9gTBdgCcWYNpo6Ppu6+gae/WLzMe8V7a3i6qcX2krFpkuxhGJMG/VOieai4en8Z+FW9h6oOVy+/1AdN72Yx/yNu7nt1aVs3X3Ah1Ea03EsoRjjgVvP6seBmnqen78FgPoG5c7XlvHNzv380dkY7ScvL+Fgjd27YgKfJRRjPDAwLZbzh3Tnha82U1ldy58+LOCTtTu5/6IhfP+UnvzjqhGs21HJ/01bZVsJm4BnCcUYD912Tj8qquu48cU8nv5iM9eO6cU1Y3oDcNbAbtw9YQDTlm3npQVbfRuoMe3Mo+XrRWQacOyMpPM2EK6qt3pyDGP83bDMBM7sn8IX3+xi3IBUfj15yBHv33Z2P1YUlvPA9DWclB4XMNsvG3M08aQbLiIfqOpFx3l/mqpe2uYDeFlubq7m5eX5OgwTgNbvqOT5r7Zw76RBxEWEHvP+voO1XPzolxysqWf6HWfQLTbCB1Ea0zYiskRVc1uq5+mQV0vZyAaNTZcwoHssf7xsqNtkAhAfGcoTPxxNRXUtt726jNr6tm0p/NCstdz26lJPQjWm3dgcijEdZHCPOP5w6VC+3ryHP89ae8Kfr61v4NVF25i+ssQuRTZ+ydMtgNNE5OJm3hMgxsP2jQkol43KZNm2cp7+YjMjeyYyaWiPVn928eY97DtYC8C0Zdu5a8KA9grTmDbxtIfyv0B5M4+9wL0etm9MwPnV5MGMyErgf99awYad+1v9udlrdhAeEsToXom8s3S7XYZs/I6nPZS7cF3lJc28Hw4s8fAYxgSU8JBg/v2DUUz+15dc/OiXxEWEEiQgIkSHB/P0j3LplRx9xGdUlTlrdnBm/1QuyEnjZ2+tIG/rXk62K8aMH/E0oYSo6j3NvelcVnxCRORh4CKgBtgIXK+q5W7q3Q3chGvif5VTz1bpM51CekIkL15/Cq9+vY36hgZUoUHhg5XFPP3FJh68ZOgR9VcXV7C9/CB3TujPxJw0fv1ePu8sLbKEYvyKpwmlPa7ymgPcq6p1IvIQrmGzXzStICIZwB3AEFU9KCJvAlcBL7TheMb4xNDMeP6YeWTiCA6C/y4p4mfnDSQxOuxw+ew1OwgSOHdQN6LDQ5iYk8b0lSX89qKTiAgN7ujQjXHL767yUtXZqlrnvFwIZDZTNQSIFJEQIAoo7oj4jGlPN57Rl+raBl79etsR5bNXl5LbK4nkmHAALh+VSWV1HXPW7PBFmMa45WlCSRORi5t5TMHzq7xuAD48ulBVtwOPANuAEmCfqs5214CI3CwieSKSV1ZmS4kb/zYwLZYz+6fw4vwt1NS57lUp3FPF2tJKzj+p++F6p/VNpkd8BO8sLfJVqMYcwxtXee1z82i80uv/3H1IRD4WkXw3jylN6twH1AGvuPl8IjAF6AOkA9Ei8kN3x1LVp1Q1V1VzU1NTPTlXYzrEjWf0YWflIaavdHW6Zzu9kPOGfJtQgoOES0dmMO+bXbbBl/EbbU4oIvJzVZ2rqnOBbo3PndcTnedu1zlR1QmqmuPm8Z7T9rXAZOAH6v7ayAnAZlUtU9Va4B1gbFvPxRh/Mn5AKv27xfDMF5tRVWavLmVQWuwxV35dNiqD+gbl/eU22mv8gyc9lKuaPD/6fpOJbW1URCbimoS/WFWrmqm2DThNRKJERIBzgYK2HtMYfyIi3HhGH9aUVPBhfimLt+w5onfSqF+3WIZnxvP20u0+iNKYY3mSUKSZ5+5en4hHgVhgjogsF5EnAEQkXURmAqjqIuC/wFJclwwHAU95cExj/MolIzNIig7jF2+vpEHh/CFpbutdOjKDgpIK1u+o7OAIjTmWJwlFm3nu7nXrG1Xtp6pZqjrCeUx1yotVdVKTer9V1UHOUNk1qnqorcc0xt9EhAbzw9N6UVldR4/4CHIy4tzWmzS0ByIwc1VJB0dozLE8SSjDRaRCRCqBYc7zxtdDW/qwMeb4rjmtF+EhQUzMScM1snusbnERnNwriQ9XlXZwdMYcq803Nqqq3U1lTDtKjQ3no7vG0S0u/Lj1Jg1N4/4P1rBh5376dbP1WI3vePXGRhGJFpEfiMgMb7ZrTFfVOyWaqLDj/903Mce1YvGHNuxlfMzjhCIiYSJyibP8SQmuS3qf8DgyY0yrpMVHMLpXIjPzbdjL+JYn96GcJyLP4Vpt+ArgZWCPql6vqh94K0BjTMsuyEmjoKSCzbts4y3jO570UD4CsoEzVPWHThJp276mxhiPXOBs1PVhvg17Gd/xJKGMxrV448ciMkdEbgRsot4YH8hIiGREVoJd7WV8qs0JRVWXqeovVDUbuB8YCYSJyIcicrO3AjTGtM6koWms2r6PbbubW2DCmPbllau8VPUrVb0NyAD+DozxRrvGmNa7IMeGvYxvtfk+FBEZ1cxbZcC/2tquMaZtspKiGJYZz8z8Un4yPtvX4ZguyJMdG/OA1bgSCBy5fpcC53jQtjGmDS7I6cFDs9by7Jeb2XewlpLyg5RWVDMxJ40fnNrL1+GZAOdJQvkZcDlwEHgdmKaq+70SlTGmTS4c2oNHZq/jgelrEIFuseEIwqJNazg9O4XeKdEtN2JMG4n77UZOoAGRPsD3cW14tRX4g6ou90JsXpebm6t5eW63aDEmYBTuqSI4SEiNDSc0OIgdFdWc88jnnNY3mWevO9nX4ZlOSESWqGpuS/U8npRX1c3Ae8Bs4BRggKdtGmPaLispivSESEKDXT/e3eMiuOPc/nyydiefFNge9Kb9eHKnfF8R+T8RWQT8DlgBDFLVN70WnTHGK64/vQ/ZqdH8v+lrqK6t93U4JkB50kPZAHwXmAUsAHoCt4rIPSJyjzeCM8Z4R1hIEPdffBJbd1fxzBebfB2OCVCeJJT/B0zDtdxKDK5dFhsftoa2MX7mzP6pXJCTxqOfbWB7+UFfh2MCkCf7odzf3Hsicldb2zXGtJ/7LhzMZ+t28uD0NTz+w9G+DscEGK/uh9KEDXkZ44cyE6O47ex+fJhfyuzVtu6X8a72Siju9ys1xvjczeOyGZQWy6/ezWdfVa2vwzEBpL0Simc3txhj2k1YSBCPXDmc3QdqeHDGGl+HYwKIJ5cNV4pIhfO1sulrIN2LMRpjvCwnI56p4/vy1pIi5q4va/kDxrSCJ8vXx6pqnPM19qjXnizpYozpALef059+3WK49+2VVFY3P/TV0KA8NW8jj322oQOjM52RR7/4ReQ3LVTZqaq2v7wxfigiNJg/XzGMyx+fz0Oz1vLgJUOPqVNZXcs9b65gzhrXHfaDe8RyzqDuHR2q6SQ87UmMAb5H85PwLwKWUIzxU6N6JnLj6X145svN7Kqs4fun9uTMfikEBQkby/Zz80t5bNldxa8uHMx/lxRx7zurmH13EvGRob4O3fghTxNKnapWNPemiNjkvDF+7n++M5CQ4CDezCtk1upSMhMjmXhSGm8sLiQ0JIj/3HgqY7KTObVPMpf8+yt+P2MNf75iuK/DNn7I06u8WkoYllCM8XMRocH88oJBLLj3HB69eiQ9k6J45svN9EqJ4v3bTmdMdjIAQzPj+cm4vryZZxP5xj1PeyihIhLXzHsCBJ9ogyLyMHARUANsBK5X1XI39e4Efuwc52lV/fuJHssY863wkGAmD0tn8rB09hyoIS4ihJDgI//mvOPc/sxes4N7317JR3ePIzbChr7MtzztoSwE7jrO48M2tDkHyFHVYcB64N6jK4hIDq5kcgowHJgsIv3bcgLGmGMlRYcdk0zg24n80opq/vjhWh9EZvyZpz2UU4Gr8OKkvKrObvJyIXCFm2qDgYWqWgUgInOBS4E/n8ixjDEnblTPRG5wJvJvOL03/brF+jok4yc87aHUq2qFqu5z98DzOZQbcN/LyQfGiUiyiEQBk4AsD49ljGmlqWdlExIkvLG40NehGD/ik0l5EflYRPLdPKY0qXMfUAe8ckyjqgXAQ7iGx2bh2tyrrplj3SwieSKSV1ZmE4nGeENKTDjnDenO20u3U1PX4OtwjJ/wyaS8qk44XqMici0wGThXm9n0XlWfBZ516v8BKGqm3lPAU+DaU/54xzXGtN53T87iw/xSPi7YwaShPXwdjvEDniaUxkn55uZQZp1ogyIyEfgFML5xjqSZet1UdaeI9AQuw3WTpTGmg4zrn0qP+AjeWFxoCcUAHiYUVf2dtwJp4lEgHJgjIuCafJ8qIunAM6o6yan3togkA7XAT1V1bzvEYoxpRnCQcOXoTP7l7ACZkRDp65CMj7XX8vVtpqr9VDVLVUc4j6lOeXGTZIKqnqmqQ1R1uKp+4ruIjem6rsx1XQvz3zy3I86mi/G7hGKM6TyykqI4PTuFN/MKqW+wKcquzhKKMcYj3zs5i+3lB/lqwy6375dVHuL9FcXc+85Kfv1uPs1cZ2MCgO1bYozxyPkndSchKpQ38goZNyAVVWV5YTkzVpYw75sy1u/YD0BIkFDXoNxwRh/6pET7OGrTHiyhGGM8Eh4SzKUjM/jPwq08MH0Ns/KTZh65AAAgAElEQVRL2V5+kLDgIE7tm8SlIzM5vV8ysRGhnP3I58xdt5M+KX18HbZpB5ZQjDEeu+rknrwwfwsvLdjCmf1Tuee8AUwY0v2YfVP6pEQzd30Z151uCSUQWUIxxnhsYFosH955Jj3iIomPan4F4vEDUnl98Taqa+uJCD3hxciNn7NJeWOMVwxKiztuMgFXQqmubWDxlj0dFJXpSJZQjDEd5tS+SYSFBDF3na2rF4gsoRhjOkxUWAin9kmyHR8DlCUUY0yHGj8glW927md7+UFfh2K8zBKKMaZDjR+QCsA866UEHEsoxpgO1a9bDOnxETaPEoAsoRhjOpSIMH5gKl9t2EVtvW3OFUgsoRhjOtz4AalUHqpj2bbyI8qra+ttra9OzBKKMabDje2XQnCQMHf9TgBq6xt4et4mRj8wh+8+uYDCPcfurVddW8/vZ6zh7Ec+p3RfdUeHbFrBEooxpsPFRYQyumcic9eXMX/jLib94wt+P7OA4VkJrC2pZNI/vmDasqLDvZXlheVc+M8vePqLzWzdfYDHPtvg4zMw7tjSK8YYnxg/MJWHP1rH1U8vIispkmd+lMuEId0p3FPFPW8u5+43VvDp2jKyEiN5ct4musWG89INp/BhfimvL97GT8b3JTMxytenYZqwhGKM8YlJQ3vwxuJCLh2ZwS1nZR9e2ysrKYrXbx7DE3M38rc566lrUK4YncmvJw8hPjKUft1ieHtJEY9+uoE/XT7Mx2dhmpKuNAGWm5ureXl5vg7DGNNKa4orKK+qYWy/lCPKf/tePv9ZtI1PfzaeXsm2t0p7E5ElqprbUj2bQzHG+K0h6XHHJBOAn57dj5Ag4Z+f2FyKP7GEYozpdLrFRXDNab2YtqyIjWX7fR2OcVhCMcZ0SlOdeZd/fPyNr0MxDksoxphOKSUmnGvH9uaDlcWsKa7wdTgGSyjGmE7s5jP7khQVxi2vLGHPgRpfh9PlWUIxxnRaidFhPPWj0ZTsq2bqy0s4VFfv65C6NEsoxphObXSvJB6+Yhhfb9nDvW+vsrXAfMhubDTGdHpTRmSwdXcVf52znj4p0dx+bn9fh9QlWUIxxgSE28/px+ZdB/jLnPX0Sonm4uHpvg6py/G7IS8ReUBEVorIchGZLSJuvytE5FoR+cZ5XNvRcRpj/IuI8KfLh3Jy70T+560VLN2219chdTl+l1CAh1V1mKqOAKYDvzm6gogkAb8FTgVOAX4rIokdG6Yxxt+EhwTz5DW5pMVFcPNLeRTtPXYZfNN+/C6hqGrTC8qjAXczbN8B5qjqHlXdC8wBJnZEfMYY/5YUHcZz1+VyqK6Bm17MY/+hOl+H1GX4XUIBEJHfi0gh8APc9FCADKCwyesip8xdWzeLSJ6I5JWV2R7WxnQF/brF8tjVo/hm537ufG0Z9Q125VdH8ElCEZGPRSTfzWMKgKrep6pZwCvAbe6acFPm9jtGVZ9S1VxVzU1NTfXeSRhj/Nq4Aancf/FJfLJ2J//z1go+WFHM4i17KNxTRU2d7WXfHnxylZeqTmhl1VeBGbjmS5oqAs5q8joT+NzjwIwxAeWa03pRtKeKJ+dtYtqy7YfLo8KCeeZHuW5XMjZt53f7oYhIf1X9xnl+OzBeVa84qk4SsAQY5RQtBUar6p7jtW37oRjTNe07WMuOimpK9lWzY181j8/dSE1dA7PuOpPYiFBfh+f3Wrsfij/eh/InERkINABbgakAIpILTFXVm1R1j4g8ACx2PvP/WkomxpiuKz4ylPjIUAZ0jwWgX/cYrnh8Pn+YWcAfL7NdH73F7xKKql7eTHkecFOT188Bz3VUXMaYwDGqZyI/HteXJ+duYmJOD8YPsPlVb/DLq7yMMaa93T1hAP26xfCL/65k38FaX4cTECyhGGO6pIjQYP5y5XDK9h/iwelrfB1OQLCEYozpsoZnJTB1fF/eWlLEC19tprbeLif2hCUUY0yXdse5/TmlTxL3f7CGsx/5nFcWbbV9VdrIEooxpksLDwnmjZtP47nrckmJCee+afmM//PnPPvlZiqrbW7lRPjdfSjtye5DMcYcj6ry5YZd/OvTDXy9eQ8x4SFcmZvJdWN70ys52tfh+Uxr70OxhGKMMW6sKCzn+a82M2NVCXUNyrmDunPr2dmM6tn1Fja3hOKGJRRjzInaWVHNfxZu5eWFW9lbVcvp/ZK57ez+nNY3CRF3ywoGHksoblhCMca01YFDdby6aBtPztvErv2HGN0rkTF9k0mLj6BHfARp8RFkp8YQERrs61C9zhKKG5ZQjDGeqq6t5828Ql6Yv4Wtu6uOWBo/MjSYswamMjEnjbMHdSMuQNYJs4TihiUUY4w31dU3sGt/DSX7DlJcXs3CTbv5aHUpOysPERosTBjcnd9NOYlusRG+DtUjllDcsIRijGlvDQ3KssJyZuWX8PLCrcSEh/LPq0Z06qXyW5tQ7D4UY4zxoqAgYXSvRO67cAjv/fQM4iND+MGzi/j7x+sDfudIv1tt2BhjAsXAtFjev+0Mfv1uPn//+BsWbNzNWQO7kRQdSmJUGMkxYQxKiyM6PDB+FQfGWRhjjJ+KDg/hL98dzmnZyTw4fQ2LNh+5dVNYSBBn9kvh/JO6M2Fwd5Jjwqmrb2Bn5SFK9lVTXVtPbu9EwkP8/+oxm0MxxpgOVFVTx54DNew9UEvZ/mq+2rCbWfmlbC8/SJBAckw4u/cfounoWFxECBcOS+fyURmM7pXY4fe/2KS8G5ZQjDH+SFVZXVzB7NWllFZUkxYfefjeloYGZfrKEmbll3Kwtp6MhEgSo0PZX13H/kP17D9Ui6qrpxMeEkx4SBDhoUFEhga7HmHBRIQGc8c5/RmaGd+m+DrzFsDGGNOliAg5GfHkZLj/hX/u4O48eEkds/JLmbW6lPoGpW9KCNHhIcSEBxMkwqG6BmrqGzhU28Chunqqa+s5WFvP/kN1lFUeoqa+/VdQtoRijDGdQHR4CJePzuTy0Zm+DqVZdtmwMcYYr7CEYowxxissoRhjjPEKSyjGGGO8whKKMcYYr7CEYowxxiu61I2NIlIGbG1l9RRgVzuG46/svLsWO++upa3n3UtVU1uq1KUSyokQkbzW3BkaaOy8uxY7766lvc/bhryMMcZ4hSUUY4wxXmEJpXlP+ToAH7Hz7lrsvLuWdj1vm0MxxhjjFdZDMcYY4xWWUIwxxniFJZSjiMhEEVknIhtE5Je+jqctROQ5EdkpIvlNypJEZI6IfON8TXTKRUT+6ZzvShEZ1eQz1zr1vxGRa5uUjxaRVc5n/ikdvX1cM0QkS0Q+E5ECEVktInc65QF97iISISJfi8gK57x/55T3EZFFzjm8ISJhTnm483qD837vJm3d65SvE5HvNCn3258LEQkWkWUiMt15HfDnLSJbnO/D5SKS55T5/vtcVe3hPIBgYCPQFwgDVgBDfB1XG85jHDAKyG9S9mfgl87zXwIPOc8nAR8CApwGLHLKk4BNztdE53mi897XwBjnMx8CF/j6nJ24egCjnOexwHpgSKCfuxNLjPM8FFjknM+bwFVO+RPALc7zW4EnnOdXAW84z4c43/PhQB/nZyHY338ugHuAV4HpzuuAP29gC5ByVJnPv8+th3KkU4ANqrpJVWuA14EpPo7phKnqPGDPUcVTgBed5y8ClzQpf0ldFgIJItID+A4wR1X3qOpeYA4w0XkvTlUXqOs776UmbfmUqpao6lLneSVQAGQQ4OfuxL/feRnqPBQ4B/ivU370eTf+e/wXONf5C3QK8LqqHlLVzcAGXD8TfvtzISKZwIXAM85roQucdzN8/n1uCeVIGUBhk9dFTlkg6K6qJeD6xQt0c8qbO+fjlRe5KfcrznDGSFx/rQf8uTvDPsuBnbh+MWwEylW1zqnSNNbD5+e8vw9I5sT/PfzB34GfAw3O62S6xnkrMFtElojIzU6Zz7/PbQvgI7kbJwz066qbO+cTLfcbIhIDvA3cpaoVxxn+DZhzV9V6YISIJADTgMHuqjlfT/T83P3h6fPzFpHJwE5VXSIiZzUWu6kaUOftOF1Vi0WkGzBHRNYep26HfZ9bD+VIRUBWk9eZQLGPYvG2HU5XFufrTqe8uXM+Xnmmm3K/ICKhuJLJK6r6jlPcJc4dQFXLgc9xjZUniEjjH41NYz18fs778biGSE/038PXTgcuFpEtuIajzsHVYwn080ZVi52vO3H9AXEK/vB97uvJJX964OqxbcI1Mdc4CXeSr+Nq47n05shJ+Yc5csLuz87zCzlywu5rpzwJ2Ixrsi7ReZ7kvLfYqds4YTfJ1+frxCW4xnv/flR5QJ87kAokOM8jgS+AycBbHDk5favz/KccOTn9pvP8JI6cnN6Ea2La738ugLP4dlI+oM8biAZimzyfD0z0h+9zn38j+NsD1xUR63GNQd/n63jaeA6vASVALa6/Nm7ENVb8CfCN87XxG0eAx5zzXQXkNmnnBlwTlBuA65uU5wL5zmcexVlxwdcP4AxcXfOVwHLnMSnQzx0YBixzzjsf+I1T3hfX1TobnF+y4U55hPN6g/N+3yZt3eec2zqaXNnj7z8XHJlQAvq8nfNb4TxWN8blD9/ntvSKMcYYr7A5FGOMMV5hCcUYY4xXWEIxxhjjFZZQjDHGeIUlFGOMMV5hCcUEFBGpd1ZgXSEiS0VkbAv1E0Tk1la0+7mI5Hov0s5PRF4QkSt8HYfxH5ZQTKA5qKojVHU4cC/wxxbqJ+BahdYvNbnj2xi/ZwnFBLI4YC+41vcSkU+cXssqEWlcNfZPQLbTq3nYqftzp84KEflTk/auFNe+I+tF5EynbrCIPCwii529Jn7ilPcQkXlOu/mN9Zty9rR4yGnzaxHp55S/ICJ/FZHPgIecfS7eddpfKCLDmpzT806sK0Xkcqf8fBFZ4JzrW87aZojIn0RkjVP3EafsSie+FSIyr4VzEhF51GljBt8uPmgMYItDmsAT6ay6G4Frf5RznPJq4FJ1LRaZAiwUkfdxLVGRo6ojAETkAlxLdZ+qqlUiktSk7RBVPUVEJgG/BSbgWoVgn6qeLCLhwFciMhu4DPhIVX8vIsFAVDPxVjht/gjXOlSTnfIBwARVrReRfwHLVPUSETkH1/IyI4BfO8ce6sSe6Jzbr5zPHhCRXwD3iMijwKXAIFVVZxFJgN8A31HV7U3KmjunkcBAYCjQHVgDPNeq/xXTJVhCMYHmYJPkMAZ4SURycC0/8QcRGYdrqfMMXL8UjzYBeF5VqwBUtem+Mo2LTS7BtVYawPnAsCZzCfFAf1xrIT3nLFb5rqoubybe15p8/VuT8rfUtYIwuJaUudyJ51MRSRaReCfWqxo/oKp7nRV4h+BKAuBag2oBUIErqT7j9C6mOx/7CnhBRN5scn7NndM44DUnrmIR+bSZczJdlCUUE7BUdYHzF3sqrjWZUoHRqlrrrFAb4eZjQvNLdR9yvtbz7c+OALer6kfHNORKXhcCL4vIw6r6krswm3l+4KiY3H3OXayCa9Ok77uJ5xTgXFxJ6DbgHFWdKiKnOnEuF5ERzZ2T0zOztZpMs2wOxQQsERmEa9XY3bj+yt7pJJOzgV5OtUpc2wU3mg3cICJRThtNh7zc+Qi4xemJICIDRCRaRHo5x3saeBbXlszufK/J1wXN1JkH/MBp/yxgl6pWOLHe1uR8E4GFwOlN5mOinJhigHhVnQnchWvIDBHJVtVFqvobYBeu5czdnpMTx1XOHEsP4OwW/m1MF2M9FBNoGudQwPWX9rXOPMQrwAcikodrFeK1AKq6W0S+EpF84ENV/V/nr/Q8EakBZgL/d5zjPYNr+GupuMaYynDNwZwF/K+I1AL7gR818/lwEVmE64+7Y3oVjvuB50VkJVAFXOuUPwg85sReD/xOVd8RkeuA15z5D3DNqVQC74lIhPPvcrfz3sMi0t8p+wTXCrYrmzmnabjmpFbhWoF37nH+XUwXZKsNG+MjzrBbrqru8nUsxniDDXkZY4zxCuuhGGOM8QrroRhjjPEKSyjGGGO8whKKMcYYr7CEYowxxiu61H0oKSkp2rt3b1+HYYwxncqSJUt2qWpqS/W6VELp3bs3eXl5vg7DGGM6FRHZ2pp6NuRljDHGKyyhGGOM8QpLKMYYY7zCEooxxhivsIRijDHGKyyhtIKqsudAja/DMMYYv+Z3CUVEkkRkjoh843xNbKben0VktYgUiMg/nX0b2sX976/m8sfnU1VT116HMMaYTs/vEgrwS+ATVe2Pa8OfXx5dQUTGAqcDw4Ac4GRgfHsFNDGnB1t2H+CB6Wva6xDGGNPp+WNCmQK86Dx/EddOcUdTXPuBhwHhQCiwo70CGpOdzNTx2bz2dSGz8kvb6zDGGNOp+WNC6a6qJQDO125HV1DVBcBnQInz+EhVC9w1JiI3i0ieiOSVlZW1Oai7JwxgaEY8v3xnJTsqqtvcjjHGBCqfJBQR+VhE8t08prTy8/2AwUAmkAGcIyLj3NVV1adUNVdVc1NTW1yKpllhIUH8/aoRHKpt4GdvrqChwTYmM8aYpnySUFR1gqrmuHm8B+wQkR4Aztedbpq4FFioqvtVdT/wIXBae8ednRrDby4awpcbdvHcV5vb+3DGGNOp+OOQ1/vAtc7za4H33NTZBowXkRARCcU1Ie92yMvbrjo5i/OHdOfPs9bxzBebqKtv6IjDGmOM3/PHhPIn4DwR+QY4z3mNiOSKyDNOnf8CG4FVwApghap+0BHBiQh/vmIYY/sl8+CMAib/60sWb9nTEYc2xhi/JqpdZy4gNzdXvbV8vary0eodPDB9DdvLD3LZqAz+b9JgUmLCvdK+Mcb4CxFZoqq5LdXzxx5KpyAiTMxJY84947j1rGw+WFHMeX+dy3vLt9OVkrQxxjSyhOKhqLAQfj5xEDPvOJNeydHc+fpyfvxSHqX77NJiY0zXYgnFS/p3j+XtW8byqwsH8+WGXZz3t7m8s7TI12EZY0yHsYTiRcFBwk1n9mXWneMY3COOe95cwf3vr7YrwYwxXYIllHbQOyWaV286lRvP6MML87fwo+e+Zq+tVmyMCXCWUNpJSHAQv548hEeuHE7elr1Meewr1pVW+josY4xpN5ZQ2tkVozN54yenUV1bz+WPz2fLrgO+DskYY9qFJZQOMLJnIu/cOpYggbveWG5zKsaYgGQJpYNkJkbx+0uHsrywnMc+2+jrcIwxxussoXSgi4anc8mIdP756TcsLyz3dTjGGONVllA62O+m5NA9Npy731huWwobYwKKJZQOFh8Zyl++O4Ituw/w+xkdskCyMcZ0CEsoPjAmO5kfn9mXVxZt46FZazlwyHoqxpjOzxKKj/zs/AFcNiqDxz/fyLl/ad2ikqrKzkpbI8wY458sofhIeEgwf/3uCN6+ZSypseHc+fpyvvvkAtaWVritX9+g/N+0VZz6h09YXbyvg6M1xpiWWULxsdG9Ennvp6fz0OVD2VR2gIsf/YqXF249ordyqK6eO15bxmtfF6IKizbZhl7GGP9jCcUPBAUJ3zu5J7PvHsfY7GR+/W4+t76ylH0HazlwqI6bXsxjxqoSfnXhYHrER7CiyC45Nsb4nxBfB2C+lRwTznPXnszTX2zi4Y/WsWr7FyRFh5G/fR8PXzGMK3OzyNuylxV2D4sxxg/5XQ9FRK4UkdUi0iAizW45KSITRWSdiGwQkV92ZIztKShI+Mn4bN6cOgZVWFtSyeM/HM2VuVkADM9KYMvuKlu92Bjjd/yxh5IPXAY82VwFEQkGHgPOA4qAxSLyvqqu6ZgQ29+ononMvnscew7UkJUUdbh8RFYCACuKyjlrYDdfhWeMMcfwux6Kqhao6roWqp0CbFDVTapaA7wOTGn/6DpWdHjIEckEYGhmPCKwotCu9DLG+Be/SyitlAEUNnld5JQFvJjwEPp3i2F54V5fh2KMMUfwyZCXiHwMpLl56z5Vfa81Tbgpc3tXoIjcDNwM0LNnz1bH6M+GZybwydqdqCoi7v4pjDGm4/kkoajqBA+bKAKymrzOBIqbOdZTwFMAubm5x78VvZMY0TOBt5YUUbT34DFDYsYY4yuddchrMdBfRPqISBhwFfC+j2PqMMMzXRPztgS+Mcaf+F1CEZFLRaQIGAPMEJGPnPJ0EZkJoKp1wG3AR0AB8KaqrvZVzB1tYFos4SFBllCMMX7F7y4bVtVpwDQ35cXApCavZwIzOzA0vxEaHERORrzd4GiM8St+10MxrTMiK4H84n3U2v70xhg/YQmlkxqelUB1bQPrSit9HYoxxgCWUDqtEZnf3jFvjDH+wBJKJ5WVFElSdJjNoxhj/IYllE5KRBieGW9Xehlj/IYllE5seFYC3+zcz37bk94Y4wcsoXRiw7MSUIVVRbZQpDHG9yyhdGKNE/MvzN/MwZp6H0djjOnqLKF0YonRYfzvdwby0eodXPb4fLbsOuDrkIwxXZgllE7up2f34/nrTqa4/CAXPfolc9bs8HVIxpguyhJKADh7UDem334GvZOj+fFLeTzzxSZfh2SM6YIsoQSIrKQo3po6hjP7p/D45xtpaAiIlfqNMZ2IJZQAEhEazMXD09l9oIb1O21JFmNMx7KEEmDGZCcDsGDjbh9HYozpaiyhBJjMxCh6JkVZQjHGdDhLKAFoTN9kFm3eY/MoxpgOZQklAI3JTmbfwVrWlFT4OhRjTBdiCSUA2TyKMcYXLKEEoO5xEfRNiWbBJksoxpiO43cJRUSuFJHVItIgIrnN1MkSkc9EpMCpe2dHx+nvTstO5uvNe6izLYKNMR3E7xIKkA9cBsw7Tp064GeqOhg4DfipiAzpiOA6i7HZyew/VEd+sc2jGGM6ht8lFFUtUNV1LdQpUdWlzvNKoADI6Ij4OovT+to8ijGmY/ldQjlRItIbGAksaub9m0UkT0TyysrKOjI0n0qJCWdA9xjmb9zl61CMMV2ETxKKiHwsIvluHlNOsJ0Y4G3gLlV1O7ajqk+paq6q5qampnoj/E5jTN9k8rbspabO5lGMMe0vxBcHVdUJnrYhIqG4kskrqvqO51EFnjHZyby4YCsri8rJ7Z3k63CMMQGuUw55iYgAzwIFqvpXX8fjr07tk4yIzaMYYzqG3yUUEblURIqAMcAMEfnIKU8XkZlOtdOBa4BzRGS585jko5D9VmJ0GIPT4ux+FGNMh/DJkNfxqOo0YJqb8mJgkvP8S0A6OLROaUx2Mi8v3Ep1bT0RocG+DscYE8D8rodivGtM32Rq6hpYtq3c16EYYwKcJZQAd0rfJIIEG/YyxrQ7SygBLi4ilKEZ8Syw+1GMMe3MEkoXcFp2MssLy6mqqfN1KMaYAGYJpQsYm51Cbb2St2Wvr0MxxgQwSyhdQG6vREKCxOZRjDHtyhJKFxAdHsKIrAS7wdEY064soXQRY7KTWbV9H5XVtb4OxRgToDxKKOJyqohc5tzhfqqzLIrxM2P6JlPfoCzessfXoRhjAlSb75QXkfOBfwPfANud4kygn4jcqqqzvRCf8ZJRvRIJCwli/obdnDOou6/DMcYEIE+WXvkHMEFVtzQtFJE+wExgsAdtGy+LCA1mVM8Em5g3xrQbT4a8QoAiN+XbgVAP2jXtZGx2CmtKKiivqjmifNm2vRTuqfJRVMaYQOFJQnkOWCwivxCRq53HL3DtnPisd8Iz3jQmOxlVWLjp23mUGStLuOKJBfx+RoEPIzPGBII2JxRV/SNwNa5Vf8cAY53nP3DeM35meGYCkaHBLHSGvT5cVcIdry+jvkFZU+J2w0tjjGk1j5avV9UCoEBEIoGeqrrOO2GZ9hAWEkRu70Tmb9zFrPxSbn9tGSOyEjilTxKPf76RyupaYiNstNIY0zYe34ciIhcBy4FZzusRIvK+p+2a9jEmO5n1O/Zz26tLGZoZzwvXn0xur0QA1u+o9HF0xpjOzBs3Nt4PnAKUA6jqcqC3F9o17eD07BQATsqI58UbTiE2IpRBPeIAWFNiCcUY03be2LGxTlX32f2MncPwrARevvEURmQlHB7eSo+PIC4ihLU2j2KM8YA3eij5InI1ECwi/UXkX8D8tjYmIleKyGoRaRCR3BbqBovIMhGZ3tbjdUVn9k89Yq5ERBjUI44CSyjGGA94I6HcDpwEHAJeAyqAuzxoLx+4DJjXirp3Ana9qxcMTotlXWklDQ3q61CMMZ2UxwlFVatU9T5VPVlVc53n1R60V9Caq8VEJBO4EHimrccy3xrcI44DNfUU7rUbHI0xbePxHIqIpAI/x9VLiWgsV9VzPG27BX93jht7vEoicjNwM0DPnj3bOaTOq3FivqCkkl7J0T6OxhjTGXljyOsVYC3QB/gdsAVYfLwPiMjHIpLv5jGlNQcUkcnATlVd0lJdVX3K6Tnlpqamtqb5Lmlg91hEsHkUY0ybeeMqr2RVfVZE7lTVucBcEZl7vA+o6gQPj3k6cLGITMLVK4oTkf+o6g89bLfLigwLpk9yNGtLLaEYY9rGGz2Uxh2bSkTkQhEZiWsZ+3ajqveqaqaq9gauAj61ZOK5wT3iKLB7UYwxbeSNhPKgiMQDPwP+B9ck+d1tbczZqKsI1/pgM0TkI6c8XURmeiFe04xBabFs21PF/kN1vg7FGNMJeTzkpaqN94DsA872QnvTgGluyouBSW7KPwc+9/S4xtVDAVhXWsHoXkk+jsYY09l44yqvPrjuRendtD1VvdjTtk3HGtTDdcFcQUmlJRRjzAnzxqT8u7j2P/kAaPBCe8ZHMhIiiY0IsSu9jDFt4o2EUq2q//RCO8bHRITBaXGsLbWJeWPMifNGQvmHiPwWmI1r+RUAVHWpF9o2HWxQj1jeXlJEQ4MSFGQLfhpjWs8bCWUocA1wDt8Oeanz2nQyjUuwFO09SM/kKF+HY4zpRLyRUC4F+qpqjRfaMj42KM01Mb+mpMISijHmhHjjPpQVQIIX2jF+YGCaawkWu97TbC4AACAASURBVGPeGHOivNFD6Q6sFZHFHDmHYpcNd0JRYSH0To5mrd0xb4w5Qd5IKL/1QhvGjwzuEUv+duuhGGNOjDfulD/uQpCm88ntlcTMVaVs3nWAPim2lL0xpnXaPIciIpUiUuHmUSki9udtJ/adnDQAPswv8XEkxpjOpM0JRVVjVTXOzSNWVeO8GaTpWBkJkYzISuDDVaW+DsUY04l44yovE4AmDU1j1fZ9FP5/9u47vM3qfPj49/bee68sZ0/IZgQCYYURZoGyWkpTaGlp6S4to/xKGW1foKUtYZRC2ZSyR0iAEEKWE7L3tuPEI97bks77hx45dixv2ZLt+3NduiI9Onp0HkfW7bPuU6JbAiulOkcDinLrggmpgHZ7KaU6TwOKciszLowJ6VF80IfdXmv2l/DEZ3v67P2UUp6lAUW16YIJqWzILSO/rLZP3u//3t/GIx/v1G42pfopDSiqTRdYs70+2tL7rZQth8vZlFcOwAebtZtNqf5IA4pq0/DECMakRHpsHMUY0+ZzL605RHCAH9lJERpQlOqnNKCodl0wIZWcg6UUVNT16Dx7CquY9celvLsxv9VzVfU23v76MBdNSuPKqRlszNPZZUr1Rz4XUETkKhHZKiIOEZnWTrkYEXlDRHaIyHYRmd2X9Rws5k9MwRj4eGv3u72MMdz37lYKKuq5952tlNc0tnj+nQ35VDfY+ebMLC6c6Jxdpq0UpfofnwsowBbgcuCLDso9BnxkjBkDTAa293bFBqORyZFkJ0X0aJHjx1sLWL67mGtnZFFa08Aji3e0eP7lNYcYkxLJyVkxZMaFMTE9WgOKUv2QzwUUY8x2Y8zO9sqISBQwB+de9hhjGowxZX1Rv8Fo/oQUVu8/xt6iqi6/trbBzv3vbWNMSiT3LxjPjbOH8uLqQ2zKc/53bc4rZ/Phcq6dkYWIc4fI+RNTtdtLqX7I5wJKJw0HioB/icjXIvK0iLjNYigiC0UkR0RyioqK+raWA8SVUzOJDAnk0idWsGRbQZde+49lezlcVsu9l4wnwN+PO88dRUJEML99awt2h+GlNQcJCfTj0pPSm17j6vbqq0WVRZX1fLj5CHsKq7A72p44oJRqn1cCiogsEZEtbm4LOnmKAOBk4B/GmJOAauBX7goaYxYZY6YZY6YlJiZ66AoGl6z4MN774WlkxYVxy/M5/Hnxzk598R46VsM/l+3l4slpzBoeD0BUSCC/vXAsm/LKeWr5Pt7ekM/Fk9KIDg1s8X4T06N5v48WVd71v83c9uJ65v1lGePv+YhLn1jBPW9vobiqvuMXK6WaeCWgGGPmGWMmuLm93clT5AF5xpjV1uM3cAYY1Usy48L4722ncNXUDP766R6+/dzaVoPrJ7r//W0E+Am/mT+mxfFLJqcxe3g8D364g5oGO9fOzGr12vkTU9mYW9br3V4Hiqv5ZHsB35yZxZ+umsx1M4cQGujPy2tzuezvK9hTqBuNKdVZ/bLLyxhzFMgVkdHWobOBbV6s0qAQEujPw1dO4oHLJrJybzG/eWtzm2U/31nIJ9sKuP2sbFKjQ1s8JyLcf+l4Av2FMSmRnJTZegfpvur2eu6rAwT4CT8+eyRXTs3gdxeN4+WFs3j9e7OpbXBw2d+/YsWe4l6tg1IDhc8FFBG5TETygNnA+yLysXU8TUQ+aFb0h8CLIrIJmAI80Pe1HXxEhG/OzOK2M7N5f9MR1h0sbVWm3mbn3ne2MiwhnO+cNsztebKTIll04zT+dNXkpsH45vqi26u8tpHXcnK5eHIaSVEhLZ6bnBnDWz84hbToUG56dg2vrDnUa/VQaqDwuYBijPmfMSbDGBNsjEk2xpxnHc83xsxvVm6DNTYyyRhzqTGm9Teb6jXfmzOcpMhg7n9vW6sV8IuW7ePAsRruu2Q8wQH+bZ5j7ugkJqRHt/m8q9vrH5/v7ZXxjFfWHKKmwd5m0MuIDeP122Yze0Q8v3pzM//+6oDH66DUQOJzAUX1D+HBAfzsvNFsyC3j3U3Hu6VyS2r422d7mD8xhTmjejYJ4urpmcwcFsdDH+1g1gNLue0/61i2qwiHB2ZiNdod/PurA8waHsf4tLaDWlRIIP/61nTOHJ3IHz/czqFjOpVZqbZoQFHddsXJGYxLjeKhD3dQ12gH4L53t+HvJ/z2wnE9Pn9ceBCvfm82S+6cw7dOGcqqfce46dk1XPnPr9hT2PU1Mc19uOUo+eV13HLa8A7LBvj78cBlE/EX4a63Nrebk0ypwUwDiuo2Z+AYy+GyWv614gBLtxewZHsBPzp7JGkxoR2foJOykyL57UXjWPWbs3n4iknsLapm/uPL+cfne7HZHR2+/lhVfYtyxhie+XI/wxLCOWtMUqfqkBYTyi8vGMPy3cW8uf5wt69FqYEswNsVUP3bKdkJzBubxBOf7SEmLJARieHcfKr7MYmeCg7w5xvTMzlzTCJ3v7WVhz7awYdbjvDg5ZMYlxbVqnxtg52HPtrBc18dICEiiPkTU7l4choAG3PL+P2C8fj5tZ4Q0JbrZw7hra8Pc//72zhjdCIJEcEeuzalBgJtoage+/X8sdQ12skrreX+BRMICujdj1VSZAj/uP5knvjmyRwurWX+48v53gs5bLb2UwHYlFfGRX9dznNfHeDaGZnMHBbPq2tzueqfK7l20SqiQwO5cmpGl97Xz0948IpJVNfbuP+99mepV9Y18os3Nva4a06p/kRbKKrHRiRG8LuLxlFa08Ap2Ql98p4iwoWTUjk1O55nVxzgXyv28/HWAuaOTmRUSiTPLN9PYmQw//nOTE4b6axTVb2NJdsK+HDLEc4cnURYUNc//qOSI7ntzGweX7qbS09KZ+5o911m976zjf+uzwPg4Ssnd/9ClepHZDANME6bNs3k5OR4uxqqF1TUNfLCyoM8vXwfpTWNLJiSxu8vmUB0WGDHL+6iepud+Y8tp7y2kddvPYVhCS3TyL23KZ/bX/qa2LBAGmwO1v52XreCl1K+QkTWGWPa3E7ERbu81IAQFRLID+Zm8+Uvz2LJnWfw2DUn9UowAedYzpM3TMMYuP7p1Rwuq216Lr+slt+8uZnJmTE88c2TqW6w9yj1v1L9iQYUNaCEBweQnRTR6++TnRTBv2+eQUVdI9c/vZqiynrsDsOdr23A5jA8dvUUZo+IZ0h8GG+sy+v1+ijlCzSgKNVNE9Kjee7b0zlaXscNz6zmL5/sZNW+Eu69eDxDE8IREa48OYOV+47p3i5qUNCAolQPTB0Sx1M3TmNfUTVPfLaXCyakcNW047PHLp+agQhNA/RKDWQaUJTqodNGJvCP609m3tgkHrhsYotkl+kxoZw6IoE31uV5JGWMUr5MA4pSHnD22GSevmk6seFBrZ67cmoGeaW1rN5f0u459hZVcedrG7R7TPVbGlCU6mXnjU8hMjig3cH5Q8dquO6p1by5/jDffT6HmgZbH9ZQKc/QgKJULwsN8ueiyal8sPkIVfWtA0V+WS3ffHoVdTY7v71wLDsLKvn5G5s0CaXqdzSgKNUHrpyaQW2jnRdXHaTBdjxRZWFFHd98ahXlNY28cPNMbjl9OL84bwzvbzrCP5bt9WKNleo6Xb6rVB84OSuW8WlR/PHDHTy6ZDdTh8Qyc1gc72zMp7Cynhe+M4OJGc59WW49YzjbjlTwyMc7GZsSxdxOZkRWyts09YpSfaSyrpEVe4pZta+E1ftL2HG0giB/P5779gxmj4hvUba2wc4V//iK3NIaHr/mJGaPiCck8Pjul8YYcg6W8t91eewqqOTRq08iKz6sry9JDRKdTb3icwFFRK4C7gXGAjOMMW4jgIj8BLgFMMBm4NvGmLr2zq0BRfmSspoGAGLCWs8MA+ful5f/4yuKKusJCvBj+tBYTstOpN5m5831hzlUUkNYkD/+IkSHBfL6rbNJjW69D43DYRChxXRmpbqiPweUsYADeBL4mbuAIiLpwJfAOGNMrYi8BnxgjHmuvXNrQFH9TU2DjdX7S/hydzFf7i5mZ0ElInDKiHiuODmD88ansLeoim8+tZqkqGBe+97sFvu0fLqjgLvf3kp8RDCPXzOFIfHh7bybUu51NqD43BiKMWY7dOqvqQAgVEQagTAgv5erplSfCwsKYO7opKY0+YWVdWAgKSqkqcykjBie/dZ0bnx2NTc8s4ZXvjuLepud+97dxvubjzAiMZz9RVVc+PiX/OGyCSyYku6ty1EDnM8FlM4wxhwWkT8Bh4BaYLExZrG7siKyEFgIkJWV1XeVVKoXJEWGuD0+Y1gci26Yxi3/zuGqJ7/iSHkd9TYHPzt3FAvnjKCwso47XtnAHa9sYMWeYu69ZLym1Fce55VpwyKyRES2uLkt6OTrY4EFwDAgDQgXkevdlTXGLDLGTDPGTEtMTPTcRSjlY+aMSuSv3zyJfUXVTEyP5uMfz+H2s0YSFOBHRmwYry6cxe1zs3l9XR6XPrGCwop2hxyV6jKfG0NxEZHPaXsM5SrgfGPMd6zHNwKzjDHfb++cOoaiBoOKukYigwPa7DZevruI772wjpSoEF5eOIvkKPetHqVcBvoGW4eAWSISJs7fmrOB7V6uk1I+ISoksN0xyNNHJvLvm2dQUFHHNYtWcbS8ZUulvKaRF1YeYOfRyl6uqRpofC6giMhlIpIHzAbeF5GPreNpIvIBgDFmNfAGsB7nlGE/YJGXqqxUvzN9aBzPf2cGRZX1XLNoJUfKa9ldUMlv/reZWX9cyu/e3srlf1/BZzsL3b5+f3E1r6w5pBmUVQs+2+XVG7TLS6mW1h0s5aZn1yAClXU2ggL8uHRKGgumpPPAB9vZcbSS+y4Zz/WzhgDQYHPw1PJ9PLZ0Nw02B49cOYmrpmW6PfemvDIiQwIZlqBTlfu7frsOpTdpQFGqtfWHSvnD+9s5a0wS187IIs5KwV9db+OHL3/NpzsK+d6c4Zw7PpnfvLmFnQWVzJ+YwuHSWvLL6/jsZ2cSEdxyxtiB4mrOf+wL4sODWfrTM1qs8lf9jwYUNzSgKNU1NruD37+3jedXHgQgNTqE3y+YwDnjkvn6UCmX/f0rfjB3BD8/b0zTaxwOwzWLVrH5cDm1jXZ+es4ofnj2SG9dgvKAfruwUSnlOwL8/bjvkvGMSo7kcFktP5ib3dQaOSkrlstOSuep5fu5ZnoWmXHOXGLPfXWANQdK+NNVk1myrYC/f76Xb0zP1Nlkg4DPDcorpXyLiHD9rCH88vwxrbq2fnH+aPxFePDDHYBzsP7hj3dw1pgkrjg5nd/MH4vdYXjoox3eqLrqYxpQlFLdlhodyq1njOD9zUdYufcYv3hjI4H+fjxw2UREhKz4MG4+bRhvrj/MhtyyFq/dU1jFHz/YzuGyWi/VXnmaBhSlVI8snDOc1OgQvvt8DmsPlHLPxeNJiT7evXX7WdkkRATz+3e3YoyhrtHOnxfv5ILHvuDJL/ax4G8rWHew1ItXoDxFA4pSqkdCg/z51QVjqKq3NXV1NRcRHMAvzhvN+kNlPPDBds579Av++ukeLpqUxqsLZxEe7M+1i1bx5vo8L12B8hQdlFdK9dglk9MI9Pfj1BEJblfpXzE1g3+vPMBTy/czLCGcF2+ZyanZCQC89f1T+f6L67nztY3sKqji2hmZFFbWU1BRR0FFPfU2O5HBAUSGBBIZEkBMWBDDE8KJDXe/j4zyHp02rJTqE/uKqlix9xhXTc1otS6l0e7g3ne28uLqQ50+X3x4ENlJEYxMjuDsscnMGZmIv59uItYbdB2KGxpQlPJdxhg+3VFISXUDyVEh1i2YkEB/KuoaqaqzUVln41h1PXsLq9lTWMWeoip2Hq2kqt5GanQIV03N4KppmU1TmJVnaEBxQwOKUgNPg83Bku0FvLI2l+W7izAGbjltGL+9aJy3qzZg6MJGpdSgEBTgx/yJqcyfmMrhsloe/WQXT3+5n8mZMVw8Oc3b1RtUdJaXUmrASI8J5YHLJzJ1SCy/eXMzh47VeLtKg4oGFKXUgBLo78dj10xBBH74ytc02h3ertKgoV1eSqkBJyM2jIeumMRtL67nT4t38usLxrYqU9Ng40BxDQeOVXPwWA0OYwgN9CcsyJ/QIH/GpkYxKjnSC7XvvzSgKKUGpAsmpnLdzCyeXLaPcalRhAb6syW/gi2Hy9mWX8HRirp2X+/vJ/zwrGxun5tNgH/rzhyb3eH2+GCmAUUpNWD97qJx5Bwo5Y5XNgDgJ5CdFMHsEfGMSAxnaEI4Q+PDGRIfRlCAH7UNdmoa7FTX2/j753t5dMluvthVxKNXn0RWfBjGGL7ae4znvjrA0u0FXDAxlYevmER4sH6Vgk4bVkoNcHmlNSzbVcSYlCjGpkYSFtT5L/+3Nxzmt29twRi4ftYQlm4vYHdhFXHhQZw+MoF3N+YzIjGCJ2+YyvDEiF68Cu/SdShuaEBRSnVVXmkNd762kTX7S5iYHs1NpwzlokmphAT6s2JPMbe/tB6b3fCXq6dwzrhkb1e3V/TbgCIijwAXAw3AXuDbxpgyN+XOBx4D/IGnjTEPdnRuDShKqe6wOwyFlXWkRIW0ylWWV1rDbf9Zz+bD5ZyWnUBseBARwQFEhgSQFRfGlW5SzbhjjMFhQAA/H0sh058DyrnAp8YYm4g8BGCM+eUJZfyBXcA5QB6wFrjWGLOtvXNrQFFK9Ya6RjsPf7STnIMlVNXZqKizUVXfSF2jg7ToEH523mgunZLeFChKqxt4NSeXV9fmcqS8FrvD0Gh3fhcHBfgxJC6MIfHhDI0PY3hiBDOGxTIiMcJt4s2+0G9XyhtjFjd7uAq40k2xGcAeY8w+ABF5BVgAtBtQlFKqN4QE+nP3xa1TvXy1p5gHPtzOna9t5Jkv97NwznC+3F3MOxvzqbc5mDksjnPGJePvJwT6CQH+flTWNXLwWA0Hj9WwfHcR9TbnOpqEiGBmDY9j+tA46hrtHCqpabpV1tmwOwwOh8FuDMaAiNXase7847qpnDYyoVd/Dj4XUE5wM/Cqm+PpQG6zx3nATHcnEJGFwEKArKwsT9dPKaXadEp2Au/84DTe3ZTPwx/t5I5XNhAa6M+VUzO4cfZQRqe0v87F4TAcKqlh1b5jrNp3jJX7jvHepiMAxIQFMiQujInp0cSGBeHvJ/iJ4CfOYGIMGJz/OowhJTq416/XKwFFRJYAKW6eussY87ZV5i7ABrzo7hRujrntuzPGLAIWgbPLq1sVVkqpbvLzExZMSee88SmsPVDCpIwYokMDO/3aoQnO6c3XzMjCGEN+eR0RwQGdPkdf8kpAMcbMa+95EbkJuAg427gf5MkDMps9zgDyPVdDpZTyrJBAf04fmdijc4gI6TGhHqqR5/ncMk9r9tYvgUuMMW1ldlsLjBSRYSISBFwDvNNXdVRKKdWazwUU4G9AJPCJiGwQkX8CiEiaiHwAYIyxAbcDHwPbgdeMMVu9VWGllFI+OChvjMlu43g+ML/Z4w+AD/qqXkoppdrniy0UpZRS/ZAGFKWUUh6hAUUppZRH+Fzqld4kIkXAwU4WTwCKe7E6vkqve3DR6x5cunvdQ4wxHc55HlQBpStEJKczuWsGGr3uwUWve3Dp7evWLi+llFIeoQFFKaWUR2hAadsib1fAS/S6Bxe97sGlV69bx1CUUkp5hLZQlFJKeYQGFKWUUh6hAeUEInK+iOwUkT0i8itv16c7RORZESkUkS3NjsWJyCcistv6N9Y6LiLyuHW9m0Tk5Gavuckqv9vaUsB1fKqIbLZe87h4a1/SE4hIpoh8JiLbRWSriNxhHR/Q1y4iISKyRkQ2Wtd9n3V8mIistq7hVSszNyISbD3eYz0/tNm5fm0d3yki5zU77rO/FyLiLyJfi8h71uMBf90icsD6HG4QkRzrmPc/58YYvVk3wB/YCwwHgoCNwDhv16sb1zEHOBnY0uzYw8CvrPu/Ah6y7s8HPsS5adksYLV1PA7YZ/0ba92PtZ5bA8y2XvMhcIG3r9mqVypwsnU/EtgFjBvo127VJcK6Hwistq7nNeAa6/g/gdus+98H/mndvwZ41bo/zvrMBwPDrN8Ff1//vQDuBF4C3rMeD/jrBg4ACScc8/rnXFsoLTXtVW+MaQBce9X3K8aYL4CSEw4vAP5t3f83cGmz488bp1VAjIikAucBnxhjSowxpcAnwPnWc1HGmJXG+cl7vtm5vMoYc8QYs966X4lza4N0Bvi1W/Wvsh4GWjcDnAW8YR0/8bpdP483gLOtv0AXAK8YY+qNMfuBPTh/J3z290JEMoALgaetx8IguO42eP1zrgGlJXd71ad7qS6elmyMOQLOL14gyTre1jW3dzzPzXGfYnVnnITzr/UBf+1Wt88GoBDnF8NeoMw49w6ClnVtuj7r+XIgnq7/PHzBo8AvAIf1OJ7Bcd0GWCwi60RkoXXM659zn9sPxcs6vVf9ANLWNXf1uM8QkQjgv8CPjTEV7XT/DphrN8bYgSkiEgP8Dxjrrpj1b1evz90fnl6/bhG5CCg0xqwTkTNdh90UHVDXbTnVGJMvIkk4NyPc0U7ZPvucawulpYG8V32B1ZTF+rfQOt7WNbd3PMPNcZ8gIoE4g8mLxpg3rcOD4toBjDFlwOc4+8pjRMT1R2PzujZdn/V8NM4u0q7+PLztVOASETmAszvqLJwtloF+3RjnhoMYYwpx/gExA1/4nHt7cMmXbjhbbPtwDsy5BuHGe7te3byWobQclH+ElgN2D1v3L6TlgN0a63gcsB/nYF2sdT/Oem6tVdY1YDff29dr1Utw9vc+esLxAX3tQCIQY90PBZYDFwGv03Jw+vvW/R/QcnD6Nev+eFoOTu/DOTDt878XwJkcH5Qf0NcNhAORze5/BZzvC59zr38QfO2Gc0bELpx90Hd5uz7dvIaXgSNAI86/Nr6Ds694KbDb+tf1wRHgCet6NwPTmp3nZpwDlHuAbzc7Pg3YYr3mb1gZF7x9A07D2TTfBGywbvMH+rUDk4CvreveAtxtHR+Oc7bOHutLNtg6HmI93mM9P7zZue6yrm0nzWb2+PrvBS0DyoC+buv6Nlq3ra56+cLnXFOvKKWU8ggdQ1FKKeURGlCUUkp5hAYUpZRSHqEBRSmllEdoQFFKKeURGlDUgCIidisD60YRWS8ip3RQPkZEvt+J834uItM8V9P+T0SeE5ErvV0P5Ts0oKiBptYYM8UYMxn4NfDHDsrH4MxC65OarfhWyudpQFEDWRRQCs78XiKy1Gq1bBYRV9bYB4ERVqvmEavsL6wyG0XkwWbnu0qc+47sEpHTrbL+IvKIiKy19pr4nnU8VUS+sM67xVW+OWtPi4esc64RkWzr+HMi8hcR+Qx4yNrn4i3r/KtEZFKza/qXVddNInKFdfxcEVlpXevrVm4zRORBEdlmlf2Tdewqq34bReSLDq5JRORv1jne53jyQaUATQ6pBp5QK+tuCM79Uc6yjtcBlxlnssgEYJWIvIMzRcUEY8wUABG5AGeq7pnGmBoRiWt27gBjzAwRmQ/cA8zDmYWg3BgzXUSCgRUishi4HPjYGPMHEfEHwtqob4V1zhtx5qG6yDo+CphnjLGLyF+Br40xl4rIWTjTy0wBfme990Sr7rHWtf3Wem21iPwSuFNE/gZcBowxxhgriSTA3cB5xpjDzY61dU0nAaOBiUAysA14tlP/K2pQ0ICiBpraZsFhNvC8iEzAmX7iARGZgzPVeTrOL8UTzQP+ZYypATDGNN9XxpVsch3OXGkA5wKTmo0lRAMjceZCetZKVvmWMWZDG/V9udm//6/Z8deNM4MwOFPKXGHV51MRiReRaKuu17heYIwptTLwjsMZBMCZg2olUIEzqD5ttS7es162AnhORF5rdn1tXdMc4GWrXvki8mkb16QGKQ0oasAyxqy0/mJPxJmTKRGYaoxptDLUhrh5mdB2qu566187x393BPihMebjVidyBq8LgRdE5BFjzPPuqtnG/eoT6uTude7qKjg3TbrWTX1mAGfjDEK3A2cZY24VkZlWPTeIyJS2rslqmWmuJtUmHUNRA5aIjMGZNfYYzr+yC61gMhcYYhWrxLldsMti4GYRCbPO0bzLy52PgduslggiMkpEwkVkiPV+TwHP4NyS2Z2rm/27so0yXwDXWec/Eyg2xlRYdb292fXGAquAU5uNx4RZdYoAoo0xHwA/xtllhoiMMMasNsbcDRTjTGfu9pqselxjjbGkAnM7+NmoQUZbKGqgcY2hgPMv7ZuscYgXgXdFJAdnFuIdAMaYYyKyQkS2AB8aY35u/ZWeIyINwAfAb9p5v6dxdn+tF2cfUxHOMZgzgZ+LSCNQBdzYxuuDRWQ1zj/uWrUqLPcC/xKRTUANcJN1/P+AJ6y624H7jDFvisi3gJet8Q9wjqlUAm+LSIj1c/mJ9dwjIjLSOrYUZwbbTW1c0/9wjkltxpmBd1k7Pxc1CGm2YaW8xOp2m2aMKfZ2XZTyBO3yUkop5RHaQlFKKeUR2kJRSinlERpQlFJKeYQGFKWUUh6hAUUppZRHDKp1KAkJCWbo0KHeroZSSvUr69atKzbGJHZUblAFlKFDh5KTk+PtaiilVL8iIgc7U067vJRSSnmEBhSllFIeoQFFKaWUR2hAUUop5REaUJRSSnmEBhTVLav2HaOkusHb1VBK+RANKKrLGmwOrn96Nf/+6oC3q6KU8iEaUFSXFVTUYXMYymsbvV0VpZQP0YCiuuxoRR0AVfU2L9dEKeVLNKCoLjta7gwo1RpQlFLNaEBRXVagLRSllBsaUFSXaQtFKeWOBhTVZUcqXAHF7uWaKKV8iQYU1WUF5drlpZRqTQOK6jLXLK/qBg0oSqnjfC6giMgjIrJDRDaJyP9EJKaNcneIyBYR2SoiP+7reg5WDodpGpTXMRSlVHM+F1CAT4AJxphJwC7g1ycWEJEJwHeBGcBk4CIRGdmntRykSmoaaLQbte/q+gAAIABJREFUEiKCabQbGmwOb1dJKeUjfC6gGGMWG2Ncf/quAjLcFBsLrDLG1FhllwGX9VUdBzPXDK/spHBAWylKqeN8LqCc4GbgQzfHtwBzRCReRMKA+UCmuxOIyEIRyRGRnKKiol6s6uDg6u4akRgB6MC8Uuo4r+wpLyJLgBQ3T91ljHnbKnMXYANePLGQMWa7iDyEs3usCtholW3FGLMIWAQwbdo045ELGMSOlLcMKDowr5Ry8UpAMcbMa+95EbkJuAg42xjjNggYY54BnrHKPwDkebqeqrWCijr8BIYlaJeXUqolrwSU9ojI+cAvgTOMMTXtlEsyxhSKSBZwOTC7r+o4mB0tryMxMpioUOdHp0oXNyqlLL44hvI3IBL4REQ2iMg/AUQkTUQ+aFbuvyKyDXgX+IExptQLdR10jlbUkRIdSniwM6BoC0Up5eJzLRRjTHYbx/NxDr67Hp/eZ5VSTY6W1zE8MZzwIFcLRQOKUsrJF1soyocdragjJSqECG2hKKVOoAFFdVp1vY3KOhvJ0SHa5aWUakUDiuo0Vw6v1OgQggL8CPL300F5pVQTDSiq01xZhpOjQgAID/bXFopSqokGFNVprhZKSlNACdCAopRqogFFdZprlXxKtDOgRAQH6CwvpVQTDSiq0woq6ogKCSDMmjIcHhygqVeUUk00oKhOO1pe19Q6AWdA0UF5pZSLBhTVaUcr6poG5AEidFBeKdWMBhTVaUfL60ht3kIJ0kF5pdRxGlBUp9jsDoqr6ptmeIGry0sDilLKSQOK6pSiqnocBpJbjKE4u7za2GFAKTXIaEBRneKaMpx6wqC8w0Bdo+4rr5TSgKI66cRV8kBTgkjt9lJKgQYU1UknrpIHmlLY68C8Ugo0oKhOOlpeR5C/H3HhQU3HwrWFopRqxucCiojcLyKbrN0aF4tIWhvlbhKR3dbtpr6u52BztKKO5OhgRKTpmO6JopRqzucCCvCIMWaSMWYK8B5w94kFRCQOuAeYCcwA7hGR2L6t5uBytLyuRXcXOGd5AdQ06Gp5pZQPBhRjTEWzh+GAuzmp5wGfGGNKrL3kPwHO74v6DVYFJ6ySBx2UV0q15HMBBUBE/iAiucB1uGmhAOlAbrPHedYxd+daKCI5IpJTVFTk+coOAsYYjpywSh7QXRuVUi14JaCIyBIR2eLmtgDAGHOXMSYTeBG43d0p3Bxzu7rOGLPIGDPNGDMtMTHRcxcxiJTXNlJvc7RqoeigvFKquQBvvKkxZl4ni74EvI9zvKS5PODMZo8zgM97XDHlVtOU4RNbKEHOMZRqzTislMIHu7xEZGSzh5cAO9wU+xg4V0RircH4c61jqhfkldQCtOryCvD3IyTQT/dEUUoBXmqhdOBBERkNOICDwK0AIjINuNUYc4sxpkRE7gfWWq/5vTGmxDvVHfi25JcjAqNTolo9p7s2KqVcfC6gGGOuaON4DnBLs8fPAs/2Vb0Gs0155WQnRjTN6mpO95VXSrn4XJeX8i3GGDbllTEpI8bt87onilLKRQOKateR8jqKqxqYnBnt9nnt8lJKuWhAUe3alFcG0HYLJdhfZ3kppQANKKoDG/PKCfATxqREun1ex1CUUi4aUFS7NuWVMSY1kpBAf7fPa5eXUspFA4pqk8Nh2JRX3mZ3F2gLRSl1nAYU1aYDx6qprLMxOcP9gDxYAaXBjsOh+8orNdhpQFFt2ny4HGh7QB4gwpXCvrHlwLzDYSivaey9yimlfI4GFNWmjbnlhAT6MTIpos0ybWUcfnvjYWb+cQn5ZbW9WkellO/QgKLatCmvjAlp0QT4t/0xaWtPlG35FdQ1Ong9J69X66iU8h0aUJRbNruDLfntD8iDc6U8tG6h5FoJJV/LydXxFaUGCQ0oyq3dhVXUNTqY1M6APLS9J0puaQ0hgX4cLqtlxd7iXqunUsp3aEBRbm3Ocw3Itx9QIprGUFoOyueW1LBgcjoxYYG8ujbX3UuVUgOMBhTl1sa8MiJDAhgaH95uufBg1yZbx1so5bWNVNTZGJEUzqVT0lm8tYDS6oYO33P57iIe/sjd9jdKqf5AA4pyy7mgMRo/P3e7LR/nblA+t6QGgIzYMK6enkmD3cH/vj7c4Xu+sPIgf/98ry6UVKqf0oCiWqm32dlxtKLDAXlwP204r9QZUDJjwxibGsXkjGheXZuLMe0Pzm+yutl2FVR2t+rdUtdop9Da5lgp1X0+F1BE5H4R2SQiG0RksYiktVHuIxEpE5H3+rqOA932I5U02k27K+RdwoL8EWkZUFwzvDLjQgG4enoWOwsq2WgFDHcKKuqa9q7febRvA8pfP93N/Me/7DDgKaXa53MBBXjEGDPJGDMFeA+4u61ywA19V63Bo6OU9c2JiHOTrYbjg/K5pTVEBgcQHRoIwMWTUwkN9OfVtYfaPM/G3LKm+zv7uIWyNb+C4qp6ynRlv1I94nMBxRhT0exhOOD2z0ZjzFKgb795BontRyqJDQskNTqkU+Wde6K0HEPJiAtDxDn+EhkSyIWTUnlnQ36b4yMb88qa0uT3dQtlf3E1AHmluqpfqZ7wuYACICJ/EJFc4DrabqF09lwLRSRHRHKKioo8U8EBLrekhiHx4U0BoSPhJ6Swzy2tJTM2tEWZq6dnUt1gZ/G2o27PsSmvnFHJkUzKiO7TgNJgczRNIjhcVtNn76vUQOSVgCIiS0Rki5vbAgBjzF3GmEzgReD2nryXMWaRMWaaMWZaYmKiJ6o/4B0qqSErLqzT5SOapbA3xpBXWkPmCa+fmhVLQkQwn+5oHdSNMWzMLWNyZgyjU6I4Vt1AcVV9zy6ikw6V1OBayK8tFKV6JsAbb2qMmdfJoi8B7wP39GJ1VDM2u4P8slounpza6deEBwU0LWwsrmqgrtHRqoXi5yecOTqRT7YVYLM7WuQHO3CshgorTb4rEO08WklCdrAHrqh9ru4u0ICiVE/5XJeXiIxs9vASQFe69aEj5XXYHKZLLZTmXV65rinDbl5/5uhEymsb2ZhX1uK4axLA5MwYRiU7txre0UfdXvuLqwBIiw7RgKJUD/lcQAEetLq/NgHnAncAiMg0EXnaVUhElgOvA2eLSJ6InOed6g4srvEEdwGhLRHB/lQ32Dp8/enZifj7CZ+d0O21IbesKU1+YmQw8eFB7Dxa0er1vWF/cTXx4UGMS4tqWj+jlOoer3R5tccYc0Ubx3OAW5o9Pr3PKjWIuFoYXW2huMZQXH/lZ5zQ5QUQHRbIyVkxfL6rkJ+dN7rp+Ka8ciamH0+TPzolkp0FVd2+hq7YV1TNsIRw0mNCWb2/pE/eU6mByhdbKMqLDpXUEOAnpEa3DghtiWje5VVSQ0JEEGFB7v9WOXN0ElsOVzStTG+0O9hyuGWa/FHJkewuqOyTtPf7i50BJSM2jMo6G+W1uhZFqe7SgKJaOFRSS3psKP4d5PBqLiwogLpGBza7g9zSGtJj227dnDnaOdPu813Obq9dBZXU2xxMzjweUMakRFLTYG9qLfWWqnobhZX1DEsMJ91qUWm3l1LdpwFFtZDbxSnD0CzjcIOd3JLWa1CaG5caRVJkMMt2OgPKxlxnOpbmaV5GpzgH5nt7Pcr+IucMr+EJEU1ddId1YF6pbtOAolrILakho50WhjuujMMVtY3kl9W2O6Av4pw+/MXuImx2B5vyyogJC2wRxFwzvXo7oOyzZngNTwxvumad6aVU9/VKQBGRMb1xXtW7quptHKtu6EYLxRlQ9hZVYXMYMjsISHNHJ1FZZ2P9oTI25jnHT5qvyg8PDiAzLpQdvZzTa39xNSLOCQixYYGEBvpzuEwDilLd1VstlMW9dF7Vi1xTfrsaUFwtFNfaEVeW4bacOjKBAD/hg81H2FVQ6Tar8ejkKHb1dpdXcTXpMaGEBPojImTEhuoYilI90O1pwyLyeFtPAR2nqVU+5/gaks7P8ILjLZQdR5xrRzpqoUSFBDJ1SCyvrD2E3WGY7Car8ZiUSD7bWUi9zU5wgH+X6tNZrhleLumxodrlpVQP9KSF8m1gC7DuhFsO0PF+r8rnHOpmC8U1KL/jaCUikBbTcUCaOyaJukYHAJMyW7dQRqVEYncY9hZWt3quufyy2hbpUzrLGMP+omqGNwsoGbGh2uWlVA/0JKCsBbYYY/594g1NK98v5ZbUEBlyfB+TznJ1ee0prCI1KoSggI4/Vq7pw2nRISRFtk6TP8Y106ug/RXzP39jI9cuWkWj3eH2eWMMX+wqoq7R3uJ4cVUDlfW2li2UmDDKahpbZE5WSnVeTwLKlcAGd08YY4b14LzKS5xp58M6nbbexdXlZXMYMjrZuhmdHElmXCjThsa5fX5YQjiB/sLOo22vmG+0O1h3sJSjFXUs3V7gtsyKPce48dk1PPXFvhbHXa2aYYkRTcd06rBSPdPtgGKMKTHG1ACISKiIjO7oNcq3dTVtvYurhQIdj5+4iAivfW829y+Y4Pb5QH8/RiRGtJvTa2t+BXWNDkTghVUH3ZZ58ou9ALyak9ti5b0rKeSJXV6gixuV6q4ez/ISkYtxtlQ+sh5PEZF3enre/q7B5qDeZu+4oI8wxjgXNcZ3PaAEB/g1razvyoB+anQo0WFtd6+N7mD3xpwDztxbN8wawoo9x9hT2LI1sy2/guW7i5mSGUNeaS0r9hY3PbevuJqgAL8W4z3HV8trC0Wp7vDEtOF7gRlAGYAxZgMw1APn7dd++PJ6vvNcjrer0WlFlfXU21rvY9IZzn3lnQPznW2hdMbolEjyy+uoqHOfXyvnQCmZcaH88KyRBPoLL65u2Up5avk+woL8WXTjVGLCAnllbW7Tc/uKqhkaH9YixUxiRDDBAX4+PTD/6tpDPL50t7eroZRbnggoNmNMuQfOM2CU1zSydHshK/YWU1TZNzsP9tShbqStb87V7eUuy3B3jU2NAmBzXuuPlzGGnIOlTBsSR2JkMBdMSOWNdXnUWGn088tqeXdjPldPzyQpMoTLT8pg8dajlFQ7JyCeOGUYnIExPca316K8npPHf9ro3lPK2zwRULaIyDcBfxEZKSJ/Bb7ywHn7raU7CrA5DMbAZzsKe/W9tuaXe2RWUnenDLu4Bua7G5DcmTksjpBAPxZvbb0P/aGSGoqr6pk6JBaAG2YPobLOxrsb8wH414r9GOA7pznnh1w9PZNGu+HN9XnYHYaDx6oZlhDR6rzpsaE+PSifW1pDYWV9q1lrSvkCTwSUHwLjgXrgZaAC+LEHzttvfbTlKClRIaTHhPJJG7OPPKGspoFLn1jBXxbv6vG5DpXUIHJ8HKGrwoMDCPQXkqNaTwHurrCgAM4YlcjHWwtapbJfe6AUgOnWLLFpQ2IZkxLJ8ysPUlHXyMtrcrlwYmpTjq7RKZGclBXDK2tzySutodFuWgzIu2TEhvnsGEpdo52CCmeL15dbUWrw6nFAMcbUGGPuMsZMN8ZMs+7Xdfd8InK/iGwSkQ0islhE0tyUmSIiK0Vkq1X26p5dhefUNNhYtquI88YnM29sEst3t14D4SnLdhXRaDd8uOVIj/cOyS2pJSUqpNur0iOCA0iP6Vra+844f0IKRyvqWm0bvO5gCVEhAYxMcrYyRITrZw1ha34Fv/rvJqrqbSycM7zFa66dnsWewir+uy4PgGGJ7gJKKMeqG6ht8L0WQPNA52pRKuVLPDHLK1FEHhGRD0TkU9etB6d8xBgzyRgzBXgPuNtNmRrgRmPMeOB84FER8Yl0L8t2FlFvc3DehBTmjUumrtHBij3FHb+wG1zdaUfKW3/hdlVuSU2Puqu+c9ow7jzX8zPHzxqTTKC/8NGWlt1eaw+UcvKQWPyaBbBLT0onPMifDzYf5dTseCakt1yBf+GkVMKD/Hlq+X6AVmMo0GwtStnxL2y7w/CTVzfw2c7e7b7sSPP9YQ4d04CifI8nurxeBHYAw4D7gAM4V9F3izGm+cKDcKDVn97GmF3GmN3W/XygEEjs7nt60sdbjxIbFsiMoXHMHBZPRHAAn2zzfLeX3WFYtquIeWOdX7gfbmk9ztAVh0pqejRDa+6YJC6Z3Kox2WPRoYGcMiKBj7YexRjnR6GspoE9hVVN3V0uEcEBXH5yBgAL54xoda7w4AAumZJGbaOdqJAA4sODWpVJt6YR5zZrDby94TD/+/owL60+5LHr6o48q1Ui0rJ+SvkKTwSUeGPMM0CjMWaZMeZmYFZPTigifxCRXOA63LdQmpedAQQBe9t4fqGI5IhITlFRUU+q1aEGm4Ol2wuZNzaZAH8/ggL8OGN0Iku2F3p8O9sNuWWU1jSyYEoap2Yn8OGWI01fuF1V12inoLKu2wPyve38CSkcPFbTlM143UHn+IlrQL65n5wzioevnMSckQluz3XN9CzAuULeXUYA15iLa2C+0e7g0SXOabqr9x3D3gfbErclt7SW4AA/shMjtMtL+SRPBBTXIoEjInKhiJwEZLT3AhFZIiJb3NwWAFjjMJk4Wz+3t3OeVOAF4NvGGLfJnIwxi6yxnWmJib3biPlqbzGV9TbOn5DSdOycsckUV9X3uEvqRJ/vLMTfT5gzMpELJqSQW1LL1vz281615XBZLcZAVrznpvx60jnjkhGhqRW29kApgf7iNktxXHgQ35iW2Wb6mEkZ0cwYFsesYe5TviRFBhPoL03jFW+sy+NQSQ2XTE6jos7G9iPd+xl7wqFjNWTEhpIVF9aUGVopX+KJgPJ/IhIN/BT4GfA08JP2XmCMmWeMmeDm9vYJRV8CrnB3DhGJAt4HfmuMWdXzy+i5j7ceJTzIn1Ozj/91fOboRPz9hCUenu316Y5CpmbFEh0WyDnjUvD3Ez7ccqRb52pag+LBRYmelBARzPShcXxsBZR1B0sYnxZNaFDXJxCICK8unMWv5491+7yfn5AW48w6XG+z89eluzkpK4bfWOVX7j3W/QvpodxSZ2qczLgwDpXUdLtFqlRv8cQsr/eMMeXGmC3GmLnGmKnGmG6nXhGRkc0eXoJzfObEMkHA/4DnjTGvd/e9PMnuMCzeWsDcMUmEBB7/oosJC2L60FiWbPPcgG5BRR1b8ys4c4yzxRUXHsSs4XF8uPlop75kjpTXtuiCy+vhGpS+cMGEFHYWVLLzaCUb88qZ5qa7q7M6Sn7p2mjrlTW55JfX8dNzRpMSHcLwhHBW7vNiQLEmTmTFhVHTYG9apKmUr/DELK9hIvIXEXlTRN5x3Xpwyget7q9NwLnAHdb7TBORp60y3wDmAN+yphdvEJEpPbuSnll3sJRj1Q0turtc5o1NZmdBpcdm5nxuzTY6a0xS07ELJqSyr7iaXQVtZ+etrrdx1/82M/uPn3LJE1/y5W7n7LNDJTUEB/iRGBnskfr1hvPGO3+uj3y8kwabo80sxZ6QHhPKgeJq/vbZHmYOi+PU7HgAZo2IZ83+EmxtpMrvTeU1jVTU2ciMDWuajafjKMrXeKLL6y2cM7v+Cvy52a1bjDFXWN1fk4wxFxtjDlvHc4wxt1j3/2OMCTTGTGl2c5tKv698tOUoQQF+nDk6qdVz54xLBvBYt9dnO4pIjQ5hdHJk07FzxzvHGT7Y7L7ba+2BEi54bDkvrTnElVMzKKtp5PpnVnPDM6tZc6CUzLiup63vS2kxoUzOiG76GbobkPeUjNgwSmsaKaqs56fnjm76ucweHk9VvY0t3Ryr6gnXlOHMuNCmlqQGFOVrPBFQ6owxjxtjPrNmeS0zxizzwHn7jbpGOx9tOcLp2QktUrm7DIkPZ2RSRIcBpdHu6LDLqsHm4Ms9xcwdk9QiACRFhjB9aFyr9Rq1DXYe+GA733hyJQbDK9+dxZ+umszSn57B7y4ax5bD5WzMLetWUsi+dv6EVMC5fqQ3W1OuqcNzRiUyo9ng/czhzvveGEdxDcJnxIY1ZXT21RX9avDyREB5TETuEZHZInKy6+aB8/YLdofhx69sIL+8jutnD2mz3LxxyazeX8JHbQycbzlczhkPf8bZf1nGOxvz25xmnHOghKp6G3PdtIRc4wx7i6qoqrfxz2V7Of3hT1n0xT6unZHFR3fMYeZwZ/dNcIA/3zltGMt+MZdfnj+G287M7sbV963zxjtber3ZOgGYnBlDYmQwvziv5ULNpMgQspMivDKO4mqhZMWHERYUQEJEsC5uVD6n9Z/TXTcRuAE4C3B1Lhvr8YBmjOH3727lo61H+d1F49x+ybvcctowvtp7jFv/s56fnjOK28/KbmphfLD5CHe+toG4sCAC/fz40ctf8/fP9vCTc0Zx7rjkFi2RT3cUEuTvxykj4lu9x/kTUrjv3W38+s3N7DxaSXltI6ePTOBHZ49stQjQJSokkNvObL0I0BcNT4zgdxeNaxrT6C3ZSRGsvWue2+dmD4/nv+vzaLQ7CPT3xN9jnXOopIbo0ECiQpz7x2TGhWqXl/I5nggolwHDjTGDbsrJk1/s498rD3LLacOastq2JT4imFcXzuJX/93Enz/Zxa7CKh66YiJPfbGf/7dkFydnxfDkDdOIDw/ivc1HePSTXXzvhXWMSo7gvPEpnDUmickZMXy2s5CZw+Oasvs2lxodyrQhsazZX8K8scncflY2UzJ9IiONx3T0c+5ts0fE88Kqg2zKK2PqkN6bGHCi3JLaFpuXZcWFNS3wVMpXeCKgbARicKY/GTTe+vowD364g4smpTatUehISKA//+/qKYxOieLhj3ewfHcRZTWNXH5yOn+8fGJTYsZLJqcxf0IK//v6MK/n5PHEZ3v466d7iA8P4lh1A9fNbLtr7YnrTqayzkZ2UuvU7KrnZlldhiv3HuvbgFJa02ISRlZcGO9uzO/zlpJS7fFEQEkGdojIWpwp7AEwxlzigXP7hMeW7ObtDYexG4PNbnAYQ2FlPbOGx/Hnb0xukaCwIyLCbWeOYGRSBPe8s5XbzhjBwjnDW82wCvD346ppmVw1LZOymgaW7Spi6fZCdhVUMn9iapvnT44KITmq25eqOhAXHsSYlEhW7Svh9j7q1HU4DHmltcwbm9x0LDM2DIdxbiQ2JL51kkulvMETAeUeD5zDp6VEBzMuLYoAP8HPTwjwE2LDgvj+3Oxup3ufNy6ZeeOSOy6Ic3HkginpLJiS3q33Up41a3g8r6w9RL3N3uH/f6PdwbGqBoqr6imuqmdcahRJXdwzprCyngabo0U2aNf93BINKMp39DigDIYpwldPz+JqK6mgUrNHxPPcVwfYmFvOjGFx7C2qYtGyfSzZ7typ08XhMFSesJvmqOQI3v/R6V3qpmpag9JsandWvK5FUb6n2wFFRCpxk1oeEMAYY7TjRQ1Is4bFIwKvrDnEv1bs56OtRwny92P+xFSiQwNblI0JCyQhIpiEiGAKK+u4++2tPL18f5dm1rnWoDRvoaREhRDoL50KKD97fSOZsWHcMW9kh2WV6oluBxRjTGTHpZQaeKLDAhmXGsWbXx8mMiSA7585gm+dMqxTiy2/3F3MY0t3cdGk1E5vaJZb4lzA6FpwCeDvJ6THhLbYdMud8ppG3lyfR2p0KD86O9unsyGo/k+nhyjVDb+9cBz3XDyOr351Fj8/b0ynV+7fe8l4/EW4++0tnc4WfKikhpSokBZJR8HZYukojf2KvcU4jHOLAldgUqq3aEBRqhtmj4jn26cOIzIksOPCzaTFhPKTc0bx2c6iTu+ymVta02INikuWlca+PV/sKiLAmoW4cl/vbEWtlIsGFKX62LdOGcq41Cjue3crlXWNHZbPa2N75sy4MMpqGqlo4xzGOLeJPmdcMgkRQV7dy0UNDp6YNqyU6oIAfz8euHwil/19BT95dSPZSRFN04rrGu08cuXkpvGVBpuDIxV1ZLgZb8lqmjpcw/i06FbP7y6s4kh5HXecnYifn7By3zGMMTqOonqNtlCU8oIpmTHcNHsoS7YX8MyX+1ixp5hjVQ1szC1vMb7i2p7ZXTbo5gHFnS92FQHOrMmzh8dTUFHPAU0oqXqRtlCU8pJ7Lh7HT84ZRVRIQFOr4enl+/i/97fz8dajnD8htSlYuNtN09UN1tZg+7JdRYxMiiAtJpTZI46njBmWoAshVe/wuRaKiNwvIpusXRgXi0iamzJDRGSdVWariNzqjboq1RMiQnRoYIsuqG+dMpQxKZHc9+42quttzTbWah1QosMCiQoJcDswX9tgZ/X+Es4Y5dwmenhCOEmRwV7dwlgNfD4XUIBHrN0apwDvAXe7KXMEOMUqMxP4lbvAo1R/E+Dvxx8um8CR8joeW7qb3JJaAv2F5DbStWTFu5/ptWr/MRpsDs4Y7QwoIsLsEfGs3Hus09OVleoqnwsoxpjm+6uG42Y1vjGmwRjjSkQZjA9eh1LdNXVIHNdMz+SZL/ezbFcR6TGh+LeRgDQrLszt4sZlO4sICfRrsQ/O7OHxFFfVs7eoqtfqrgY3n/wiFpE/iEgucB3uWyiISKaIbAJygYeMMfltlFsoIjkiklNUVNR7lVbKg355/hiiQwPZfqSi3RX1mbFh5JXUttrh84tdRcwaHt9iMWTzcRSleoNXAoqILBGRLW5uCwCMMXcZYzKBF4Hb3Z3DGJNrjJkEZAM3iYjb1L3GmEXGmGnGmGmJiYm9dUlKeVRseBC/umAM4H78xCU7KYIGu4O/fLKrKajkltSwr7i6afzEJSsujLToEB1HUb3GK7O8jDHu91dt7SXgfdpJkW+MyReRrcDpwBseqJ5SPuHKkzM4UFzN2WPb3ubg0pPSWXewlL99tod9xVX8+aopLGs2Xbg5EWHWiHg+31mEw2G6tI+PUp3hc9OGRWSkMWa39fASYIebMhnAMWNMrYjEAqcCf+nDairV6/z8hF+cP6bdMoH+fvzx8omMSIzggQ+3k1e6kpBAfzJiQxnuZnrw7OHxvLn+MLsLqxidovldlWf5XEABHhSR0YADOAjcCiAi04BbjTG3AGOBP4uIwZku/0+U9eKSAAAgAElEQVTGmM3eqrBS3iQifHfOcIYmhHPHK19T02DnuplZblfEHx9HKdaAojzO5wKKMeaKNo7nALdY9z8BJvVlvZTydeeMS+b1W2fzf+9t55o2NoTLiA0jMy6UlfuO8a1Th/VxDdVA53MBRSnVfePTonl54ax2y5wyPIFXc3KZ/oclpEWHkBYTysikCG49cwRhQfqVoLpPPz1KDTI/mjeStJhQ8stqyS+vZVdBZVMq/TvPHe3l2qn+TAOKUoNMekxoq+2Ab39pPU8t38/1s4eQFOl+Vb5SHfHJhY1Kqb71s3NH02h38Nele7xdFdWPaUBRSjE0IZxrZmTy8ppDHCiu9nZ1VD+lAUUpBcCPzh5JoL8ff1q809tVUf2UBhSlFABJkSHccvow3tt0hM155d6ujuqHNKAopZosnDOc2LBAHvqoVYIKpTqkAUUp1SQyJJAfzM3myz3FLN+t2blV12hAUUq1cMPsIWTGhXL321upa7R7uzqqH9GAopRqITjAnz9eNon9xdU8tnR3xy9QyqIBRSnVymkjE/jGtAwWfbGPLYd1gF51jgYUpZRbd80fR1x4EL94YxONdoe3q6P6AQ0oSim3osMCuX/BBLYdqWDRF/u8XR3VD2hAUUq16fwJKVw4MZXHluxmT2GVt6ujfJwGFKVUu+69ZDyhQf7c8crXlNc0ers6yof5XEARkftFZJOIbBCRxSKS1k7ZKBE5LCJ/68s6KjWYJEYG8+jVU9hdUMU1T62iuKre21VSPsrnAgrwiDFmkjFmCvAecHc7Ze8HlvVNtZQavOaOSeLpm6axv7iKq59cydHyOm9XSfkgn9sPxRhT0exhOGDclRORqUAy8BEwrQ+qptSgNmdUIs/fPJObn1vLVU9+xUu3zCI4wI/Nh8vZcriCXQWVJEYGMzI5gtHJkYxMjiQ6NNDb1VZ9SIxx+33tVSLyB+BGoByYa4wpOuF5P+BT4AbgbGCaMeb2Ns61EFgIkJWVNfXgwYO9WXWlBrwNuWXc9Owaqutt2BzO7w8RyIwN4/+zd9/xbVfn4sc/j7yHvFdsJ3HsJM6eDiRhJWwohQZKSxdQaKF03I57uYXS3+W2l7a0tNwW6OJSSmmBFigpZRUCCWElgew4iZ0423YS7z1lnd8f+sqRY8krsiXbz/v10svS0dFX5ySyH51d3dROc8ep1fVXz8/kfz+9gBCbBKq4yg9EZIsxpt8v7gEJKCLyJpDh5al7jDEveuS7G4g0xtx72uu/DkQbY34mIjfTR0DxVFBQYDZv3nxmhVdKUXyikWc+PMqkpGjmZscza0IcMRGhGGMoq2tl/8km3tlfyR/fP8wt50zhvz4+K9BFVmdgoAElIF1expiLB5j1aeAV4N7T0pcB54nIV4FYIFxEmowxd/mxmEopH/Iz7Pz31bN7pYsI2YnRZCdGs3JGGoLw+PuHmJgUxRfPmRKAkqqRFHRjKCIyzRjj3kDoaqDXPtrGmM955L8ZVwtFg4lSQeaej82krK6FH768h8yEKC6b7a1jQo0VwTjL634RKRSRncClwDcBRKRARB4LbNGUUoMRYhN++emFzMtO4Jt/3cb2Y3WBLpIaRkE5KD9cdAxFqcCoampn1W/ep7HNwYOfms+FM9IDXSQ1CAMdQwnGFopSaoxJiY3gL7eezYT4KG55YjM/emUPHY6+N5xs6+zi+//YxfW/+4A/fXCYmuaOESqtGiptoSilRkxbZxc/fnUvT244wvzseB7+zCImJUf3ylfZ2M7tf97M1qN15KbGcLCymbAQYWV+Gp89exIr8tN8vse9LxaSGBPOty6ePpxVGVeCepaXUmp8igwL4YfXzGF5XjJ3Pr+TKx96l08uzua6RdnMyYpDRCg60cCtT2ymurmd33xuEVfOncDe4w28sLWU1dvKeWPPSf5621KW5ib3uv72Y3X8aYNrrdlZOUksn5oy0lUc17SFopQKiGM1Ldz/ryLW7D5JR5eT6emxrMxP4y8bjxAbGcr/3VjAvOyEHq9p6+xixQNvk5kQyd/vWI5IzwWTtz7xEVuO1pIQFUaXMbz+rfOJDtfvzWdKx1CUUkFtYlI0v/7sIj6652J+tGoO9sgwfv/OQXJTY3nxa+f2CibgauH820XT2Hq0jrVFFT2eKyyr562iCr507hR+et08jtW08vPX941UdRTa5aWUCrD46DA+d/ZkPnf2ZKqa2kmICiM0xPd33esLsnn0nQM88HoxK/PTsFnbujz01n7iIkO5cXkOcZFhfGHpZP74wSE+Nm8CiycnjlR1xjVtoSilgkZKbESfwQQgLMTGty+ZTtGJRl7aWQ7A3uMNvLHnJF88Zwpxka4NKf/z8nwmxEXy3b/vpN3R1dcllZ9oQFFKjTofn5fJjAw7D67ZR2eXk0fWlhAbEcotHtu72CPD+PG1cympaOJXb+7v42rKXzSgKKVGHZtNuPOyfI5Ut/CTV4t4tfA4Ny2fTHx0z+3yV+Sncd2ibH7z9gFue3Izx2paAlTi8UEDilJqVLpwRhqLJyfy+PuHiAoL4dZzc73m+8m1c/nPy/N5d38VFz+4nl++uY+2Tu0CGw4aUJRSo5KIq5UC8IWlk0mKCfeaLzzUxldXTGXtf1zApbMz+OWb+7n4wfUcrmoeyeKOC7oORSk1qm0+XMOcrHgiw0IGlH/DgWq+8pct5KTE8PxXlhHWzyQApetQlFLjREFO0oCDCcCyvGR+cu1cdhyr4+G3dLDenzSgKKXGnSvnTuC6Rdk8sq6EzYdrAl2cMUMDilJqXPrvq2eRlRjFt5/dTmNbZ6CLMyZoQFFKjUv2yDD+91MLKKtt5b//uac73ek0VDW1s/d4A+/tr+LF7WX88f1DPLnhsM4O60fQbb0iIv8DXAM4gQrgZmNMuZd8XcAu6+FRY8zVI1dKpdRYUJCTxNdXTuWhtSXsO9lITXMHFY1tdHZ5n6z0t4+O8dvPLfa65X5Lh4OwENu4HuQPulleIhJnjGmw7v8bMMsY8xUv+ZqMMbGDubbO8lJKna6zy8n3VxdSWtdCuj2StLhIMuIiSLVHkhwbTkpsOEkxEWw9Ust3nt0OwC8+tYBLZrlOnTxc1cwf3jvEc1uOkRgdzs+vn885Y2zb/IHO8gq6gOJJRO4GJhlj7vDynAYUpdSIOlbTwh1PbaGwrIGbl+dwsqGNf+0+QZjNxsfnZ7LtWC0HK5u5eXkOd10xY1Czz4LZqA4oIvIj4EagHlhpjKn0kscBbAccwP3GmH/4uNZtwG0AkyZNWnzkyJFhK7dSauxr6+zif17ew1ObjhIXGcrnl07m5uU5pMVF0trRxU//VcQTHxwmLzWGBz+1gPkTe2/DP9oEdUARkTeBDC9P3WOMedEj391ApDHmXi/XyDTGlItILrAWuMgYc6Cv99UWilLKXw5UNpERF0lMRO+h6Pf2V3Hn8zuoaGzn9vNz+beLpo3q1kpQB5SBEpHJwCvGmDn95HsCeNkY83xf+TSgKKVGSn1rJ/e9vIfntpQyNS2Wn31yHosmjc5zWUbtSnkRmebx8GqgyEueRBGJsO6nAOcAe07Pp5RSgRIfFcYD18/niS8uoaXdwXW//YAfvLSbd/dXcriqud8zWo7VtHD3C7u46Bdv89i7B0fFmS5B10IRkb8D+bimDR8BvmKMKRORAuv+l0RkOfB7K48N+KUx5g/9XVtbKEqpQGhs6+QnrxXx9Kaj3WkikG6PZFZmHIsnJ7IkJ4l52fGcqG/j1+tKeGFbGSEi5GfY2VVWz8SkKO68bAZXzZ3QfUolgKPLSV1rJ41tDhqsn22dXYSGCOEhNkJDbISFCHlpsd2Hjw3WmOjy8jcNKEqpQDpR38bh6mZKa1sprW3haE0LO0vrKaloAiA8xIbD6SQsxMZnzprEVy7IIyM+knf2VfKT14rYe7yBOVlxZMRFcbKhjRMNbVQ1tTOQP+NPfHEJK/LThlTugQaUoFvYqJRSY1VGfCQZ8ZG90muaO9hypJbNh2sIC7Fx4/LJpNlP5Tt/eirnTk3hH9vL+P36g5TWtpAeF8msCXGkx0eSHBNOXFQo9ogw4qLCiAyz0dllcHQ56ewydDqdzM2KH/b6aQtFKaVUn0btoLxSSqnRSQOKUkopv9CAopRSyi80oCillPILDShKKaX8QgOKUkopv9CAopRSyi/G1ToUEanEtZ3LQKQAVcNYnGCl9R5ftN7jy1DrPdkYk9pfpnEVUAZDRDYPZCHPWKP1Hl+03uPLcNdbu7yUUkr5hQYUpZRSfqEBxbdHA12AANF6jy9a7/FlWOutYyhKKaX8QlsoSiml/EIDymlE5HIRKRaREhG5K9DlGQoReVxEKkSk0CMtSUTWiMh+62eilS4i8pBV350issjjNTdZ+feLyE0e6YtFZJf1modERAgCIjJRRNaJyF4R2S0i37TSx3TdRSRSRD4UkR1WvX9gpU8RkU1WHf4mIuFWeoT1uMR6PsfjWndb6cUicplHetD+XohIiIhsE5GXrcdjvt4ictj6HG4Xkc1WWuA/58YYvVk3IAQ4AOQC4cAOYFagyzWEepwPLAIKPdJ+Btxl3b8L+Kl1/0rgNUCApcAmKz0JOGj9TLTuJ1rPfQgss17zGnBFoOtslWsCsMi6bwf2AbPGet2tssRa98OATVZ9ngVusNJ/B9xh3f8q8Dvr/g3A36z7s6zPfAQwxfpdCAn23wvgO8DTwMvW4zFfb+AwkHJaWsA/59pC6eksoMQYc9AY0wH8FbgmwGUaNGPMO0DNacnXAH+y7v8J+IRH+pPGZSOQICITgMuANcaYGmNMLbAGuNx6Ls4Ys8G4PnlPelwroIwxx40xW637jcBeIIsxXner/E3WwzDrZoALgeet9NPr7f73eB64yPoGeg3wV2NMuzHmEFCC63ciaH8vRCQb+BjwmPVYGAf19iHgn3MNKD1lAcc8HpdaaWNBujHmOLj+8ALuw6V91bmv9FIv6UHF6s5YiOvb+pivu9Xtsx2owPWH4QBQZ4xxWFk8y9pdP+v5eiCZwf97BINfAv8JOK3HyYyPehvgDRHZIiK3WWkB/5zrmfI9eesnHOvT4HzVebDpQUNEYoG/A98yxjT00f07ZupujOkCFohIArAamOktm/VzsPXz9sUz4PUWkauACmPMFhFZ4U72knVM1dtyjjGmXETSgDUiUtRH3hH7nGsLpadSYKLH42ygPEBl8beTVlMW62eFle6rzn2lZ3tJDwoiEoYrmDxljHnBSh4XdQcwxtQBb+PqK08QEfeXRs+ydtfPej4eVxfpYP89Au0c4GoROYyrO+pCXC2WsV5vjDHl1s8KXF8gziIYPueBHlwKphuuFttBXANz7kG42YEu1xDrkkPPQfkH6Dlg9zPr/sfoOWD3oZWeBBzCNViXaN1Psp77yMrrHrC7MtD1tcoluPp7f3la+piuO5AKJFj3o4B3gauA5+g5OP1V6/7X6Dk4/ax1fzY9B6cP4hqYDvrfC2AFpwblx3S9gRjA7nH/A+DyYPicB/yDEGw3XDMi9uHqg74n0OUZYh2eAY4Dnbi+bdyKq6/4LWC/9dP9wRHg11Z9dwEFHte5BdcAZQnwRY/0AqDQes0jWAtkA30DzsXVNN8JbLduV471ugPzgG1WvQuB/7LSc3HN1imx/shGWOmR1uMS6/lcj2vdY9WtGI+ZPcH+e0HPgDKm623Vb4d12+0uVzB8znWlvFJKKb/QMRSllFJ+oQFFKaWUX2hAUUop5RcaUJRSSvmFBhSllFJ+oQFFjSki0mXtwLpDRLaKyPJ+8ieIyFcHcN23RWTcnUHeFxF5QkQ+GehyqOChAUWNNa3GmAXGmPnA3cBP+smfgGsX2qDkseJbqaCnAUWNZXFALbj29xKRt6xWyy4Rce8aez+QZ7VqHrDy/qeVZ4eI3O9xvevFde7IPhE5z8obIiIPiMhH1lkTt1vpE0TkHeu6he78nqwzLX5qXfNDEZlqpT8hIg+KyDrgp9Y5F/+wrr9RROZ51OmPVll3ish1VvqlIrLBqutz1t5miMj9IrLHyvtzK+16q3w7ROSdfuokIvKIdY1XOLX5oFKAbg6pxp4oa9fdSFzno1xopbcBq4xrs8gUYKOI/BPXFhVzjDELAETkClxbdZ9tjGkRkSSPa4caY84SkSuBe4GLce1CUG+MWSIiEcD7IvIGcC3wujHmRyISAkT7KG+Ddc0bce1DdZWVPh242BjTJSIPA9uMMZ8QkQtxbS+zAPh/1nvPtcqeaNXt+9Zrm0Xku8B3ROQRYBUwwxhjrE0kAf4LuMwYU+aR5qtOC4F8YC6QDuwBHh/Q/4oaFzSgqLGm1SM4LAOeFJE5uLaf+LGInI9rq/MsXH8UT3cx8EdjTAuAMcbzXBn3ZpNbcO2VBnApMM9jLCEemIZrL6THrc0q/2GM2e6jvM94/Pxfj/TnjGsHYXBtKXOdVZ61IpIsIvFWWW9wv8AYU2vtwDsLVxAA1x5UG4AGXEH1Mat18bL1sveBJ0TkWY/6+arT+cAzVrnKRWStjzqpcUoDihqzjDEbrG/sqbj2ZEoFFhtjOq0daiO9vEzwvVV3u/Wzi1O/OwJ8wxjzeq8LuYLXx4A/i8gDxpgnvRXTx/3m08rk7XXeyiq4Dk36jJfynAVchCsIfR240BjzFRE52yrndhFZ4KtOVstM92pSPukYihqzRGQGrl1jq3F9y66wgslKYLKVrRHXccFubwC3iEi0dQ3PLi9vXgfusFoiiMh0EYkRkcnW+/0f8AdcRzJ782mPnxt85HkH+Jx1/RVAlTGmwSrr1z3qmwhsBM7xGI+JtsoUC8QbY14FvoWrywwRyTPGbDLG/BdQhWs7c691sspxgzXGMgFY2c+/jRpntIWixhr3GAq4vmnfZI1DPAW8JCKbce1CXARgjKkWkfdFpBB4zRhzp/UtfbOIdACvAt/r4/0ew9X9tVVcfUyVuMZgVgB3ikgn0ATc6OP1ESKyCdeXu16tCst/A38UkZ1AC3CTlX4f8Gur7F3AD4wxL4jIzcAz1vgHuMZUGoEXRSTS+nf5tvXcAyIyzUp7C9cOtjt91Gk1rjGpXbh24F3fx7+LGod0t2GlAsTqdiswxlQFuixK+YN2eSmllPILbaEopZTyC22hKKWU8gsNKEoppfxCA4pSSim/0ICilFLKL8bVOpSUlBSTk5MT6GIopdSosmXLlipjTGp/+cZVQMnJyWHz5s2BLoZSSo0qInJkIPm0y0sppZRfaEBRSinlFxpQlFJK+UXQBRTrBLndIuKUPs7wFtdZ4M+LSJGI7LXOvlBKKRUgQRdQgEJcp92900++XwH/MsbMAOYDe4e7YEoppXwLullexpi9ANZpc16JSByu0+Nutl7TAXSMQPGUUkr5EIwtlIHIxXVGwx9FZJuIPGYdAKTGkNLaFq777QdUNrb3n1kpFXABCSgi8qaIFHq5XTPAS4TiOgHvt8aYhbiOS73Lx3vdJiKbRWRzZWWln2qgRsIHB6rZcqSWXWV1gS6KUmoAAtLlZYy5+AwvUQqUGmM2WY+fx0dAMcY8CjwKUFBQoHv1jyKHqlzHqmsLRanRYVR2eRljTgDHRCTfSroI2BPAIqlhcLCyCdCAotRoEXQBRURWiUgpsAx4RURet9IzReRVj6zfAJ6yztleAPx45EurhpO7hVKhAUWpUSEYZ3mtBlZ7SS8HrvR4vB3wuU5FjW5dTsPh6hZAWyhKjRZB10JRCqCstpUOhxPQgKLUaKEBRQWlg1Wu8ZMJ8ZHa5aXUKKEBRQWlg5Wu8ZOluclUNrZjjE7QUyrYaUBRQelQVTNxkaHkZ9hp7eyiuaMr0EVSSvVDA4oKSgermpiSGkuaPQLQcRSlRgMNKCooHapsJi8lhlQroFQ0tAW4REqp/mhAUUGnpcNBeX0bUzwCSmWTtlCUCnYaUFTQOVzlWn+SmxpLmj0S0C4vpUYDDSgq6LinDE9JiSEhKoxQm+jUYaVGAQ0oKui4pwxPSYnBZhNSYiO0haLUKKABRQWdQ1XNZMZHEhUeAkBanAYUpUYDDSgq6BysbCI3Nbb7caq2UJQaFTSgqKBijOFgVTNTUk4dwJlqj9AxFKVGAQ0oKqhUNXXQ2OYgN/VUQEmzR1DT3E6XU7dfUSqYaUBRQcV9BsrpLRSngepmbaUoFcw0oCi/WL+vkld2Hj/j67hPaczzHEPR7VeUGhU0oCi/+M26En7y2t4zvs7BqmbCQ21kJkR1p3Vvv6IBRamgpgFF+UVlUzulta00tTvO6DoHK5vJSY4mxCbdabpaXqnRQQOK8gv3H/t9JxvP6DoHq5p6jJ8ApMRql5dSo4EGFHXG2jq7aGxztUz2nRh6QHF0OTla3dJjDQpAVHgI9ohQDShKBbmgCygicr2I7BYRp4gU+MiTLyLbPW4NIvKtkS6rcvH8Q190BgHlWG0rDqch97QWCkCqrpZXKuiFBroAXhQC1wK/95XBGFMMLAAQkRCgDFg9IqVTvXhuLX8mXV6HrE0hPdeguOlqeaWCX9C1UIwxe62AMVAXAQeMMUeGq0yqb+4/9LMz4yg+gxaKe1PI3JTYXs+l2iP0TBSlglzQBZQhuAF4xteTInKbiGwWkc2VlZUjWKzxwx1Qzp2aQnVzB1VD/MO/fl8l6XERJMaE93ou1R6hpzYqFeQCElBE5E0RKfRyu2aQ1wkHrgae85XHGPOoMabAGFOQmpp6pkVXXlQ2tiMCy6emAAyplbL9WB3v7q/i5uVTvD6fZo+kuaOL5jOclqyUGj4BGUMxxlzsp0tdAWw1xpz00/XUEFQ2tZMUHc7szDjAFVDOsYLLQD2ydj8J0WF8Ydlkr8+7FzdWNbUTExGMQ39KqdHe5fUZ+ujuUiOjsrGdVHsEKbERJMeED7qFsru8njf3VnDLOVOI9REsdLW8UsEv6L7qicgq4GEgFXhFRLYbYy4TkUzgMWPMlVa+aOAS4PbAlXZsKqtr5cXtZVQ2tlPZ2E5VUzuOLsPvv7CYZGuRoSd3QAHIz7BTPMiZXr9eV4I9IpSbluf4zJOm+3kpFfSCLqAYY1bjZQqwMaYcuNLjcQuQPIJFGzd+s66EpzYdxR4RSoo9gqiwEPYcb2Db0TounpXeK39lY3v32pHp6Xae3XwMp9Ng89g+xZf9Jxt5rfAEX1sxlfioMJ/5dINIpYJf0AUUFXhHa1qYmxXPS984F3D9EV/yozcprW3pldcYQ2XTqRbKjAw7LR1dlNW1MjEput/3+vW6EqLCQrjlXO+D8W5J0eGE2GREAorTaXi3pIrzp6Ug0n9QVEq5jPYxFDUMyupayU48tdtvSmw4kWE2Smtbe+VtaHPQ4XB2B5TpGXZgYCvmD1c1888d5Xx+6WSSvEwV9mSzCSmx4VQ0Dv/U4bf3VXDT4x/y0eHaYX8vpcYSDSiqB2MM5XWtZHlsHy8iZCdGc8xLC8XdYugOKOmugDKQFfO/ffsAYSE2vnRe360Tt1T7yKyW33GsHoDD1c3D/l5KjSUaUFQP1c0dtHU6yfJooQBkJ0Z5baF0BxRrsD42IpTsxKh+WyjGGP61+wRXzcvs3p6+P6mxI7Nafnd5AwBlXuqrlPJNA4rqwR00shN7jn9MTIzmWI2XFkpTzxYKuMZR+tt1uLS2lfrWThZNThhw2UaqhbK73NVCKavTgKLUYGhAUT24v5V7dnmBq4XS0OagvrWzR7r7D7xnK2N6up0DlU10OJw+36ewzPVHe05m/IDLlmaPpKqpgy6nGfBrBqumuYPj9a5xmnINKEoNigYU1UNZnasVcnqXl3vG1undQJWN7YSH2IiLOjVhMD/DjsNpOFTlewxiV1k9oTYh3xrEH4hUewRdTkNtS8eAXzNY7tZJSmyEtlCUGiQNKKqHstpW7BGhvdaEuGd9nT4wX9HYRqo9osf02vzumV4NPt+nsLyBael2IsNCBly2kViLUljmKvMls9I4XteGcxhbQ0qNNRpQVA9lda29WifgGkMBeg3MVza2k2LvuXo+NyWWUJv4nOlljKGwrJ65WXGDKlvaCGy/sru8nqyEKGZlxtPR5dQt85UaBA0oqofS2p5rUNwSosOICQ/pNTBf2djePcPLLTzURl5qrM89vY7Xt1HT3MGcrIGPn8DItFD2lDcwOzOObGsMydvMNqWUdxpQVA9lta29BuTh1FqU0//AVnmskvc0vY89vboH5AcZUFJi3S2U4Vnc2NTu4FB1M3Oy4sm0/g10HEWpgdOAorrVt3bS2O7w2uUFMDEpqsf2K44uJ9XNHV4DyowMO8dqWnvNCgNXQLEJzMwYXJdXTEQoE5Oi2HCgelCvG6i9xxswxnXypPvfQNeiKDVwGlBUt1NThr3vweVuoRjjGqiuae7AGLwGlKW5rn071xVV9HqusLyBaWl2osIHPiDvds38LN4vqfJ5emOHw9ndAhqs3dbrZmfGE2tNTHDPelNK9U8Diurmbn14G0Nxpze1n1qLUnHaKnlPCycmkBEXyau7jvd6bldZPbMHOSDv9omFWTgN/HNHudfnH1m7n6sefo8dx+oGfe3d5Q2kxIaTHueqT1ZCFOV1euywUgOlAUV1c48X+Orycq+eP1bjyudtlbybzSZcPieD9fsqexzbW9HQRmVjO3MHOX7iNjUtlnnZ8azeVtbruZYOB09uPAK4djEerMLyBmZlxndPgc5KjNIuL6UGQQOK6lZW20pkmI1kHzv/Tkxyz3xytWROrZLvHVAArpiTQbvDybriU91eu4Y4IO9p1cIsdpc39JqW/PyWUupaOlmRn8obe072uQ7mdO2OLvafbGRO5qmWU1ZCFGV1p7r4lFJ904CiupXVtZKZEOXzDJDs09aiuANKipcuL4CCnCRSYsN5rfBEd1phWQMiMGvC0Lq8AK6al0mITXq0Urqchj+8d4iFkxL4308tIDo8hN+sOzDga+4/2YTDaZjtsRVMVoKri6+h1e0PlqEAACAASURBVNHHK5VSbhpQVLeyOu9Tht3io8KwR4Z2r5avbGzHHhHqc3A9xCZcNjuDdUUVtHV2Aa4WSm5KDDE+zo4fiFR7BOdNS+HFbWXdK9nX7DnBkeoWbjsvl8SYcL6wdDIv7yzncB/bv3hyb7ky27OFYnX9lerAvFIDogFFdXMtauz7lMWJHmtRKn2sQfF0xZwJtHR0sX5fJeD6w30m3V1uqxZmUV7fxoeHawB49J2DTEqK5tLZGQDcet4UQkNs/PbtgbVSCssasEeEMsnjlEl3cNVxFKUGJugCiohcLyK7RcQpIgV95Pu2la9QRJ4RkYEdqqG8aulwUNPc4XOGl1t2YlT3anlv266c7uzcJBKjw3ht13Gqmto5Xt825AF5T5fOyiAmPITVW8vYcqSGrUfruPXcKYRY59in2SO5YclE/r61tMfiREeXk1d2Hmf7abPAdpfXMzMzDpvtVHdf91oUXdyo1IAEXUABCoFrgXd8ZRCRLODfgAJjzBwgBLhhZIo3Nrm3au+rywtcuw6716JUNfbfQgkLsXHJrHTe2lvB1iOuI3VnD2LLel+iwkO4fM4EXt11nEfWlhAfFcb1Bdk98tx+QR4Aj64/QLuji6c2HWHlL97ma09v5ZO//YDH3zuEMYYup2Hv8cYe3V0AyTGuo491G3ulBiboAooxZq8xpngAWUOBKBEJBaIB7wsT1IC4u7F8TRl2y06MorWzi5rmDq/7eHlzxdwJNLY7+P07BwGGvAbldKsWZtHY7mBdcSWfXzqJ6PCe4zJZCVFcuyiLv350jPN/to57VheSFB3Orz+7iJUz0vjhy3v4xjPb2F1eT2tnV6+zWUSETGuml1Kqf0MfGQ0gY0yZiPwcOAq0Am8YY97wlldEbgNuA5g0adLIFXKUcf/R7L/LyzXGsL+iicZ2R78tFIBz8lKwR4ay5UgtOcnRxEWG9fuagViWl0x6XAS1zZ3ctCzHa56vrpjKyzuPMyUlhl9cv4BzpiYjIlwxJ4PfvXOAn79ezNvFrvEdb4EuK2Hk1qJUN7Xz8NoSvnPpdL/9Gyk1kgISUETkTSDDy1P3GGNeHMDrE4FrgClAHfCciHzeGPOX0/MaYx4FHgUoKCjQBQU+lNa2EmqTfs93d69F2XbUNQYxkIASHmrjkpnpvLCtzC8D8m4hNuHej8+mrqWTtDjv5c5JiWHHvZcSFtKzMW6zCV9dMZX52Ql845ltxJgQ8lJje70+KyGKvccHvp7lTPzolb3d/0afXJzd/wuUCjIBCSjGmIvP8BIXA4eMMZUAIvICsBzoFVDUwJTVtjIhIbJ7UNsXdwtl61HXeMhAAgrA5XMy/B5QAK6cO6HfPKcHE0/nTE3h9W+dT1VTu9d8WQlRVDV10NbZNajDwAbrw0M1vGCtq9l4sFoDihqVRmWXF66urqUiEo2ry+siYHNgizS69bcGxS02IpTE6DC2uQPKAMZQAFbkp3H7Bbl8YkHWGZVzOKTaI3wGRs+ZXt5aMP7g6HLyXy8WkpUQRW5qDJsODc9uykoNt6AblBeRVSJSCiwDXhGR1630TBF5FcAYswl4HtgK7MJVj0cDVOQxwXUOSt9rUNyyE6OpanKd6+5r25XThYfauPuKmWTEj67Z3SOxFuXPG49QdKKR/3fVTFbmp3GsprXHMQFKjRZBF1CMMauNMdnGmAhjTLox5jIrvdwYc6VHvnuNMTOMMXOMMV8wxuhZrUPU4XBysrGt3wF5N/c4iggk+dj3a6xwt1CGa+pwZWM7D76xj/OmpXDZ7Izubf83HawZlvdTajgFXUBRI+94fSvG9D9l2M09jpIcE05oH+MTY0F6XCQ2Gb7FjT95bS9tji5+cPVsRIQZGXbio8LYeFC7vdToM7b/GqgBcXfnZA9gDAVOTS32tSnkWBIWYiMjLnJYurw2H67hha1lfPm8XHKt8RmbTTh7ShIbdRxFjUIaUBSl/ZyDcrqJVgtloDO8RrusxKjufyN/eubDY8RHhfH1C6f2SF+am8yxmlZdUKlGHQ0oirLaVkRgQvzgWijjJqAM0+LGbUdrWZKT2GuF/6lxlIG1Utw7LisVaBpQFKW1raTbIwkPHdjHITsxGhH6XQQ5VmQlRnGioQ1Hl3PAr2npcPC91bs45GP7/NrmDg5WNbNwUmKv5wYzjrJmz0nm//AN6lo6Blw2pYaLBhRFWV3LgLu7wLUx428/t5iblk8exlIFj6yEaLqchpONA59I+NBbJTy96SjPbj7m9Xn3bscLJyX0es5mE86aksTGAcz0er+kisY2B4VlI7OaX6m+aEAZ59odXRSfaGRy8sDWoLhdPidjwF1ko11mgqslNtBur30nG3nsXddGmB8c8N7K2Hq0FpvA/OzeAQVc3V5Ha1r6na7sPhhspLaHUaovGlDGudd2naC2pTMoV7AHi+xBrEUxxvD9fxQSExHK55dOYldpHfWtnb3ybTtax4yMOJ8nVy7NTQLoc9W809p2H2DvCQ0oKvA0oIxzf9pwmNyUGM6dmhLoogStTGs69dMfHmXNnpPdxxl78/etZXx4qIa7rpjBVfMycRrXPl2eupyG7cfqWDTZe+sEYGZGnGsc5YDvbq9jtS00tTsQoTuwKBVIZxxQRKTXPtsion+dRoGdpXVsO1rHF5ZN7nFSoeopOjyUL583haLjDXz5yc0s/p81fP3prby++wSdHgP1dS0d/OTVvSyalMCnCyaycFICkWE2PjhQ1eN6+ysaaWp3sHBi7wF5t+5xlD5aKLvLXa2Sc/JSKKlo7FEWpQJhyAFFRFZae26Vi8gbIpLj8bTXs0lUcHlywxGiw0O4Tne27dc9H5vF5u9fwp9uOYurF2Sy4UA1t/95C+fcv5ZfvFFMaW0LP3u9mLrWTu77xFxsNiEiNIQlOUl8UNIzKGw94hqQXzTZd0AB1zjKkeoWjtd772rbU95AiE24ekEmnV2GA5VN/qmsUkN0Ji2UnwGXGWNScW3MuEZEllrP6dfdIFfT3ME/d5Rz7aIsPcxpgMJDbVwwPZWfXDuPTd+7iMduLGBOVjyPrCvhvJ+t4+lNR7l5eQ6zPI4SXpaXTPHJRio9ZohtO1pLUkw4Of1MhOgeR/Ex22t3eT1TU2NZMNHVdTbQgfmmdocGHzUsziSghBtjdgMYY54HPgH8SURWAbrSKsj97aNjdDic3OjjpEPVt9AQGxfPSufxm5fw3ncv5Bsrp3L57Ay+fcn0HvnOyXP1/nquKdl6tJaFExMQ6ft718yMOBKiw3h3f5XX5/ccb2B2Zhy5KTGEh9goGuA4ysNr93P1w+9pF5nyuzM5D6VTRDKMMScAjDG7ReQi4GUgzy+lU8Oiy2n4y8YjLMtNZnq6PdDFGfWyEqL4zqX5Xp+bnRmHPTKUDw5U8fH5mdS1dHCgsplrF/XfzWizCedOTeGd/ZU4nabHOFdVUzsnG9qZlRlHaIiNaemx7BlgC2VXaT3NHV0cqmrW/3/lV2fSQrkLSPdMMMaUAiuA+8/gumqYvbX3JGV1reNmYWIghYbYOHtKcvd6lG3uBY0Tfc/w8rQiP43KxvZewWKPNSDv7l6bOSGOohMDa6EUW/kGml+pgRpyQDHGvGmM2QEgIlEikm+l1xljfuSvAir/e3LDESbER3LxzPT+M6szds5U1+B6aW0L247WuRY0DjCgXDA9FYD1+yp7pLtneM2a4AooMzLsVDa2U9XU92r+ysZ2qptd27QU6WJI5Wf+mDb8cWA78C/r8QIR+eeZXlcNzfdW7+JrT2/1+fyxmhbeK6nic2dPGvNnmQSL5dY4ygcHqtl2tJb8PhY0ni7VHsGcrDjeLq7okb7neANZCVEkRLsOOHMHlv4G5t2tE5toC0X5nz/+ovw3cBZQB2CM2Q5M8cN11SB1OJy8uK2MNbtP0trhffHd+yWuAd7L52SMZNHGtenpsaTEhvN+SRXbj9axyMv+XX1ZMT2NrUd7rrjfXV7fYzbZDCug9DcwX2StqD9naoq2UJTf+SOgOIwx9ael6SyvANh8pIbmji46upxsOVLrNc+Gg9Wk2iPIsw50UsNPRFiWl8JrhSdobHd43WG4LyvyU+lyGt6zZnu1dDg4VNXc3SoB11HM6XERA2qhpMSGszwvhfL6Nq/bwig1VP4IKIUi8lkgRESmicjDwAd+uK4apPXFlYSFCKE26bU6G1z7TH1woJqlucn9TllV/rU8L5kOh2ua7mBbKAsmJhAXGdrd7VV0ohFjXDPIPM2cENfvTK/ik43kZ9iZMcE1u6tYu72UH/kjoHwDmA20A88ADcC3hnoxEbleRHaLiFNECvrI900RKbTyDvn9xpK3iyspmJzE/IkJbPBylsaBymYqG9tZnpccgNKNb+71KInRYUxJiRnUa0NDbJw3PZX1+yoxxpwakD8toMzIiONAZVN34Dpdl9Ow72Qj+elxzMywush0U0nlR2ccUIwxLcaYe4wxS4wxBdb9tjO4ZCFwLfCOrwwiMgf4Mq6xm/nAVSIy7Qzec9Qrr2ul+GQjK/JTWZ6XzM7SehrbenZnbLBaLctyNaCMtIlJUUxOjmZJTtKQWocrpqdSYU0f3lPeQHxUGFkJPY8PmDnB3ucWLEdrWmjrdDIjw056XAQJ0WG6qaTyK3/M8koVkQdE5FURWeu+DfV6xpi9xpjifrLNBDZawcwBrAdWDfU9xwL3tNIV+Wksy0umy2n46HDPLTs2HKwmMz5y0GefqDMnIvzl1rP58bVzh/T6C/Jd04ffLq5kT3k9sybE9QpM7jEVX62OYis9P8OOiJCfbu9OU8of/NHl9RRQhGtm1w+Aw8BHfrhuXwqB80UkWUSigSuBid4yishtIrJZRDZXVlZ6y9IvYwwVjWfS6Bp+bxdXMCE+kunpsSyalEh4qK3HpoROp2HjwRqW5un4SaBMTIomJTZiSK9Ns0cyOzOOtUUVFJ1o7NXdBTAlJYbwUJvPVkfRiUZE6F4dP3NCHMUnGvVMeuU3/ggoycaYPwCdxpj1xphbgKV9vUBE3rTGP06/XTOQNzTG7AV+CqzBtf5lB+DwkfdRqyuuIDU1dVAVc7v7hV1c/fD7GBOcv3gdDifvl1SzIj8VESEyLITFkxJ7jKMUn2ykprlDu7tGsRX5qWw5Uku7w9lrQB5cYy3T02N9zvQqPtHI5KRoosJDANdiyOaOLkoHeBKlUv3xR0Bxd9QfF5GPichCoM+NiowxFxtj5ni5vTjQNzXG/MEYs8gYcz5QA+wfehX6NicrnhMNbRyubhmutzgjW47U0tTu4ILpad1py/OS2XO8gVprVfQGa+uPZTogP2qtyD/1/+uthQKugXlfLZTiE64ZXt153YshtdtL+Yk/Asp9IhIP/DvwH8BjwLf9cN0+iUia9XMSrkH8Z4brvdyzorxNxQ0Gb++rINQmnDP1VLBYPjUZY04dIbvhYDWTkqLJTtTxk9FqoTV9ODzU5nMd0cwJcVQ1tffYLh+grbOLw9XN5GecCkTT02MR0anDyn/8McvrZWNMvTGm0Biz0hiz2Bgz5K1XRGSVdXDXMuAVEXndSs8UkVc9sv5dRPYALwFfM8Z4X8nnB1NSYkiPi+j+lh9s1hdXUpCTiN3jXJN52QlEh4fwwYFqupyGjQertbtrlAsNsfGxeZksy00mzMe2OTOtFsjp3V77TzbhNK5uLrfo8FAmJ0UPaOpwaW0L+05q4FF9O5Pt6wEQkSm41qLkeF7PGHP1UK5njFkNrPaSXo5r8N39+LyhXH8oRIRlucm8V1KNMSaoBrWP17dSdKKRu66Y0SM9LMTGkpwkNhyoZk95A41tDpZP1YAy2v141Zw+P3+zs+KJCLXx0o5yzp9+asywyGOGl6cZGXEDOkflzud2UlLZxMa7LyJEj4tWPvijy+sfuGZ2PQz8wuM2pizLS6aqqZ2SiuA66W59sXu6cO8JB8vzktlf0cSL28sAXX8yFvT3ZSY+KozPnj2JF7aVcdRjzK/4RCMRoTZyknsuqpwxwc6h6mafe78B1Ld08uHhGiob24O221cFB38ElDZjzEPGmHXWLK/1xpj1frhuUFmWe2rH2GDydnElGXGR5Hs5KMm9y+2fNx4hLzWGtLjIkS6eCoA7Lsgj1CY8su7UPJXik41MS4/t1bqYkWHHGPrszlq/v5Iup8Em8OL28mErtxr9/BFQfiUi94rIMhFZ5L754bpBZWJSFFkJUUE1jlLV1M57JVXd04VPNyszjrjIUNodTp3dNY6kxUXymbMm8cLWU62UohONXk9nnGEN0vc1ML9270mSYsJZtTCbfxWeoK3Td2tGjW/+CChzcW2Dcj+nurt+7ofrBhXXjrHJbDxUHRQLwYwx3PncDjq6nNxyrvfTAkJswtlWN5e7haXGhztW5GGzCb9eV0JNcweVje09BuTdJiVFExUW4nPqsKPLybriSlbmp3Htoiya2h2sLarwmlcpfwSUVUCuMeYCa5bXSmPMhX64btBZlptMXUtnUBxM9OSGI6wrruR7V8zo81zwS2amEx0eoi2UcSY9LpLPnjWJv28t5c29JwF6TBl2s9mE/Ay7z4F59zksF81MY2luMmn2CP6xrWxYy65GL38ElB3A4PbjHqWWBcl6lOITjfzo1b2szE/lpuU5fea9viCbDXdfRFJM+MgUTgWNr1zgaqXc9/IeAK8tFHd60YkGrztBvFV0klCbcN60FEJswsfnZ/J2cSX1LXqOiurNHwElHSgSkddF5J/umx+uG3QyE1w7xm70sjX8SGnr7OLfntlGXGQoD1w/v99ZPyJCfFRYn3nU2JQRH8lnlkykoc1BQnQYaXbv+4jNyLBT29LZazEkwNq9FZydm9S9xumaBZl0dDl5rfD4sJZdjU7+CCj34ur2+jFjeNqw27LcZDYdqqErQOMo979WRPHJRh64fv6QNxpU48cdK6YSHmIjP93u88vHvImuDobntpT2SD9a3cL+iiYunJHenTY3K54pKTH8Y7t2e6ne/LFSfr23mz8KF4yW5SXT2OZgd/nppx4Pv7VFJ3nig8PcvDyHlR77OinlS0Z8JA99ZgF3XpbvM8/CiQlcOTeDX725v8dsr7eKXGMvF8889VkTEa5ZkMmmQzUcr9dNJVVPQw4oItIoIg1ebo0iMmZ3m3MvDhzp6cPH61v592d3MCPD3mtVvFJ9uXzOBApyknw+LyL8zzVzsEeG8u/Pbaezy3Xi49qiCvJSY5h82mLIaxZkYQy8tEPXpKiehhxQjDF2Y0ycl5vdGON9K9QxIC0ukrzUmBFd4OjocvLNZ7bT7nDy688tIjIsZMTeW40PybER3PeJORSWNfCbdQdoanew8WA1F81M75V3SkoM87PjdZGj6sUfYyjjzrK8ZD46XEO7Y2QWeD301n4+PFzDj1bN8bnLrFJn6oq5E7h6fiYPr93Po+sP0NlluHCG967Vj8/PZHd5A0eqm0e4lCqYaUAZghXT02jp6GLJfW/y9ae38sLWUmqsc0f87f2SKh5eV8L1i7NZtbDPY2aUOmM/uHo2iTHhPLS2hLjIUAomJ3rNd4G18WQgZzyq4KMBZQgunpXOYzcWcPmcDDYdquE7z+5g8X1reGrTEb++T2VjO9/863byUmP5wTWz/XptpbxJjAnnx6tc595fkJ9GqI9t8qemxZIcE87GgzUjWTwV5M54+/rx6uJZ6Vw8Kx2n07C7vIEfv7qXH72yl5X5aWQmRPnlPX748h4a2zr5y5fOIjpc/6vUyLhkVjq/umEB87N9r1cWEZbmJrPxYPAd6aACR1soZ8hmE+Zmx/OzT87DaQw/fGmPX66793gDL+0o59Zzp3Rv4KfUSLlmQRY5KTF95lmam8Tx+jaO1gTn0dhq5GlA8ZOJSdF848Jp/Gv3Cdb5YfO8B9fswx4Zyu3n5/mhdEr531JrCr2Ooyg3DSh+9OXzcslLjeHef+7uscW3MYbnNh/jP5/fQYfD2e91dhyrY82ek3z5vFzio3XbFBWcpqbFkhKr4yjqFA0ofhQeauN/PjGHozUt/HpdCQDHalq48fEPufP5nTy7uXRAi8F+sWYfidFhfPGcnGEusVJDJ+I6HsE9jqJU0AUUEXlARIpEZKeIrBYRryODInK5iBSLSImI3DXS5fRleV4KqxZm8bv1B3jwjWIu++U7bD1Syw+vmc309Fj+792Dff7yfXiohnf2VXLHirzuDfmUClZLc5N1HEV1C7qAAqwB5hhj5gH7gLtPzyAiIcCvgSuAWcBnRGTWiJayD9+7ciZRYSE8tLaEJTlJvP7t87lxWQ5fOjeXohONvF/ivc/ZGMPP3ygm1R7BF5bmjGyhlRqCZbmuLV10HEVBEAYUY8wbxhiH9XAj4G0131lAiTHmoDGmA/grcM1IlbE/qfYIHrtpCb/7/CKe+OISshOjAbhmYSYpsRE8+u5Br697v6SaDw/V8PWVU4kK1+1VVPDLS9VxFHVK0AWU09wCvOYlPQs45vG41ErrRURuE5HNIrK5srJyGIro3VlTkrh8zoQe8/MjQkO4eflk3tlX2esM75YOBz96dS9ZCVHccNbEESunUmdCx1GUp4AEFBF5U0QKvdyu8chzD+AAnvJ2CS9pXj/NxphHjTEFxpiC1NRU/1TgDHzu7MlEhtl4zKOV0uFw8pW/bKX4RAM/vGY2EaHaOlGjh46jKLeALL82xlzc1/MichNwFXCR8f61pxTw/BqfDYyKrU8TY8K5fvFE/vbRMe68PJ+UmAj+47kdvLOvkp9eN9fr7q5KBTPPcZTTt7pX40vQdXmJyOXAd4GrjTG+vvJ8BEwTkSkiEg7cAIyaY4dvPXcKnU4nf/rgMD98eQ//3FHOdy+fwaeXTAp00ZQaNB1HUW7BuEHUI0AEsMYaf9hojPmKiGQCjxljrjTGOETk68DrQAjwuDFmd+CKPDg5KTFcMjOd368/iMNp+NK5U/jKBbmBLpZSQ3L6OIp73LDLaTjZ0Mbx+laO17dxor6N5Nhw3TV7DAu6gGKMmeojvRy40uPxq8CrI1Uuf7v9glzW7D3JtYuy+N6VM3VzPTWqLc1N5pWdx3lq01GO1bSw7VgdhWX1tHT0PjOoYHISE5OifV7LGEPRiUZe332C13efJCY8hPtWzdE97UYBGU8zMwoKCszmzZsDXYxuR6tbyEqMIsSmwUSNbgcqm7joF+sBCA+xMTMzjoUTE5iebmdCQiQT4iMxBq741bv8x6XT+fqF07xe57F3D/LnjUc4Ut2CCCyelMjh6mYaWh1859LpfPm8XP19CQAR2WKMKegvX9C1UMaTScm+v6UpNZrkpcby51vPIjYilFmZcT5nKp49JYkXtpbxtZVTe7XK9x5v4L5X9rJ4ciK3n5/HJbPSSbVHUN3Uzj2rC7n/tSLe3HOSX3xqvg7+B6mgG5RXSo1O501LZeGkxD6nvV+7KIuDVc3sKK3v9dwT7x8mMszGH24q4LNnTyLVHgG4zrv/7ecX8b+fnk/xyUauevg96ls6h60eaug0oCilRswVcycQEWrjha2lPdKrm9pZvb2M6xZlkxAd3ut1IsKqhdk89JmFNLY5KCzvHZBU4GlAUUqNmLjIMC6Zlc5LO8p7HOXw9KajdDic/e6wPTvTNTC/72Rjn/lUYGhAUUqNqGsXZVHb0snbxa6D6DocTp7ceIQLpqcyNc3e52tTYyNIiA5j38mmkSiqGiQNKEqpEXXetFRSYsNZva0MgFd2lVPZ2D6g839EhOlpdvZrCyUoaUBRSo2osBAbH5+fyVt7K6hv6eSP7x8mLzWG86cNbK+9aemxFJ9s1M0og5AGFKXUiLt2YTYdXU5+8NJudpbW88VzpmAb4PqS/Aw7jW0OTja0D3Mp1WBpQFFKjbg5WXFMS4vlhW1lxEWGcu0ir6dPeDXNGmfRgfngowFFKTXiRIRVVhD5zNmTiA4f+Brr6emxgAaUYKQr5ZVSAXHDkkmUVDRx67lTBvW65NgIkmPCNaAEIQ0oSqmASIoJ58FPLRjSa6en23XqcBDSLi+l1KgzPT2WkoomnekVZDSgKKVGnWnpdpraHZTXtwW6KMqDBhSl1KgzPd2a6XVCx1GCiQYUpdSoozO9gpMGFKXUqJMQHU6aPUIH5oOMBhSl1Kg0Pd3O/gptoQQTDShKqVFpWnos+0824XTqTK9gEXQBRUQeEJEiEdkpIqtFJMFHvsdFpEJECke6jEqpwMtPt9Pa2UVpbWugi6IsQRdQgDXAHGPMPGAfcLePfE8Al49UoZRSwWVauu7pFWyCLqAYY94wxjishxuBbB/53gFqRqxgSqmgMs0900vHUYJG0AWU09wCvHYmFxCR20Rks4hsrqys9FOxlFKBFhcZxoT4SF2LEkQCspeXiLwJZHh56h5jzItWnnsAB/DUmbyXMeZR4FGAgoICHb1TagzRPb2CS0ACijHm4r6eF5GbgKuAi4xu1qOU8mF6eiwbD1bT5TSEDPCALjV8gq7LS0QuB74LXG2MaQl0eZRSwWtaup12h5OjNfqnIhgE4/b1jwARwBoRAdhojPmKiGQCjxljrgQQkWeAFUCKiJQC9xpj/hCgMiulAiDfmul138t7yEuLxR4Rij0ylEnJ0SyYmEhSTHiP/K0dXWw7VsuxmhY+Ni+T2Ihg/BM4egXdv6YxZqqP9HLgSo/HnxmxQimlglJ+hp2CyYnsKK3nvZIq2h3OHs9PSYlh4aQEkmPC2XKkll1l9XR2uXrRn9xwhD9+cQlp9shAFH1MkvE0RFFQUGA2b94c6GIopYZJh8NJQ1snJRVNbD1ay7ajdWw7Wkt9ayfzshNYkpPE2VOSaOvs4jvP7iA5Npw/3XIWeamxgS56UBORLcaYgv7yBV0LRSmlhio81EZKbAQpsREszU0GwBiDw2kIC+k5ZJyZEMUtT3zEJ3/7AY/dtITFkxMDUeQxJegG5ZVSyp9EpFcwAZg/MYEXvrqc+KgwKkGnMgAAIABJREFUPvt/G1mz52QASje2aEBRSo1bk5Nj+Psdy5kxIY7b/7yZpzcdDXSRRjUNKEqpcS05NoJnvnw2F0xP5Xurd/Hgmn16Vv0QaUBRSo170eGhPHpjAdcvzuaht/Zz9wu7cHQ5+3+h6kEH5ZVSCggLsfGzT84jPS6SR9aVcLSmhR9cPbt7V2PVP22hKKWURUT4j8vy+el1c9lVVs/lv3qXe1bvoqqpPdBFGxU0oCil1Gk+vWQS6+9cyefPnsTfPjrGigfe5pG1+6lr6fD5mnZH17jvJtOFjUop1YcDlU3c/1oRa/acJDzUxsfmTuCzZ0+iYHIizR1drC2q4F+Fx1lXVMn0DDt/vvUs4iLDAl1svxrowkYNKEopNQBFJxp4etNRVm8to7HdwaSkaE40tNHhcJISG8F501J4aUc587LjefLWs8fUPmEaULzQgKKUOlMtHQ5e2lHOK7tOkJcawxVzJrB4ciIhNuFfhcf52tPbWDwpkSduWUJ0+NgIKhpQvNCAopQabi/tKOebf93G2VOSefzmJYTYhKITDWw/VsfR6hY+vWTiqJs5pnt5KaVUAHx8fiYOp5PvPLuDlT9/m5rmDjqswXqbwJ82HOZrK6dyx4o8IkJDAltYP9OAopRSfrZqYTY2Ef6+tYz89FgWTExk/sR4IsNC+OFLe/jlm/t5eedx7r92LgU5SYEurt9ol5dSSo2wdUUVfP8fhZTVtfLJxdl8+5LpZCVEBbpYPukYihcaUJRSwaK53cGv3trPEx8cBgOfWzqJr62cSkpsBJ1dTgrL6vnocA0n6tu5eXkOk5KjA1ZWDSheaEBRSgWb8rpWfvXmfp7bcoyosBDmZMWzs7Se1s4uAMJCBJsIX10xldsvyCUybOTHXTSgeKEBRSkVrEoqmvjlm/s4Ut3C4smJnDUliSU5SXQ5Dfe9soeXdx5ncnI0/+9js0iMCWNPeQN7jjewp7yB2pbOHtcKsQnR4SHERoQSGxFKTEQot1+Qy+zM+CGVbdQGFBF5APg40AEcAL5ojKk7Lc9E4EkgA3ACjxpjftXftTWgKKVGq/dLqvivFws5UNncnZYQHcbszDhSYyMQke50h9PQ3O6gqd1Bs3V74Pr5LBniBIDRHFAuBdYaYxwi8lMAY8x3T8szAZhgjNkqInZgC/AJY8yevq6tAUUpNZp1OJy8uus4sRGhzMqMY0J8ZI9AMlxG7ToUY8wbHg83Ap/0kuc4cNy63ygie4EsoM+AopRSo1l4qI1PLMwKdDF8Cvbdhm8BXusrg4jkAAuBTT6ev01ENovI5srKSr8XUCmllEtAWigi8iau8Y/T3WOMedHKcw/gAJ7q4zqxwN+BbxljGrzlMcY8CjwKri6vMyy6UkopHwISUIwxF/f1vIjcBFwFXGR8DPKISBiuYPKUMeYF/5dSKaXUYATdGIqIXA58F7jAGNPiI48AfwD2GmMeHMnyKaWU8i4Yx1AeAezAGhHZLiK/AxCRTBF51cpzDvAF4EIrz3YRuTJA5VVKKUUQtlCMMVN9pJcDV1r33wOGf66cUkqpAQu6dSjDSUQqgSMDzJ4CVA1jcYKV1nt80XqPL0Ot92RjTGp/mcZVQBkMEdk8kIU8Y43We3zReo8vw13vYBxDUUopNQppQFFKKeUXGlB8ezTQBQgQrff4ovUeX4a13jqGopRSyi+0haKUUsovNKAopZTyCw0opxGRy0WkWERKROSuQJdnKETkcRGpEJFCj7QkEVkjIvutn4lWuojIQ1Z9d4rIIo/X3GTl32/tr+ZOXywiu6zXPCQjcSDDAIjIRBFZJyJ7RWS3iHzTSh/TdReRSBH5UER2WPX+gZU+RUQ2WXX4m4iEW+kR1uMS6/kcj2vdbaUXi8hlHulB+3shIiEisk1EXrYej/l6i8hh63O4XUQ2W2mB/5wbY/Rm3YAQXKdE5gLhwA5gVqDLNYR6nA8sAgo90n4G3GXdvwv4qXX/SlxHBAiwFNhkpScBB62fidb9ROu5D4Fl1mteA64IdJ2tck0AFln37cA+YNZYr7tVlljrfhiuoxyWAs8CN1jpvwPusO5/Ffiddf8G4G/W/VnWZz4CmGL9LoQE++8F8B3gaeBl6/GYrzdwGEj5/+zdd3zdVf348dc7e69mtE333hQayrKMUtkyBAQHoAgIiKL+vq4vCiqiaPXrABEQmbIEBZRZdqEUSrrT3aYrTdskzd7r/fvj87npTXJvmrQ3ubc37+fjcR+599xzP/ec9ibve3aXtKB/zq2F0tlcYKuqFqpqM/AMcFGQy9RnqroYKO+SfBHwmHv/MeBir/TH1fExkCbOiZhnA2+qarmqVgBvAue4z6Wo6lJ1PnmPe10rqFR1r6qucO/XAJ6D18K67m75a92H0e5NgfnA825613p7/j2eB850v4FeBDyjqk2quh3YivM7EbK/FyIyAjgfeMh9LAyCevsR9M+5BZTOcoHdXo+L3LRwkKPOSZe4P7PddH917im9yEd6SJHOB6+Ffd3dbp9VQAnOH4ZtQKWqtrpZvMvaUT/3+SpgCH3/9wgFfwR+ALS7j4cwOOqtwCIRWS4iN7hpQf+ch9zmkEHmq58w3OdV+6tzX9NDhnQ5eK2H7t+wqbuqtgGzRSQNeAGY6iub+7Ov9fP1xTPo9RaRC4ASVV0uIqd7kn1kDat6u05R1WIRycbZmX1jD3kH7HNuLZTOioCRXo9HAMVBKkug7Xebsrg/S9x0f3XuKX2Ej/SQIL4PXhsUdQdQ1UrgPZy+8jQR8Xxp9C5rR/3c51Nxukj7+u8RbKcAF4rIDpzuqPk4LZZwrzfq7L6OqpbgfIGYSyh8zoM9uBRKN5wWWyHOwJxnEG56sMt1mHUZQ+dB+YV0HrD7rXv/fDoP2C1z0zOA7TiDdenu/Qz3uU/dvJ4Bu/OCXV+3XILT3/vHLulhXXcgC0hz78cDH+CcePocnQenb3bvf5POg9P/dO9Pp/PgdCHOwHTI/14Ap3NwUD6s6w0kAsle9z8CzgmFz3nQPwihdsOZEbEZpw/6tmCX5zDr8DSwF2jB+bbxdZy+4reBLe5PzwdHgL+49V0L5Hld51qcAcqtwNe80vOAAvc19+LuuBDsG/AZnKb5GmCVezsv3OsOzAJWuvUuAG5308fhzNbZ6v6RjXXT49zHW93nx3ld6za3bpvwmtkT6r8XdA4oYV1vt36r3ds6T7lC4XNuW68YY4wJCBtDMcYYExAWUIwxxgSEBRRjjDEBYQHFGGNMQFhAMcYYExAWUExYEZE2dwfW1SKyQkROPkT+NBG5uRfXfU9E8gJX0qOfiDwqIpcFuxwmdFhAMeGmQVVnq+oxwI+BXx8ifxrOLrQhyWvFtzEhzwKKCWcpQAU4+3uJyNtuq2WtiHh2jb0bGO+2aha6eX/g5lktInd7Xe9ycc4d2Swi89y8kSKyUEQ+dc+a+IabPkxEFrvXLfDk9+aeafEb95rLRGSCm/6oiPyfiLwL/MY95+JF9/ofi8gsrzo94pZ1jYhc6qafJSJL3bo+5+5thojcLSLr3by/c9Mud8u3WkQWH6JOIiL3utd4hYObDxoD2OaQJvzEu7vuxuGcjzLfTW8ELlFns8hM4GMR+Q/OFhUzVHU2gIici7NV9wmqWi8iGV7XjlLVuSJyHnAHsABnF4IqVT1eRGKBJSKyCPg88Iaq3iUikUCCn/JWu9e8Gmcfqgvc9EnAAlVtE5F7gJWqerGIzMfZXmY28FP3vWe6ZU936/YT97V1IvJD4Hsici9wCTBFVdXdRBLgduBsVd3jleavTscCk4GZQA6wHni4V/8rZlCwgGLCTYNXcDgJeFxEZuBsP/ErETkVZ6vzXJw/il0tAB5R1XoAVfU+V8az2eRynL3SAM4CZnmNJaQCE3H2QnrY3azyRVVd5ae8T3v9/INX+nPq7CAMzpYyl7rleUdEhohIqlvWKz0vUNUKdwfeaThBAJw9qJYC1ThB9SG3dfGy+7IlwKMi8k+v+vmr06nA0265ikXkHT91MoOUBRQTtlR1qfuNPQtnT6YsYI6qtrg71Mb5eJngf6vuJvdnGwd/dwT4lqq+0e1CTvA6H3hCRBaq6uO+iunnfl2XMvl6na+yCs6hSV/0UZ65wJk4QegWYL6q3igiJ7jlXCUis/3VyW2Z2V5Nxi8bQzFhS0Sm4OwaewDnW3aJG0zOAEa72Wpwjgv2WARcKyIJ7jW8u7x8eQO4yW2JICKTRCRRREa77/c34O84RzL7coXXz6V+8iwGvuxe/3SgTFWr3bLe4lXfdOBj4BSv8ZgEt0xJQKqqvgp8B6fLDBEZr6qfqOrtQBnOduY+6+SW40p3jGUYcMYh/m3MIGMtFBNuPGMo4HzTvsYdh3gS+K+I5OPsQrwRQFUPiMgSESkAXlPV77vf0vNFpBl4FfjfHt7vIZzurxXi9DGV4ozBnA58X0RagFrgaj+vjxWRT3C+3HVrVbh+BjwiImuAeuAaN/2XwF/csrcBP1fVf4vIV4Gn3fEPcMZUaoCXRCTO/Xf5rvvcQhGZ6Ka9jbOD7Ro/dXoBZ0xqLc4OvO/38O9iBiHbbdiYIHG73fJUtSzYZTEmEKzLyxhjTEBYC8UYY0xAWAvFGGNMQFhAMcYYExAWUIwxxgSEBRRjjDEBMajWoWRmZuqYMWOCXQxjjDmqLF++vExVsw6Vb1AFlDFjxpCfnx/sYhhjzFFFRHb2Jp91eRljjAkICyjGGGMCwgKKMcaYgLCAYowxJiAsoBhjjAkICyjmsGzYW01tU2uwi2GMCSEWUEyftba1c8l9S/j7B9uDXRRjTAixgGL6rKK+hcaWdooq6oNdFGNMCAm5gCIid4rIGhFZJSKLRGS4n3zXiMgW93aNrzymf5TXNQOwv6bpEDmNMYNJyAUUYKGqzlLV2cDLwO1dM7jnfN8BnADMBe5wz9M2A8ATUEqqG4NcEmNMKAm5gKKq1V4PEwFfJ4CdDbypquWqWgG8CZwzEOUzBwNKqbVQjDFeQnIvLxG5C7gaqALO8JElF9jt9bjITfN1rRuAGwBGjRoV2IIOUuV1TiA5UNdMS1s70ZEh973EGBMEQflLICJviUiBj9tFAKp6m6qOBJ4EbvF1CR9pPs8yVtUHVTVPVfOysg65WabphQNuCwWslWKMOSgoLRRVXdDLrE8Br+CMl3grAk73ejwCeO+IC2Z6pcIroJTUNDE8LT6IpTHGhIqQ66sQkYleDy8ENvrI9gZwloiku4PxZ7lpZgB4t1BsYN4Y4xFyAQW42+3+WoMTKG4FEJE8EXkIQFXLgTuBT93bL9w0MwDK65oZke60Skqsy8sY4wq5QXlVvdRPej5wndfjh4GHB6pc5qDyumYm5SSzp7LBWijGmA6h2EIxIa68rpmspFiGJMZaC8UY08ECiukTVaWivpmMpBiyky2gGGMOsoBi+qS6sZWWNmVIYgzZKbGU1FiXlzHGYQHF9IlnlXxGottCqbYWijHGYQHF9IknoKQnxpCTEkdZbRNt7T7XlHZ73RUPLKVgT1V/F9EYEyQWUEyfeALKELeF0q5woPbQrZQHFxfyyfZyFq3f399FNMYEiQUU0yeefbwyEmPISo4DDr0WpbyumceX7gBgnbVQjAlbFlBMnxzoaKHEkp0SC3DIgfmHPiikoaWN2SPTWGsBxZiwZQHF9ElFXTNx0RHEx0SSk+K0UPb3MDBfUdfMYx/t4PyZw7jwmOGU1DTZYkhjwpQFFNMnB+qaGZLotEyyktwWSg8B5aEPC6lvaePbZ05kRm4qAAXF1koxJhxZQDF9Ul7XTEZiDAAxURGkJ0T77fKqrG/msY92ct6MYUzKSWb68BREYG1Rtc/8h9Lerlz54FJeXlN82OU3xvQfCyimT7wDCkB2cpzfQfm/f7id2qZWvnXmBAASY6MYl5l42OMouyvq+biwnHc2lhzW640x/csCiumTbgElJdbnmEhVfQuPLtnBeTOHMmVoSkf6jNzUw16LUrDHadkUltYd1uuNMf3LAorpk962UF5YWURNUyvfmj+xU/rM3FT2VTce1kmPnrGXwtJaVA+9mNIYM7AsoJhea2xpo765rVsLpbSmifYuq+U/3VlBblo8U4eldEo/koF5T8umurG1Y4GlMSZ0WEAxvXbAax8vj+zkWFrbnR2Iva3cWcGxo9K6XWPacCfAFBT1LaCoKuuKq8lx175sL/Pd7fXpjnIuv/8j6ptb+3R9Y8yRs4Bieq3CR0DxrEXx7vbaV9VIcVUjx45K73aNlLhoxmYm9rmFsreqkfK6Zi6YNRzwP46yaN0+Pt1RwYdbyvp0fWPMkbOAYnrtgNc+Xh7ZyU6LYb/XwPzKXRUAHOejhQKegfm+TR32dHedPX0o0ZFCoZ8Wyoa9NQC8u8lmghkz0CygmF7z3sfLI9vHfl4rdlUQExnR0b3V1YzhKeypbOjTOMi64moiBGbkpjB6SCKFpbXd8qgqG/Y6gerdjaU2cG/MALOAYnrtQK2PMRR3TMN71tbKXZXMyE0hNirS53Vmegbm+zB9eF1xFeOykkiIcday+GqhlNY0caCumanDUthX3cjGfTW9vr4x5shZQDG9VlHfTGSEkBIX3ZEWFx1JSlxUx1qU5tZ21uyp4jgf4yce092A0pcFjgV7qpnhtnjGZiWy80Bdt3NYNrgB5KbTxwPYAkhjBpgFFNNr5XXNpCfEEBEhndKzU+I6NojcsLea5tZ2nwPyHqnx0YzKSOh1C6W0pol91Y0dU47HZybR0qYUVdR3yufp7jptYhYzclN4z8ZRjBlQFlBMrx2obe40IO+RnXzwbPkVngH50b4H5D1m5qb2eqbXOjff9OFOQBmblQjQrdtrw95qhqfGkZoQzfzJ2SzfWUFl/aHHaf6Zv5uvPrKsV2UxxvhnAcX0WnldM+mJ0d3SnYDitFBW7KpkWGocw1Lje7zWjNxUdpc39OoP/rpip+XhGeQfl+kGlNLuAcWzkPL0Kdm0KyzuxfThf3y8k/c2lfbq5EmPN9fv55/5u3ud35jBIOQCiojcKSJrRGSViCwSkeF+8r0uIpUi8vJAl3GwKq8/uHW9t5wUZ/sVVWXlLt8LGruakesucOzF9OGCPVWMHpJAarwTzDISY0iJi2J72cGZXo0tbWwrresIKMeMSCMjMYZ3DzGOsr+6kTXuIstN+3s/iP+Xd7fyuzc29Tq/MYNByAUUYKGqzlLV2cDLwO3+8gFXDVyxTNd9vDyykmNpbm1na0ktRRUNPQ7Ie8wY3vuB+XXF1R35AUSEcVlJnVooW0tqaWvXjoASGSGcNimL9zeXdhu89+Y9cL+pl7PC2tuVzftrKKlp6lOrxphwF3IBRVW9v7ImAj7/Gqjq24DNCx0grW3tVNa3+Awo2e5q+TfW7QPoVQslPTGGcZmJPLB4G48s2U5za7vPfFX1Lewqr2d6buc1LeMyEzsFlPXugPzUYckdaWdMyaa8rpnVRZV+y/H2hhJy0+JJS4hmcy9bKHsqG6hvbgOwqcnGeAm5gAIgIneJyG7gy/hvofT2WjeISL6I5JeWlgamgINQRX0LgO+A4q6Wf33dPqIjpWPw/FDuv2oO04en8PP/ruesP7zPa2v3dluMuG5v5wF5j3FZieyrbqSuydmza8PeauKjIxk9JLEjz6kTM4kQeM9Pt1djSxsfbi1lwdRsJuck9zo4eOfzzCwzxgQpoIjIWyJS4ON2EYCq3qaqI4EngVuO5L1U9UFVzVPVvKysrEAUf1DybP7oK6B49vMq2FPN9OGpxEX7XtDY1aScZP7x9RN45GvHEx0ZwU1PruBLf/ukY88wgHXuGMv0Lqvux2YmAQc3idy4t4bJQ5OJ9JrSnJYQw5zR6bzjZ/rwR9vKaGxp58ypOUwZmszmfTXddk32ZdM+p0yp8dEdW70MlNa2dr+tOWOCLSgBRVUXqOoMH7eXumR9Crg0GGU0nXlWyfubNuzRm+4ubyLCGZOzee3Wedx58QyW76zg8geWUlzZADjb3A9LjSMzqfNkgHHu1OHtZXXOliv7qrttlQ9w+uRsCvZU+zwE7K0NJSTGRHLCuAwmDU2mrrmNPe779mTT/lpGpMdzzMi0AW+h/PSldXz9sU8H9D2N6a2Q6/ISEe8TmS4ENgarLOYgz75bGUndA0pibBSJMU6rpDcD8r5ERUZw1YmjeezaueyvauTSv37Elv01FOyp8tmFNtZr6vC+6kYq61s6jZ94nDk1G4BnP+08xVdVeWdDCfMmZhEbFcmUoc5rezMwv2lfNZNzkpk6LJmtJbW0tA1ci2FNUeWAt4qM6a2QCyjA3W731xrgLOBWABHJE5GHPJlE5APgOeBMESkSkbODU9zBoWNjyITuAQUODsz3tYXS1Unjh/DsN06itV257P6lFJbVdUwx9hYXHUluWjyFZbUdrQRfLZQpQ1M4b+ZQ7nl3K1tLDv4hXldczb7qxo6AMynHDSiHGJhvbm2nsLSOyUOTmTo0hea29gE9knh3eT0H6ppoHcAgZkxvhVxAUdVL3e6vWar6OVXd46bnq+p1XvnmqWqWqsar6ghVfSN4pQ5/5XXOoHy6jy4vcLq9spNjyU3reUFjb0wbnsK/bjyZ9IRoVLsPyHuMy0pke1ldxzd2Tyujq59fOIOEmEh+8PyajinEb28oQcSZCQaQHBdNblr8IVso20praW1XJ6C4AWzjvoHp9qpqaKG6sRXVg0cJGBNKQi6gmNBUXtdESlwU0ZG+PzI3nzGBOz43HRHx+XxfjRqSwPM3nczPPjeN0yf7nkwx1p06vH5vNSMz4kmO676KH5x1Mj/73HRW7Krk0Y92APDOxv3MHpnWaWxm8tDkQwYUz9TiKUNTGJeVSExkRMeU5f7mvXdZSbWtfzGhxwKK6ZUDfhY1epw2KYvzZw0L6HtmJsXy1VPG+g1i4zITqW1q5aOtZUwd6vvsFY+LZg/nzCnZLHxjI/k7ylldVMWCqTmd8kwemsy20toeZ1Ft3FdDVIQwNjOR6MgIJmQnsXGAxjSKKg5OGNjvY5KBMcFmAcX0SkV9zwElGMZmOVOHK+pbfI6feBMR7rpkJtEREXztEWeWlGf8xGNyTjKt7er3vHpwBu3HZyURE+X86kwZluxzpld7u/LMsl1Uuet3AmF3uVcLpcZaKCb0WEAxvXKgtpkMH/t4BZNnk0jwPSDf1dDUOH5ywVRqmlrJTYtnck7nMZfJ7hhMT2Mim/bVdOQDmDYsxecWLB9uLeNH/17Lv1YU9aouvVFU0UCCO5vOs7uzMaHEAorplfI631vXB9PwtPiOlsK0XgQUgC/kjeQLeSO4bt7YbuM947OSiIoQv+MoNY0t7Kls6BRQpgz1DMx3fs3zy51AstXHUcWHq6iinlEZCWQkxlgLxYSkqGAXwIQ+VaWivtnvDK9giYwQxg5JpKiinhHpvZtdJiL89rJjfD4XExXBuKxEv3t6edK9WzaetS8b9lZzyoRMwJmN5dnXbFtJ4ALK7vIGRg1JAGxQ3oQmCyjmkGqaWmlp05BroQCcPiWL0uqmbqdIHq5JOcms2u17M8lN+5zg4N1CGZIUS1ZybKfFhq+s2UtTaztTh6WwLUAtFFVld0U9p0zIpLGljVLr8jIhyLq8zCGV1/rfxyvYfnzuVP7vitkBu96UockUVTRQ62466W3TvmoSYyK7rbWZOiyl08D888t3MykniYtnD6estrlXh4gdSkV9C/XNbYxIj+84f8aYUGMBxRzSbnf9g69tV8LNZHdMxFe318Z9NUwamtytNeS9Bcu20lpW7KrksjkjmJDtzEILRCvFM8NrZEYC2cmxlNY09WojS2MGkgUU41fBniq+/fRKvvrIp8RFR3RsTxLOPOMjXQfmVZVN+2t8rsb33oLlX8uLiIwQLp6dezCglBz51iyeoD4yI57s5Fha25XyALR8jAkkG0Mx3WzZX8PtL61jaeEBkmKjuPaUMXz1lLEB2VYl1I1IjychJrJbQCmtaaKyvsVnUPVMWV5XXMW/V+zhtElZZKfE0dauxERFBGSml2dR44j0BLJTnABVUt3UbRdmY4LJAorp5o7/rGP93mpuO28qV8wdSYqfLU3CUUSEMCmn+xYsnmnBk320UDxbsDy8ZDv7qhu5/XPTAGcW2rjMxIDM9NpdXk96QjRJsVEdxwWU1DQyjd5NlzZmIFiXl+lk54E6Ptp2gOs+M5brTx03qIKJx+ScZDbtr+l0eqQnwHRdDAl0bMFSsKea1PjoTivwx2clBaSFsruigZEZzpTh7GRnZ2cbmDehxgKK6eSZT3cTIXB53shgFyVoJg9Npryuma1eLYtN+2vITIpliJ8upinuepSLZg8nNurgiZXjs5PYXV5PY0vbEZXJe61NdopThlILKCbEWEAxHVra2nkuv4j5U3IYmhoX7OIEzWz3TJfP/mEx5/7pA377+kZW7Kzwuz0+wAx3i/3L5ozolD4+K5F2hR0HDn9gvr1dKapoYGS600KJi44kJS7K5ymUxgSTjaGYDm9v2E9ZbRNfnDt4WyfgnDr51vdO5e0NJbyzsYQHFhfS1q4smJbj9zVXzh3JpJxkZo3ofMCY90yvKYfYEdmf0tommlvbGeF2eYFzoJl1eZlQYwElzPzl3a28u7GEs6bncO6MYR397r3x9LLdDE2J47RJvs8fGUwmZCczITuZb5w2nqqGFpbvLOfYkf6PN06IieIzEzO7pY/LTEKETt1nfeU5B8V7e5ns5Fjbwt6EHAsoYea/q4vZXlZH/s4KfvXqRmbmpnLl3JF8+YTRPb6uqKKexVtK+db8iUT5OX9ksEqNj2b+FP+tk57Euyvrj2Rx4+5yZ8qwp8sLnICSv7PisK9pTH+wvxxhpN09y+OqE0fzwQ/O4H/Pm0Jbu3LbCwWsK67q8bX//HQ3AF/IG9FjPtN347OSjqiF4lkl36mF4nZ5ec9EMybYLKCEkb3VjTRbNR2fAAAgAElEQVS1tjM2K5GRGQnccOp4nvj6XCIE3ijY5/d1rW3t/DO/iNMmZTEivfddZKZ3JmQnUVhWe9hbpRRVNJCVHEtc9MHZY9nJsTS3tlPd0H3PMWOCxQJKGNle6swkGut18NSQpFjmjs3gtR4CyvubS9lX3ciVx4/q9zIORuOzkmhsaae4qqFT+vKdFfx+0SbaDhFodlfUM7LL9vxZXosbjQkVFlDCyPYyp1tlXGZSp/RzZwxjS0mt326Xp5ftIjMpttuRuCYwPDO9vP/9VZUf/3sN97yzld8t2tTj63dX1HebXJGTYosbTeixgBJGCsvqiI+OJCel8+K7s6cPBeD1gr3dXrOjrI63N5Zw5fEjibbB+H4xPstpMW4rPbgW5b1NpWzeX8uUocn89b1tvLym2OdrW9va2VvZ2O0AsWxroZgQFHJ/QUTkThFZIyKrRGSRiAz3kWe2iCwVkXVu3iuCUdZQs72sjrGZid2Oth2aGsdxo9J8dns99GEh0RERXH1yz7PAzOEbkhRLekJ0pxbKX9/fxvDUOP5988nkjU7n+8+tYX1x97Ps91U30tqunWZ4gTMoD7DfTm40ISTkAgqwUFVnqeps4GXgdh956oGrVXU6cA7wRxFJ85FvUNleVsfYrESfz507YxjriqvZdaC+I+1AbRPP5RdxybG5HftDmf4xITupY+rwil0VLNteztfnjSMhJor7vnIcqfHR3PBEPuV1nbek75gy3KXLKyk2ioSYSDsK2ISUkAsoqur9NS0R6DZiqaqbVXWLe78YKAEG9Wq85tZ2dpfXMy7Td0A5Z4bb7bXuYLfXPz7eRVNrO9efOnZAyjiYjc9K6th1+IH3t5EaH82Vxzs7EmQnx/HAVXMoqWnilqdW0NLW3vE6X4saPbKTY63Ly4SUkAsoACJyl4jsBr6M7xaKd965QAywbSDKFqp2ldfTrp1neHkbmZHA9OEpHd1ejS1tPL50B2dOyWZCdvgfnBVsE7KTOFDXzPKdFSxav5+rTxpNYuzBdcXHjEzjV5fMdHZ6fiyfOvcI4t0VDUQIDPdxFk12sm2/YkJLUAKKiLwlIgU+bhcBqOptqjoSeBK4pYfrDAOeAL6mqu1+8twgIvkikl9aWtof1QkJ28u6Txnu6twZQ1m5q5K9VQ38e8UeDtQ1c/2p4waqiIPa+CxnptdtL6wlOjKCa04e0y3PZXNGcPfnZ/LBllK++LePKattoqi8nmGp8T4nTGSlxNqOwyakHFFAEZEIEYlw78eIyHEiknGo16nqAlWd4eP2UpesTwGX+nnvFOAV4Ceq+nEP7/Wgquapal5WVvj2inmmDPcUUM6ZMQyA19bu46EPCpk1IpUTxh7yv8sEgGfq8MZ9NVw+Z4TfkxavnDuKB6/KY/P+Gi7960esLqok10d3F0BOcpztOGxCymEHFBG5GNgL7HFbFh8AvwPWiMjnjuC6E70eXghs9JEnBngBeFxVnzvc9won28vqyEiMIS0hxm+eCdlJTMxO4k9vb6GwrI7r543rNiPM9I/hafHERkUQIXD9vJ5bhQum5fDU9SdS3dDCttK6bjO8PLJTYqlrbuvoHjMm2I6khXIHcAxwMk6309WqOh84xX3ucN3tdn+tAc4CbgUQkTwRecjN8wXgVOCr7vTiVSIy+wje86hXWFrXY+vE49wZQ6lqaCE3LZ5z3YF60/8iI4Tjx2Tw+eNGMKYX/0/HjUrn+ZtOZsrQZE4eP8RnnoNrUazby4SGI9ptWFX3AYjILlXd5Kbt9HSDHeY1fXZxqWo+cJ17/x/APw73PcLR9rI6Tu3FtvPnzxrOPe9u5fp5Y21X4QH2xNfn0pe9HMdnJfH6d071+7xnqvf+6sZefZkwpr8dUUARkQh3MPxar7RInFlXZoDUNrVSUtPUqz8qk4cm8/b3TrM/QEEgIgSyh9FzFLC1UEyoOJKvqDfgBg5VXeaVPhK4+0gKZfpmhzvDy98alK7GZSXZ2EkY6OjysoF5EyIOu4Wiqp967otIPDBKVTep6g5gx5EXzfRWoWfKsJ9V8iY8pcZHExMVYVOHTcg44k50d0bXKuB19/FsEfnPkV7X9J5n2/oxQyygDCYi4q6Wt4BiQkMgRmV/BswFKgFUdRVge3kMoB0H6shNi+90AJMZHGz7FRNKAhFQWlW16/mydi7pACos692UYRN+spPjbINIEzICEVAKRORLQKSITBSRe4CPAnBd0wuqyvbSWgsog1R2Siz7bVDehIhABJRvAdOBJuBpoBr4TgCua3qhvK6Z6sZWCyiDVHZyLNWNrby2di/ri6upaWwJdpHMIHZE61AAVLUeuM29mQG23WZ4DWrThqcAcNOTKzrSMhJj+OLckdx8+oROOxob09+O+NMmIlnAD3BaKR2nNLnbsJh+VtjHNSgmvMyfksOq2z/LrvJ6dpc3sLuinlW7KvnLu9t4fnkRPzp3Chcdk0tEhK07Mv0vEF9fngSeBS4AbgSuAcJ3n/gQs72sjuhIIdfHeRlmcEhLcDYFnTXi4KGly3eW8/P/rue7z67m8aU7WXjZLDv3xvS7QIyhDFHVvwMtqvq+ql4LnBiA65pe2F5ax6iMBNuXy3QyZ3QGL958Cr+9bBaFpXX8/L/rg10kMwgEooXiGQXcKyLnA8XAiABc1/TC9rI6xmYmBbsYJgRFRAhfyBtJwZ4qnl9eRGtbu33xMP0qEJ+uX4pIKvD/gP8BHgK+G4DrmkP4dEc528vqGGcD8qYHeWMyqG9uY+O+mmAXxYS5QMzyetm9WwWccaTXM4emqvzjk138/D/rGJmRwFUnjg52kUwIyxudDjhfQGbkpga5NCacBWKW11ictShjvK+nqhce6bVNd02tbdz+4jqezd/N/CnZ/OGK2aTGRwe7WCaEDU+LJzctnvydFXztFNsVyfSfQIyhvAj8Hfgv0B6A6xkfVJVl28v59WsbWbW7km/Nn8B3F0yy6aCmV+aMTueT7QdQVTu6wPSbQASURlX9cwCuY3worWni3yuKePbT3RSW1ZESF8X9XzmOc2YMC3bRzFHk+DHp/Gd1MUUVDYzM8H1GvTFHKhAB5U8icgewCGf7FQBUdYX/l5ie1De38taGEv6zqpj3NpXQ2q4cPyadb54xgfNmDiM+xnYVNn0zZ3QGAPk7yy2gmH4TiIAyE7gKmM/BLi91H5s+WFNUyd8/3M6b6/dT39xGTkosXztlDFccP9IWpZkjMnloMsmxUeTvqOCSY21Wv+kfgQgolwDjVLU5ANcatFra2rnm4WW0K1w0O5cLjxnO3LEZRNoYiQmAyAjh2NHp5O+oCHZRTBgLREBZDaQBJQG41qD1SWE5FfUtPHDVHM6ePjTYxTFh6PjR6fzfW5upqm8hNcFmBprAC8TCxhxgo4i8ISL/8dwCcN2Q0tLWvxPYXivYS3x0JKdNyurX9zGD15wx6ajCil2dWymNLW28tGoPTa1tQSqZCReBCCh34HR7/Qr4vdctbPzxrc1cdv9S2tv75yDKtnbljXX7OWNKlh3ja/rN7JFpREUI+TvLO6UvfGMTtz6zigffL+zX9390yXbO+9MHqNqBruHqiAOKuyFkt9vhXk9E7hSRNSKySkQWichwH3lGi8hyN886EbnxyGrRs5HpCazeXcnr6/b1y/VX7KqgrLbJpgKbfpUQE8X04Sl86jWOsnxnBQ8v2U5CTCQPLC7kQG3/HCfc0tbOfe9tY/3eaqobWvvlPUzwHXZAEZEaEan2casRkeojKNNCVZ2lqrOBl4HbfeTZC5zs5jkB+JGvwBMoFx+by4TsJH6/aBOt/dD19XrBPmIiIzhjsnV3mf6VNyaD1bsraW5tp7GljR88v5rhqfE8c8OJNLS0cc87W/vlfRet209JjROsiirr++U9TPAddkBR1WRVTfFxS1bVlCO4rncwSsSZgtw1T7Oqer5KxRKYrju/IiOE/zlrEttK63hh5Z6AXltVeb1gH/MmZpIcZwOlpn/ljU6nqbWddcVV/PntLWwrrePXn5/JrBFpXHH8SJ78ZCc7D9QF/H0fW7qDmCjn17S4sjHg1zehIST3shaRu0RkN/BlfLdQEJGRIrIG2A38RlWL/eS7QUTyRSS/tPTwz/06e/pQZuam8se3tgR08LJgTzV7Khs4e4bN7DL9b84YZ6PIRz/awQOLC/lC3ghOdSeCfOfMiURFRLDwjU3dXrd6dyX5O8q7pffGxn3VLNteztdOHgPAngproYSroAQUEXlLRAp83C4CUNXbVHUkzmmQt/i6hqruVtVZwATgGhHJ8ZPvQVXNU9W8rKzD71ISEb5/9mT2VDbwzLLdh32drl4r2EtkhPDZqT6Lb0xAZSfHMXpIAi+tKmZIYgy3nT/t4HMpcVw/bywvr9nL6t2VADQ0t/Hz/67j4vuW8JW/f8KuA30PBo8v3UlsVAQ3njaeuOgI9lQ2BKw+JrQEJaCo6gJVneHj9lKXrE8Blx7iWsXAOmBef5XXY97ETE4Ym8E972ylvrlvA4stbe0UltZ2SvN0d504LoP0xJhAFtUYv/LcbVjuumRmt52qbzhtPEMSY/j1axtYtr2cc/60mEeW7ODK40cSFRHBbS+u7dMsraqGFl5YsYcLjxlOemIMw9PiLaCEsZDr8hKRiV4PLwQ2+sgzQkTi3fvpwClA93Z64MvG98+eTFltE49+tANVZWtJDU8v28VPXyzgk8IDPl/X3NrON55Yzvzfv88v/rue5lZnYH9LSS2FZXWcYwsZzQC6+Yzx/O7yY/jstO6t4qTYKL595kQ+LiznCw8spV2Vp68/kV9/fhY/OGcyH2wp48VVvR9H/NfyIhpa2rj6pDEA5KbFs8fGUMJWIFbKB9rdIjIZZ1+wncCNACKSB9yoqtcBU4Hfi4gCAvxOVdcOROHyxmQwf0o297y9lb8tLqSi3jkBOSpCePbT3fzhitmcP+vg9N/Wtna+++wq3tlYwmmTsnh4yXbyd5ZzzxeP5fWCfYhgK+PNgBqflcT4LP/HRn9x7ije2VjC2MxEvn/2ZBJjnT8TXz5hNC+s3MOdL2/gtEnZZByiVd3ervzj450cOyqNmSOcg71y0+LZsME21QhXIRdQVNVnF5eq5gPXufffBGYNZLm8/fjcKfzgX2uYmJ1E3ugM8sakMyQxluse/5Rbnl5Bed10rjppDO3tyg//tZZX1u7lJ+dP5bp543hj3T6+/9xqLvjzhyTERnLcqHSyU+KCVRVjuomJiuCxa+d2S4+MEO7+/CzO//MH/PKV9fzfF2b3eJ0l28ooLKvjD1cc05GWmxZPWW0TjS1ttog3DIVcQDkaTMxJ5oWbT+mW/sTXT+CWp1bw05fWUVrbTEVdM/9aUcT3PjuJ6+aNA5zWyPThKXzr6ZWs3FXJ9W66MUeDyUOTufG08dz77lYuOTaXeRN9T3Rpa1fuf38bQxJjOG/mwRZ7bno8AMWVDYzroZVkjk4hN4ZyNIuLjuT+r8zh8jkj+PPbW3ji451847RxfGv+hE75RqQn8M9vnMQDV83hqpPsPHhzdLll/gTGZibyvy+spaS6+3iIqvLTlwpYsvUAty6YSGzUwZbI8DRPQLFxlHBkASXAoiIj+O1lzgDmdxdM4kfnTPF55Gp0ZARnTx/a6ZfNmKNBXHQkCy+bRVlNMxfeu4S1RVWdnv/9os089ckubjxtfMdgvEeuG1D22Gr5sGQBpR+ICDefPoFbF0y087tNWMobk8HzN51EZIRw+QMf8cqavQA89EEh9767lS/OHckPz5nc7XVDU+OIENhTYVOHw5GNoRhjDsv04am8dMsp3PjEcr751ApeXjOU1wr2cd7Mofzy4pl+W+ZDU+Js6nCYshaKMeawZSbF8uT1J3D5nBG8VrCPz0zI5A9XzO7xpFFncaN1eYUja6EYY45IbFQkv71sFl84fiQzc1MPOS6Ymx7f7ZAvEx6shWKMOWIiwvFjMnq1tiQ3LZ69lY209dOBdSZ4LKAYYwbU8LR4WtuV0pr+OczLBI8FFGPMgPIsbrRxlPBjAcUYM6BGuGtRimzqcNixgGKMGVDDOxY3WkAJNxZQjDEDKjE2irSEaIotoIQdCyjGmAGXmxZvq+XDkAUUY8yAy7WTG8OSBRRjzIAb7rZQ+nKcsAl9FlCMMQNuRHo8dc1tVDe0BrsoJoAsoBhjBpxnG/sir7Uo7e3Kd59d1bFzcW81trRZSydEWEAxxgy4jqnDXgPzrxXs44WVe7jjP+uoa+pdy2VrSS3H/uJN/rO6uF/KafrGAooxZsB5HwUMTuvkT29vJjMplrLaJh7+cHuvrvPLV9bT0NLGG+v29VtZTe9ZQDHGDLghiTHERUd0zPR6tWAvm/fXcvvnpvHZaTk8sLiQ8rrmHq/x7sYS3ttUSnpCNEu2HrDNJkOABRRjzIATEfdclAandfLWFiZkJ3H+zGH84OzJ1De3ct+7W/2+vqWtnTtfWc+4zER+esE0qhpaWLunym9+MzAsoBhjgsKzuPGVtXvZUlLLrWdOJDJCmJiTzKXHjeDxpTv9rlV5fOlOCkvr+MkFUzl9cjYi8MHm0gGugenKAooxJihy0+Ipqmjgz29vYWJ2EufNHNbx3Hc+OwkE/vDm5m6vO1DbxB/f2sypk7I4Y3I2GYkxzBieygdbygay+MaHkAsoInKniKwRkVUiskhEhveQN0VE9ojIvQNZRmPMkctNi+dAXTNbSmr5tts68X7u6hNH8+8VRWzeX9Ppdb9/czP1zW3cfsHUjnPr503MZMWuCmoaWwa0DqazUDwCeKGq/hRARL4N3A7c6CfvncD7A1UwY0zgeKYOT3THTrr65hkTePbT3Vzz8DJGpicQExVBdKTw/uZSrj5pDBOykzvyzpuYxX3vbePjwnI+Oy1nwOpgOgu5FoqqVns9TAR8Tt0QkTlADrBoIMpljAmsSTlOQPjeZycR4dU68UhPjOF3XziGiTnJRERAfXMrpbVNnD45m+8smNgp73Gj00iIieSDLTaOEkyh2EJBRO4CrgaqgDN8PB8B/B64CjjzENe6AbgBYNSoUQEvqzHm8MwckcqHPzyDEekJfvOcPX0oZ08feshrxUZFcuK4ITaOEmRBaaGIyFsiUuDjdhGAqt6mqiOBJ4FbfFziZuBVVd19qPdS1QdVNU9V87KysgJbEWPMEekpmPTVvImZbC+rY3e5HS0cLEFpoajqgl5mfQp4BbijS/pJwDwRuRlIAmJEpFZVfxTAYhpjjiLzJjpfGD/YUsaXTrDeiGAIuTEUEfHuHL0Q2Ng1j6p+WVVHqeoY4H+Axy2YGDO4jc9KZHhqnI2jBFHIBRTgbrf7aw1wFnArgIjkichDwS2aMSZUiQjzJmaxZGsZrW3twS7OoBRyAUVVL1XVGao6S1U/p6p73PR8Vb3OR/5HVdXXOIsxZpCZNymT6sZW1tg2LEERcgHFGGMO1ynjMxGB9zZZt1cwWEAxxoSN9MQYThw7hPve3crDH263g7cGmAUUY0xYuf8rczh9cja/eHk9Nz+5wrZjGUAWUIwxYSU1IZq/XT2HH587hUXr9/O5ez5kfXH1oV9ojpgFFGNM2BERvnHaeJ6+/kTqm9u4+L4l/PW9bTb7q59ZQDHGhK25YzN49dZ5zJ+czW9e38il9y9lS5fdi03gWEAxxoS1zKRY/vqV47jni8ey60Ad5//5Q+57byvtdmRwwFlAMcaEPRHhc8cMZ9F3T2P+lGx++/omFi7aFOxihR0LKMaYQSMr2WmtfPmEUfz1vW28sLIo2EUKKxZQjDGDiojwswunc+K4DH74r7Ws2FUR7CKFDQsoxphBJzoygr9+eQ5DU+K44fHlFFc2BLtIYcECijFmUEpPjOHv1+TR2NLGDU/k09DcFuwiHfUsoBhjBq2JOcn8+YuzWVdczYX3fsg7G/f73K6lqr6FldY1dkgWUIwxg9r8KTk8dHUeLW3tXPtoPl/62ycU7KmiqbWN1wv28Y0n8jn+rre45L6PeOqTXcEubkiTwbR5Wl5enubn5we7GMaYENTc2s5Tn+zkT29voaK+heS4KGoaW8lMiuXCY4azeX8NHxce4MnrTuCEcUOCXdwBJSLLVTXvkPksoBhjzEHVjS38bXEhe6saOX/WMOZNyCQqMoKqhhYuuW8JlfUtvPTNUxiZkRDsog4YCyg+WEAxxhyJwtJaLvrLEnLT4vnXTSeTGBsV7CINiN4GFBtDMcaYXhqXlcRfvnQcm/fX8L1/rqKp1WaGeRsc4dUYYwLk1ElZ3Hb+NO58eT2Tf/I6qfHRZCbFkJUcy7DUeEamxzMiI4GR6QlMHppMRmJMsIs8YCygGGNMH117yhhy0+LZsr+G0tomymqbKKluYtn2cl5a1YBn38mEmEgevCqPz0zMDG6BB4iNoRhjTAA1t7azt6qBXeX13PXKBgrL6rjvS8exYFpOsIt22GwMxRhjgiAmKoLRQxKZNzGLp68/kSlDk7nxH8v57+rijjyqyvKd5fzq1Q28uX5/EEsbWNblZYwx/SQ9MYYnrzuBrz+az63PrKS0pon65lb+tWIP28vqAHhwcSE3njae7589mcgICXKJj0zIBRQRuRO4CGgHSoCvqmqxj3xtwFr34S5VvXDgSmmMMb2THBfNY9fO5YYn8vnFy+sBOGFsBjedPp4FU3NY+MYm7n9/G+uKq/jzlceSfhQP4ofcGIqIpKhqtXv/28A0Vb3RR75aVU3qy7VtDMUYEyxNrW28smYvx4/J6LYo8pllu7j9pXVkp8Ty7fkTKa5qYGtJLVtLaimrbWJURgKTcpKZmJPMhOwkhqbEkZ4YTXpCDNGR/T9yERYLG0Xkx8AoVb3Jx3MWUIwxYWPlrgpu+scK9lU3IgIj0uOZkJVEdnIc2w/UsWV/DRX1Ld1elxwXRVJsFDFREcRGRRATFUF8dCTJcdGkxEWREh9NclwUlx43gnFZffqT2aG3ASXkurwAROQu4GqgCjjDT7Y4EckHWoG7VfXFgSqfMcYE2rGj0nnr/51GUUU9Y4YkEhcd2S1PWW1TR6ulor6FirpmyuuaqW9upam1naaWdprb2qlvbmV/dSNbS1qpbmyhprGVk8ZlHnZA6a2gtFBE5C1gqI+nblPVl7zy/RiIU9U7fFxjuKoWi8g44B3gTFXd5iPfDcANAKNGjZqzc+fOQFXDGGOOCqqKKkQc5qB/uHR5jQZeUdUZh8j3KPCyqj7fUz7r8jLGmL47atehiMhEr4cXAht95EkXkVj3fiZwCrB+YEpojDHGl1AcQ7lbRCbjTBveCdwIICJ5wI2qeh0wFXhARNpxguLdqmoBxRhjgijkAoqqXuonPR+4zr3/ETBzIMtljDGmZyHX5WWMMeboZAHFGGNMQFhAMcYYExAhPW040ESkFGegvzcygbJ+LE6osnoPLlbvweVw6z1aVbMOlWlQBZS+EJH83sy7DjdW78HF6j249He9rcvLGGNMQFhAMcYYExAWUPx7MNgFCBKr9+Bi9R5c+rXeNoZijDEmIKyFYowxJiAsoHQhIueIyCYR2SoiPwp2eQ6HiDwsIiUiUuCVliEib4rIFvdnupsuIvJnt75rROQ4r9dc4+bfIiLXeKXPEZG17mv+LCIhcRC2iIwUkXdFZIOIrBORW930sK67iMSJyDIRWe3W++du+lgR+cStw7MiEuOmx7qPt7rPj/G61o/d9E0icrZXesj+XohIpIisFJGX3cdhX28R2eF+DleJcy5UaHzOnX3y7eZ2/UUC24BxQAywGucI4qCXrY/1OBU4DijwSvst8CP3/o+A37j3zwNeAwQ4EfjETc8ACt2f6e79dPe5ZcBJ7mteA84Ndp3dcg0DjnPvJwObgWnhXne3LEnu/WjgE7c+/wSudNPvB25y798M3O/evxJ41r0/zf3MxwJj3d+FyFD/vQC+BzyFc4QFg6HewA4gs0ta0D/n1kLpbC6wVVULVbUZeAa4KMhl6jNVXQyUd0m+CHjMvf8YcLFX+uPq+BhIE5FhwNnAm6parqoVwJvAOe5zKaq6VJ1P3uNe1woqVd2rqivc+zXABiCXMK+7W/5a92G0e1NgPuA5I6hrvT3/Hs8DZ7rfQC8CnlHVJlXdDmzF+Z0I2d8LERkBnA885D4WBkG9/Qj659wCSme5wG6vx0VuWjjIUdW94PzhBbLddH917im9yEd6SHG7M47F+bYe9nV3u31WASU4fxi2AZWq2upm8S5rR/3c56uAIfT93yMU/BH4Ac5xF+DUYzDUW4FFIrJcnFNpIQQ+5yG3fX2Q+eonDPdpcP7q3Nf0kCEiScC/gO+oanUP3b9hU3dVbQNmi0ga8ALOmUHdsrk/+1o/X188g15vEbkAKFHV5SJyuifZR9awqrfrFHWOQM8G3hSRbgcRehmwz7m1UDorAkZ6PR4BFAepLIG2323K4v4scdP91bmn9BE+0kOCiETjBJMnVfXfbvKgqDuAqlYC7+H0laeJiOdLo3dZO+rnPp+K00Xa13+PYDsFuFBEduB0R83HabGEe71R1WL3ZwnOF4i5hMLnPNiDS6F0w2mxFeIMzHkG4aYHu1yHWZcxdB6UX0jnAbvfuvfPp/OA3TI3PQPYjjNYl+7ez3Cf+9TN6xmwOy/Y9XXLJTj9vX/skh7WdQeygDT3fjzwAXAB8BydB6dvdu9/k86D0/9070+n8+B0Ic7AdMj/XgCnc3BQPqzrDSQCyV73PwLOCYXPedA/CKF2w5kRsRmnD/q2YJfnMOvwNLAXaMH5tvF1nL7it4Et7k/PB0eAv7j1XQvkeV3nWpwByq3A17zS84AC9zX34i6QDfYN+AxO03wNsMq9nRfudQdmASvdehcAt7vp43Bm62x1/8jGuulx7uOt7vPjvK51m1u3TXjN7An13ws6B5Swrrdbv9XubZ2nXKHwObeV8sYYYwLCxlCMMcYEhAUUY4wxAWEBxRhjTEBYQDHGGJPOQ38AACAASURBVBMQFlCMMcYEhAUUE1ZEpM3dgXW1iKwQkZMPkT9NRG7uxXXfE5FBdwZ5T0TkURG5LNjlMKHDAooJNw2qOltVjwF+DPz6EPnTcHahDUleK76NCXkWUEw4SwEqwNnfS0Tedlsta0XEs2vs3cB4t1Wz0M37AzfPahG52+t6l4tz7shmEZnn5o0UkYUi8ql71sQ33PRhIrLYvW6BJ78390yL37jXXCYiE9z0R0Xk/0TkXeA37jkXL7rX/1hEZnnV6RG3rGtE5FI3/SwRWerW9Tl3bzNE5G4RWe/m/Z2bdrlbvtUisvgQdRIRude9xisc3HzQGMA2hzThJ97ddTcO53yU+W56I3CJOptFZgIfi8h/cLaomKGqswFE5FycrbpPUNV6EcnwunaUqs4VkfOAO4AFOLsQVKnq8SISCywRkUXA54E3VPUuEYkEEvyUt9q95tU4+1Bd4KZPAhaoapuI3AOsVNWLRWQ+zvYys4Gfuu890y17ulu3n7ivrRORHwLfE5F7gUuAKaqq7iaSALcDZ6vqHq80f3U6FpgMzARygPXAw736XzGDggUUE24avILDScDjIjIDZ/uJX4nIqThbnefi/FHsagHwiKrWA6iq97kyns0ml+PslQZwFjDLaywhFZiIsxfSw+5mlS+q6io/5X3a6+cfvNKfU2cHYXC2lLnULc87IjJERFLdsl7peYGqVrg78E7DCQLg7EG1FKjGCaoPua2Ll92XLQEeFZF/etXPX51OBZ52y1UsIu/4qZMZpCygmLClqkvdb+xZOHsyZQFzVLXF3aE2zsfLBP9bdTe5P9s4+LsjwLdU9Y1uF3KC1/nAEyKyUFUf91VMP/frupTJ1+t8lVVwDk36oo/yzAXOxAlCtwDzVfVGETnBLecqEZntr05uy8z2ajJ+2RiKCVsiMgVn19gDON+yS9xgcgYw2s1Wg3NcsMci4FoRSXCv4d3l5csbwE1uSwQRmSQiiSIy2n2/vwF/xzmS2ZcrvH4u9ZNnMfBl9/qnA2WqWu2W9Rav+qYDHwOneI3HJLhlSgJSVfVV4Ds4XWaIyHhV/URVbwfKcLYz91kntxxXumMsw4AzDvFvYwYZa6GYcOMZQwHnm/Y17jjEk8B/RSQfZxfijQCqekBElohIAfCaqn7f/ZaeLyLNwKvA//bwfg/hdH+tEKePqRRnDOZ04Psi0gLUAlf7eX2siHyC8+WuW6vC9TPgERFZA9QD17jpvwT+4pa9Dfi5qv5bRL4KPO2Of4AzplIDvCQice6/y3fd5xaKyEQ37W2cHWzX+KnTCzhjUmtxduB9v4d/FzMI2W7DxgSJ2+2Wp6plwS6LMYFgXV7GGGMCwlooxhhjAsJaKMYYYwLCAooxxpiAsIBijDEmIEJu2rA4+yl9DmgGtgFfU9VKH/l24EyFbANaVdV2gjXGmCAKxRbKmzh7K83Cmev+4x7ynuHuLGvBxBhjgizkWiiqusjr4cdAwM5byMzM1DFjxgTqcsYYMygsX768TFWzDpUv5AJKF9cCz/p5ToFFIqLAA6r6oK9MInIDcAPAqFGjyM/P75eCGmNMuBKRnb3JF5SAIiJvAUN9PHWbqr7k5rkNaAWe9HOZU1S1WESygTdFZKOqLu6ayQ00DwLk5eXZohtjjOknQQkoqrqgp+dF5BqccyHOVD8rL1W12P1ZIiIvAHNxNq8zxhgTBCE3KC8i5wA/BC70nEnhI0+iiCR77uOc31AwcKU0xhjTVcgFFOBenO3E3xTn+NT7AURkuIi86ubJAT4UkdXAMuAVVX09OMU1xhgDITgor6oT/KQX4xyShKoWAscMZLmMMcb0LBRbKCYMrC+u5jvPrKS1rT3YRTHGDBALKKZfLN5Syouritlb1RjsohhjBogFFNMvqhtaADhQ1xzkkhhjBooFFNMvqtyAUl7XFOSSGGMGigUU0y88AeVArbVQjBksLKCYflHd2ApAuXV5GTNoWEAx/eJgl5cFFGMGCwsopl/U2KC8MYOOBRTTL6yFYszgYwHFBJyqHhyUt4BizKBhAcUEXENLG63tzibRNm3YmMHDAooJOE/rJCEmknKbNmzMoGEBxQScJ6CMzUykrrmNxpa2IJfIGDMQLKCYgKtucNagjMlMBGxg3pjBwgKKCThPC2WcBRRjBhULKCbgPBtDjhniBBSb6WXM4GABxQRcxxhKlqeFYjO9jBkMLKCYgKvq2kKxmV7GDAoWUEzAVTe2kBwbRXpCNFERYmMoxgwSFlBMwFU1tJASH42IkJ4YYwHFmEHCAooJuOqGVlLiowEYkhhjg/LGDBIWUEzAVTe0kBIXBUCGtVCMGTRCLqCIyEIR2Sgia0TkBRFJ85MvTUSed/NuEJGTBrqsxreqhhZS3RbK0RBQ1hRV8vCH24NdDGOOeiEXUIA3gRmqOgvYDPzYT74/Aa+r6hTgGGDDAJXPHEJ148GAMiQxhgO1oT1t+Ollu/nVqxtQ1WAXxZijWsgFFFVdpKqt7sOPgRFd84hICnAq8Hf3Nc2qWjlwpTQ98QzKA2QkxlLd2EpLW3uQS+VfaU0Tre3asWWMMebwhFxA6eJa4DUf6eOAUuAREVkpIg+JSKKvC4jIDSKSLyL5paWl/VlWA7S0tVPf3HawyyspBoCKEO72Kq1pBOCALcA05ogEJaCIyFsiUuDjdpFXntuAVuBJH5eIAo4D/qqqxwJ1wI98vZeqPqiqeaqal5WV1Q+1Md482654BuWHJDoBJZRnepXWOIEk1Md6jAl1UcF4U1Vd0NPzInINcAFwpvru2C4CilT1E/fx8/gJKGZgeVbJpyYcHJSH0P1jraqUumM8oRz0jDkahFyXl4icA/wQuFBV633lUdV9wG4RmewmnQmsH6Aimh5UNzrjEN6D8hC6f6wr61toafOcLhmaZTTmaBFyAQW4F0gG3hSRVSJyP4CIDBeRV73yfQt4UkTWALOBXw18UU1XVR1dXl1aKCE606vUq1wWUIw5MkHp8uqJqk7wk14MnOf1eBWQN1DlMr3jGUPxtFDSEmIQCd0/1iXVBwOKbWJpzJEJxRaKOYp1tFDcgBIZIaQnhO72K6W1zgyvyAixbfaNOUIWUExAVXVpoUBor5b3tFDGZiaGbNAz5mhhAcUEVHVjCzFREcRFR3akZYTwBpGlNU3ER0cyKiMhZIOeMUcLCygmoJyNIaM7pQ0J5RZKTRPZKbEh3Yoy5mhhAcUcll+9uoHlOyu6pVc3tJIa33muRyj/sS6taSIrKbZjm33bz8uYw2cBxfRZRV0zDy4u5MWVe7o9572Pl8eQxBgq6ptpaw+9P9altQdbKM2t7dQ1twW7SMYctSygmD7bfqAOgJ3l3dedem9d75GRGIMqVNaHXiulpLqRrKRYr/UyoVdGY44WFlBMn+0ocwOKG1i8eW9d75GRFAuE3lqUxpY2qhtbyU6JY0iSZ0X//2fvvsPbrM7Gj39veVse8Y4znNjZZIeQsBLIKFCgrLZQSktoC5S2tFA6KW/pj7eTAm3fFmhLKVBaZspqmWGFmeXshEw7doa3HW/LQzq/P/TIkW15RrZk+f5cly5Lj44enSexdeus++jUYaUGSgOK6rdDVkA5dryJtk5p6Wu6GZSH4Eu/4kkK6W6hBGfQU2o40YCi+s0TUNpchqJqR/txYwy13XR5QfB9WHvSrqTFRwVt0FNqONGAovqtoLKBuKjw9vse9c1tuAwkdJrlFawf1p5FjWnxUUEb9JQaTjSgqH4xxlBQ0chZk1OAjgPzvlbJAyQF6YC3p4WSHh9FbGQYUeE2DShKnQQNKKpfKupbqG9uY3F2ClHhNg57tVA8W+h2DigRYTYSosODLldWeV0zNoGUuChExL0WJciCnlLDiQYU1S+eLq7sNDtZybEUVHZtoXQelAf3h3awdXmV1zlItkcRZhPAvV1xsAU9pYYTDSiqXzwD8jmpdiak2DnsK6DEdA0owbhavryumbT4qPbHyfaooKujUsOJBhTVLwUVDYTbhLGjYpiQEkthVUN7upJah+8xFAjOgFJW10y6V0BJCeIklkoNBxpQVL8UVDaQlRxLeJiNCSmxOFpdlFnrOWp7aKEE44d11xZK8AU9pYYTDSiqXw5VNDIx1Q7AhBT3z0Kr26u2qRURiI/quhFosj2S40GUfNHlMpR3aqEk2yNpbHHiaNV8XkoNhAYU1WfGGAorG5hoBZIJybHAiRQsNU2txEeFY7MGub0l2yNpc5n2mWCBVt3USpvLdGihBOt6GaWGCw0oqs/K6pppbHGSneoOJGOTYgizSXsLpaaplcTYrt1dQNDlympPu9KphQLBt15GqeFCA4rqM88ML0+XV0SYjbGjYtoXN9Y62nxOGQaCLldWWZ07ZUx6fHT7sWALekoNN0EXUETkHhHZKyI7ROQFERnlo8w0EdnmdasVkVsDUd+RpD2gWF1egHuml1eXl68ZXhB83Um+WyjBFfSUGm6CLqAAbwKzjDFzgP3A7Z0LGGP2GWPmGWPmAacCjcALQ1vNkaegooHIMBtjRsW0H3MHlBOD8t0FlHFW99jWw9V9fj9jDDuP1uAahI25PDPTOg/KgwYUpQYq6AKKMWaNMcYzcrseGNfLS1YAecaYwsGtmTpU0UBWSmz7ynKACcl2appaqW5s8Zm63mNUbCTnTE3jxa3H+rxz4ys7i/nM/R/y/dXbu6TJP1nldc3ERoZh95qRlhAdTkSYBE0rSqnhJugCSidfBV7rpcwXgKeGoC4jXoHXDC+PrBTPTK/GHgflAT67YBwltQ4+zqvo0/u9sOUYUeE2nt96jJuf3Epzm/+m85Z1WoMCICIkxUbqoLxSAxSQgCIib4nILh+3S73K3AG0AU/0cJ5I4BJgdQ9lbhSRXBHJLS8v9+dlBKWi6qYO6VD8xeUyFFY2ts/w8vAEmANl9TS3uUiI7roGxWPFjHQSosN5fkvXveg7O97Qwnv7y1l15kTuvPgUXt9dwg2Pb6bJT3u+l9c5OnR3eSQH4QJMpYaLgAQUY8xKY8wsH7eXAERkFXAxcI3peSXcp4EtxpjSHt7rIWPMQmPMwrS0NP9eSBD60XM7uOHxXL+ft7jWQXObq32Gl0eWtRZl51H32Eh3YygA0RFhXDx3DK/vKqG+uef1KK/tKqHNZbhk7hi+enY2v/3sHD48UM61j2xoT/FyMjqvkvdI0QSRSg1Y0HV5icgFwI+AS4wxvX3Vvhrt7urgk6Ja9pXWUVrr6L1wP3j2kc/uFFBiIsNIj49ix7EawHfaFW+fXTCOplYnr+4s7rHcS9uOMSnNzswxCQBcedp4/nj1fLYcrub+dw4O9DLaufN4RXc5rgkilRq4oAsowP1APPCmNSX4LwAiMkZEXvUUEpFY4FPA84GpZvCpqG9u767p6zhFXx3qJqCAu9vrk6JaoPeAsiBrFNmpdp7bfLTbMsU1TWwsqOLSeWMROTEB4OI5YzhrcipvfdJtg7RPHK1O6hxtvlso2uWl1IAFXUAxxkw2xoz3TAs2xtxkHS8yxlzoVa7RGJNijKkJXG2Dy/7Suvb7Hx2s9Ou5CyoaiI6wkeHjW31WSizNbe5ZWD11eYF74PuK+WPZcKiKI1W+G6Avby/GGLhk7pguz62Ynk5+RQP55fUDuAq39jUocb7HUOocbbS0+XdWmVIjQdAFFDVw+0vcAeXUCUl8fLDCr4kYPTO8fOXpmphyYqC+u2nD3i5fMBaAF7b6Hpx/afsx5o5L7DJeA7B8ejoA7+wt61O9ffGsQUlL8B1QAI43aitFqf7SgBJC9pfVkxgTwWXzx1JU4+iwm+LJOlTRdcqwR5bX8d5aKADjkmI5PSeZ57cc7RL0DpbVs+tYLZfMG+vzteOTY5maEXdSAaXcSrviq4XSvqJfpw4r1W8aUELI/pI6pmXEc9akFAA+OuifcRSny3C4qtFniwFOZB0GSIjpftqwt88uGEdBZSNbDh/vcPw/24sQgc/Myez2tcunZ7DxUNWAZ3t5urzSe2ih6MC8Uv2nASVEGGPYX1rHlIw4slPtZCZG+21g/tjxJlqdpssaFA9PyyU6wkZUeFifzvnp2ZnERITxk+d38Z/tRbQ6XRhj+M+2Y5yRk0J6QtexGo8VM9Jpcxk+2D+w6yuva8YmkGL3PW0YNEGkUgOhASVElNY2U+toY9roeESEMyelsi6v0i95sNZ8UgLApLQ4n88nxkaQGBPRp+4uj7iocO67ci7NbU6+89RWltz9Lj/7z24KKhu5dF7XwXhv88ePYlRsBG/vHdhsr7K6ZpLtUR1SyHhogkilBq5v/RMq6HlmeE1JjwfgrMkpPLflKJ8U1zJrbOKAz/vRwQp+/dpeVkxPZ0FWUrflJqTE9nsV+4WzM7lg5mjW7i/jkQ8LeHxdIZFhNi6Y2X13F0B4mI1zp6axdl85TpfxGRgcrU4OVzVyqKKBo8ebyEiIYmpGPNmp9i47NXobFROBTTSgKDUQGlBChCegTM1wtyLOmpwKuNejDDSgFFQ08M0ntjApzc4fvjDP5wwvjy8tnjCgMQ2bTVg+PYPl0zM4UFpHfXNbj/nAPJbPyODFbUVsO3KcUycktx9/Z28pd/33Ew5XNeJrklu4TRCBMyeldlufpFhdi6LUQGhACRH7S+tIjYskxZq5lJEQzaQ0Ox8drOTGpZP6fb5aRyvXP56LCDx87WnE9zId+MrTxg+o3t6mZMT3uew5U9MIswlv7ylrDyg7j9bwrSe2Mj45hu8sn0J2qp3sVDtjk2IoqXFwoKyO/aX1HCyr77FbLdmuCSKVGggNKCFiX2k9Uzt9IJ81OZXVuUdpaXMRGd734TKny3Dr09soqGjg8a8tas8oHEwSYyI4bWIS7+wt44cXTOdYdRNf/ccmku2R/Ov6xV3SqqTGRfW5pZZsj9QuL6UGQAflQ4DLZThYWtcloJw5KZWmVifbjvR9UytHq5MfrN7OO3vL+NklM7vtGgoGK6ZnsLekjn0ldXztsU04Wpw8ct1pPnN09UdKXOSgzPJ665PSfv1fKDXcaEDpgzc/KeV/Xtzp15Xn/nSsuomGFmeXgHJGTgo26ft6lEMVDVz2wEc8v/UY3105lS+fPmEwqus3y2e4V81f/bf1HCir58EvLWDa6L53m3VnMFoo1Y0t3PzUFu5bs8+v51UqmPg1oIjIN/15vmCx61gN/1p/mF++sicog8qBso4D8h6JsRHMGpvYp/Uor+0s5jN/+pCSWgePfuU0blk5ZVDq6k85qXYmpsRS1dDCLy6bxZIp/tmeINkeRXVTa593luyL1blHcbS62pNsKhWKBjyGIiK3dT4E3C4i0QDGmN+dTMWCya0rp1DT1MrDHx5iVGwENy8Prg/bfSXuRIm+BrXPnJTKQ+/ncf0/NrE4O4XFOcmckplAraONvSW17CupY3PhcV7eUczc8aN48JoFjPXaMz6YiQg/u2QmZbUOrjoty2/nTbFHYow7n1eqj/Qs/eV0GR5fXwC4W5OOVifREX1bAKrUcHIyg/J3Aa8Cu3EHE4Aw3KnnQ4qIcOfFp1Db1Mq9a/aTEBPBtWdMDHS12h0orWN0QrTPhYVfX5pDTVML6/IqeWuPO/9VZJiNFq892pNiI/ja2dn86ILp/Rq8DwbLpqX7/Zye9Ct7i+s4a3JkhxT6A7F2XxlHqpo4f2YGb+wupbCy0S9dc0oFm5MJKDOB3wF24C5jTKOIrDLG3OWfqgUXm024+3NzqHW0cedLu0mIdidhDAb7SuuY2s0HVJI9kl9fMQeA0loHGw5VsfNoNenx0UwbHc/00fGkxUed9IdmKMlOtWMT+NLfNzAhJZZl09JZPj2dMyelEB7W/4D72McFjE6I5qZzJvHG7lIOVdRrQFEhacABxRhzGPictQ/8myLye/9VKzhFhNm4/4vzue7RjXxv9XayU+3MHT9qyN6/1eli7b5yzp2WRoT1weZ0GQ6W1XNGTkqvr89IiOaSuWN87jOiTpg1NpH3frCMtfvKeGdvGU9tPMxjHxdwRk4Kf7321D6l6PfIK6/ngwMVfO9TU9u7JPN1HEWFqJPu37D2gf8UsBjofhu+EBEdEcZD1y4kNS6Sn7ywkzbn0G3E9OhHh7jh8Vx+/vIn7ccOVzXS3ObqtoWiBmZ8cixfPmMij35lEdvuPI9fXj6LTQVVXPXX9ZT1Y3vlf1rpZK5enEVcVDjp8VHkl2tAUaHJXx3mBnjYGLPUT+cLagnREdx58Ux2F9Xyz/WFQ/KezW1O/v7hIWIjw3h8XSHPbDoMeKdc0YAyWGIiw7hm8QQeue40CisbuPzBj8nrw46R9c1t/HvzUS6ak9k+uJ+dateZXipknXRAEZFLgG3A69bjeSLyn5M9b7C7cPZozpmaxn1r9lNS0/dvrAP10rYiSmubeeCaBSyZkspPX9zN5sLj7bs0Tkn3nQlY+c/SqWk8fePpOFqdfO7PH3fZy6WzF7Ycpb65jWvPOLGeJyctTgOKCln+aKH8DFgEVAMYY7YBE/1w3qAmIvzvpTNpdbo6dEENBpfL8Nf38jglM4Fzp6bxp6vnMzoxmm/8azMf51UyLikGe5Rm0RkKc8aN4rlvnEl8dARfeXQTh7vZFdPlMvxjXSFzxyUy3ytLc06qnaqGFqp1i2EVgvwRUNqMMTV+OM+wMyHFzs3LJvPKzmLW7hv4lrS9eXtvGXnlDXz9nBxEhFGxkfzt2oXUN7exLr+SadrdNaQmptr519cWY4zh6//a3CVtvzGGu/67m4Nl9Xz17OwOz2Vbu17qwLwKRf4IKLtE5ItAmIhMEZE/AR/74bzDwo3n5JCTZufOl3bzcV4FD72fx7ee2MKS377Dqkc2ctwPKTz++l4eY0fFcNHsE/uETBsdz++unAvAjMyEk34P1T9ZKbH839Xz2VtSy09e6JiW5741+/nHukJuXJrTZUZdTpo7oBzSgXkVgvwRUL6Ne01KM/AUUAvcOtCTicg9IrJXRHaIyAsi4nNeroh8V0R2i8guEXnKs0J/qEWFh/GLS2dxuKqRL/5tA796dS/bj1YzY3QC6/IrueLPH1NYOfAPj9yCKnILj3PDkuwuayAumJXJv286gxuW5JzsZagBWDYtne+unMoLW4/x+Dr35Iy/vpfH/e8e5Aunjef2T0/vsr5nfHIsYTYhv6L3QX2lhhsJttxUInIe8I4xpk1E7gYwxvyoU5mxwIfAKcaYJhF5FnjVGPNYT+deuHChyc3NHZR6v/VJKWE2Yc64xPY9SXILqrjh8VxsIvxt1cL2HQ+rG1t4ZWcxb+8po7GlrcN5po9O4IoFY5k9NhER4fp/5LK5sIqPfryc2EgdJwk2Lpfhxn/msnZfOV86fQKPfVzARXMy+eMX5vvcSRJg2b1rmZEZz4PXnDrEtVVqYERkszFmYW/lTvoTSkTSgB/ibqW0txKMMcsHcj5jzBqvh+uBz3VTNByIEZFWIBYoGsj7+cvKUzK6HFs4MZnnv3kW1z26kasfWs93Vkxh6+HjvLe/nFanITvVTprXVrQul+FJaxHd5PQ4PnVKBm/tKeWWFVM0mAQpm02478p5XHr/hzz2cQHnTkvj91fO6zaYgHtgXteiqFDkj0+pJ4BngIuBm4BVQLkfzgvwVevcHRhjjonIvcBhoAlY0ykQtRORG4EbAbKy/JdAsK+yU+08/40zuf7xXO55Yx+jE6L5ylnZXDJ3DDPHJHTpEqlpauXVncU8v+Uof16bR0xEGKvOnDjk9VZ9lxgTwSPXncaLW4/xjXMn95oPLTvVzkd5FbhcpsdtlZUabk66y8tqCp0qIjuMMXOsY+8ZY87p4TVvAaN9PHWHtfIeEbkDWAhcYTpVUkSSgOeAq3BPV14N/NsY86+e6jqYXV69aWlzUVDZwKS0uB6/vXorrGxwr4LXWVwh5ckNh/nJCzv56MfLh01mZzWyDVmXF9Bq/SwWkYtwdz2N6+kFxpiVPT0vIqtwt3hWdA4mlpXAIWNMuVX+eeBMoMeAEkiR4bZ+B4YJKfZBqo0KJM/U4UPlDRpQVEjxxyyvX4hIIvA94PvAw8B3B3oyEbkA+BFwiTHG96oxd1fX6SISK+4+oxXAnoG+p1JDyTN1WGd6qVBz0i0UY8zL1t0aYNnJng+4H4jCncEYYL0x5iYRGYM7X9iFxpgNIvJvYAvQBmwFHvLDeys16NLjo7BHhunAvAo5/pjllY17LcpE7/MZYy4ZyPmMMZO7OV4EXOj1+Ge4074oNayICNlpmiRShR5/jKG8CPwd+C8wdLnclRrGslPj2Hak5+SSvjz20SHyKxr4f5+ZqTPEVNDxR0BxGGP+6IfzKDVi5KTaeXlHEc1tTqLC+7a/fFOLk/vW7KeuuY2s5Fiu1wwJKsj4Y1D+/0TkZyJyhogs8Nz8cF6lQlZOmh1joLCbbMW+vL67mLrmNqakx3H363vZeXRE5mRVQcwfAWU2cAPwG+A+63avH86rVMhqzzrcj4H5ZzYdISs5lme+fgapcVF8+6kt1De39f5CpYaIPwLK5UCOMeYcY8wy6zagtCtKjRTta1G8BubLah38c10BjlZnl/KFlQ2sz6/iyoXjSLZH8oer5nG4qpE7X9w1VFVWqlf+CCjbAZ8ZgZVSvsVHR5AWH0V+eT1tThcPf5DP8vve46cv7eZ3b+7vUn517lFsAp891b1meHFOCt9ZMYXntx7juc1Hh7r6Svnkj0H5DGCviGzCncIeGPi0YaVGiuxUOxsLqrj4Tx+yt6SOc6elYY8M5+EP8rlk7hhmjU0EwOky/HvzUZZOTSMz8cTK+m8vn8LHeZX89KVdnDU5ldGJAdnBQal2/toC+HLgV5wYQ7nPD+dVKqRNSrNTWNlInaONv375VB697jR+dflsku1R/Pj5HbQ53bPw3z9QTkmtg6sWju/w+jCbcO/n5tLqdPHHdw4E4hKU6sAfK+Xf80dFlBppbliSw/TRCVy5sb2NWwAAIABJREFUcDwxke6pw4mxEfy/S07h5ie38tjHBVy/JIdnNx0hxR7Jihldt0jISonl6kVZPLnhMDcuyWFiquZ/U4Ez4BaKiNSJSK2PW52I1PqzkkqFopy0OFadObE9mHhcNDuTFdPTuW/NfrYfqeatPaVcPn9st2nxb142mfAw4fdvdR17UWooDTigGGPijTEJPm7xxhjd5FypARIR/veyWYjAlx7eQKvTcOVp47stn27tsfOf7UXsKdbvcipw/DGGopTys7GjYvj+edOoa25j3vhRvW59cNPSScRFhXPfmn1DVEOlutKAolSQWnXmRL64OIvbPjW117KJsRHcdM4k3tpTxubCqiGonVJdaUBRKkiF2YRfXT6bpVPT+lT+K2dNJDUukt++vo+T3YlVqYHQgKJUiIiNDOfmZZPZcKiKtfvKA10dNQJpQFEqhFy9OItJaXbueGEntY7W3l+glB9pQFEqhESFh3Hv5+dSUuvgV6/orthqaGlAUSrEzM9K4oalOTy96Qjv79euLzV0NKAoFYK+u3Iqk9Pj+PFzO7TrSw0ZDShKhaDoiDDu+dwc7fpSQ0oDilIhyrvra+2+skBXR40AGlCUCmHfXTmVKelx3PrMNg6W1QW6OirEBV1AEZF7RGSviOwQkRdExOfmXSJyi4jsEpHdInLrUNdTqeEgOiKMv686jXCbjVWPbKKkxhHoKqkQFnQBBXgTmGWMmQPsB27vXEBEZuHex34RMBe4WESmDGktlRomslJieewrp1HT1Mp1j26kpqn/g/TVjS3sL63D6dIV+Kp7QRdQjDFrjDFt1sP1wDgfxWYA640xjVbZ93Bv8qWU8mHW2ET+8qVTySuv58bHc33uW98dp8tw7SMbOe/37zP3rjV8+e8b+P2b+9l2pHoQa6yGo6ALKJ18FXjNx/FdwFIRSRGRWOBCwGd+bxG5UURyRSS3vFzn5KuR6+wpqdz7+blsOFTFd5/Z1r4jZG+e3nSYHUdruGFJNpfNH0NFfQt/eucAVzz4kY7LqA78sad8v4nIW8BoH0/dYYx5ySpzB9AGPNG5kDFmj4jcjbt7rB7YbpXtwhjzEPAQwMKFC7W9rka0S+eNpbyumV+8sofvPrud3185l/Cw7r9XHm9o4Z439rE4O5mfXDgDEQHgWHUTZ9/9Dv/ZXsxtn+o5tb4aOQISUIwxK3t6XkRWARcDK0w3aVONMX8H/m6V/xVw1N/1VCoUXb8khzaX4Tev7cUYwx+umtdtUPntG/uoc7Txv5fOag8m4N6v5fTsFF7eUcR3V07p8JwauYKuy0tELgB+BFxijGnsoVy69TMLuAJ4amhqqNTwd9M5k7j909N5eUcxt3bT/bXjaDVPbzrMqjMmMm1011bIxXMzyS9v4BPdJVJZgi6gAPcD8cCbIrJNRP4CICJjRORVr3LPicgnwH+BbxljjgegrkoNW18/ZxI/udAdVG55eluHKcUul+HOl3aTYo/i1k/5nkD56VmZhNmEl3cUD1WVVZALSJdXT4wxk7s5XoR78N3zeMmQVUqpEHXj0kkIwi9f3cMrO4uZNTaBFdMzEIFtR6q57/NzSYiO8PnaZHskZ05yd3v98Pxp2u2lgi+gKKWG1g1Lczh3WhprPinl7T2l/PGdAxgDCyckccWCsT2+9jNzxvDD53aw42gNc8f7XIOsRhANKEoppmTEMyUjnm8tm0xFfTMf51Vy2sSkXlsd588czR0v7uTlHUUaUFRQjqEopQIoNS6KS+aOITMxpteyibERLJmSxis7inHpKvoRTwOKUuqkfGZuJkU1DrYe0XkxI50GFKXUSVk5I4PIcBv/3a6zvUY6DShKqZMSHx3BsmlpvLKzWJNHjnAaUJRSJ+3iOWMor2tmfX5loKuiAkgDilLqpK2YkU5qXBR3v75XWykjmAYUpdRJi40M587PnMKOozX8c11BoKujAkQDilLKLz4zJ5MlU1K5d81+3RlyhNKAopTyCxHhF5fNotXp4q7/7g50dVQAaEBRSvnNhBQ731kxhdd2lfDO3tJAV0cNMQ0oSim/umFJDlPS4/jpi7tpbPG5750KURpQlFJ+FRlu45eXz+ZYdRO3PbNdx1NGEA0oSim/W5SdzPfPm8rbe0s55553+fVre6hpbA10tdQgk2522A1JCxcuNLm5uYGuhlIjxpGqRn735n5e3HaM+KhwvnZ2DpfOG8PEVHugq6b6QUQ2G2MW9lpOA4pSarDtKa7lnjf28c7eMgBmZCbw6VmjuXB2JpPT4wJcO9UbDSg+aEBRKrCOHm/k9V0lvL6rhNxCd3biP1w1j8vm97yRlwosDSg+aEBRKniU1jr49lNb2XG0mhe+eRYzMhMCXSXVjb4GFB2UV0oFREZCNA98cQGJMRHc9K/N1DTpoP1wpwFFKRUwafFRPHjNAoqqm7jtmW266+MwF3QBRUR+LiI7RGSbiKwRkTHdlFslIges26qhrqdSyj9OnZDMTy8+hbf3lnH/uwcBKKpuYnXuEW57Zhu3PbuNj/MqGEnd88NV0I2hiEiCMabWuv8d4BRjzE2dyiQDucBCwACbgVONMT3uQapjKEoFJ2MMtz27nRe3HSMrOZbCykYAku2RtDld1DramJgSy1WnZfG5U8eRFh8V4BqPLH0dQwkfisr0hyeYWOy4A0Zn5wNvGmOqAETkTeAC4KnBr6FSyt9EhF9dPpuaplZsAteeMZEzJ6UwLSOeFqeLV3cW8/TGI9z9+l7uXbOPM3JSuHB2JufNzCA1ToNLsAi6FgqAiPwSuBaoAZYZY8o7Pf99INoY8wvr8U+BJmPMvT2dV1soSg1vB8vqeG7LMV7bWUxBZSM2gdNzUvjJhTOYNTYx0NULWUE9bVhE3gJG+3jqDmPMS17lbscdOH7W6fU/AKI6BZRGY8x9Pt7rRuBGgKysrFMLCwv9dyFKqYAwxrCnuI7XdhXzbO4R6h1t/PlLp7J0alqgqxaSgjqg9JWITABeMcbM6nT8auBcY8zXrcd/BdYaY3rs8tIWilKhp7TWwXWPbuJAaR13f3YOnz11XKCrFHKG7ToUEZni9fASYK+PYm8A54lIkogkAedZx5RSI0xGQjTPfP10FmUn873V23lw7cEeZ4QdqWrkHx8X4Gh1DmEtR4agG5QHfiMi0wAXUAjcBCAiC4GbjDHXG2OqROTnwCbrNf/rGaBXSo08CdERPPaVRXx/9XZ++/o+Nh6q4jNzxrB8ejpJ9kgA9pfW8ee1efxnexFOl6Gp1clN50wKcM1DS1B3efmbdnkpFdpcLsMD7x7kiQ2HKal1EGYTFk1MJiYyjHf2lhEbGcbVi7LYcbSaw1WNfPDD5USGB11HTdAZttOGlVJqoGw24dsrpnDz8snsPFbDmt2lvPlJKRWlzdyyYgrXnTmRJHsk7+0vZ9UjG3lp2zE+v3B8oKsdMrSFopQacYwxfPr/PsDpMrxx61JsNgl0lYLasB2UV0qpwSYifP2cHA6U1bN2f1mgqxMyNKAopUaki+eMYUxiNH99Lz/QVQkZGlCUUiNSRJiNr56dzYZDVWw7Uh3o6oQEDShKqRHrC4uyiI8O56H38wJdlZCgAUUpNWLFRYXz5dMn8NquEj4pqtUU+SdJpw0rpUa0686ayMMfHuLCP35AVLiNsaNiGDMqhsnpcZwzNY3Tc1KIiQwLdDWHBZ02rJQa8fYU17Ihv5Jj1U0UVTs4Wt3EvpJaHK0uIsNtLM5O5pypaSydmsaU9DhERtY045BIDulvGlCUUn3laHWy8VAV7+0v57395RwsqwcgMzGaJVNSOXtKGin2SIwBY23blJMWx9hRMYGs9qDQgOKDBhSl1EAdPd7IBwcqeH9/OR8erKDO0daljAicPTmVzy8cz3mnZBAdERpdZRpQfNCAopTyhzaniz3FdTRZGYtFwBj4OK+C1blHOVbdRGJMBJfPH8uXTp/A5PS4ANf45GhA8UEDilJqsLlcho/zKnkm9whv7Cqhxeni7MmpfPmMCayYno7TGA6W1bOvpI4DZfXMHpvIBTNHB3X6Fw0oPmhAUUoNpYr6Zp7ZdIQn1hdSVONgVGwE9Y422lzuz11Py2ZqRhzfWTGFC2dldggsTpeh3tFGYmxEoC4B0IDikwYUpVQgtDldvLWnjDWflJCZGM300QnMyIwnK9nO67tL+OPbBzhYVs+U9DiWTU+nsLKB/PIGCisbaXG6yEqOZVF2MosmJrNgQhKtThdF1U0UVTdxrNoBwOiEKEYnxjA6MZrUuEhiIsKItm5hJ9n60YDigwYUpVQwcroMr+4s5k/vHCC/vIEJKbHkpMWRk2YnMSaCbYer2VRQxfHG1i6vjQyzYTC0Orv/LI8Ms/G3VQs5Z2ragOqn+6EopdQwEWYTPjN3DBfPycRl8NmicLkMeeX1bD1SjT0ynDGjohmbFEOqPQqAqsYWSmoclNQ4qGxoxtHqwtHqpKnViaPVxfikwZ/OrAFFKaWChIgQ1k3vlM0mTMmIZ0pGvM/nU+OiSI2LYtbYxEGsYc80l5dSSim/0ICilFLKLzSgKKWU8gsNKEoppfxCA4pSSim/0ICilFLKL0bUwkYRKQcK+1g8FagYxOoEK73ukUWve2QZ6HVPMMb0uipyRAWU/hCR3L6sDA01et0ji173yDLY161dXkoppfxCA4pSSim/0IDSvYcCXYEA0eseWfS6R5ZBvW4dQ1FKKeUX2kJRSinlFxpQlFJK+YUGlE5E5AIR2SciB0Xkx4Guz0CIyCMiUiYiu7yOJYvImyJywPqZZB0XEfmjdb07RGSB12tWWeUPiMgqr+OnishO6zV/FJGg2AxbRMaLyLsiskdEdovILdbxkL52EYkWkY0ist267rus49kissG6hmdEJNI6HmU9Pmg9P9HrXLdbx/eJyPlex4P270JEwkRkq4i8bD0O+esWkQLr93CbiORaxwL/e26M0Zt1A8KAPCAHiAS2A6cEul4DuI6lwAJgl9ex3wI/tu7/GLjbun8h8BogwOnABut4MpBv/Uyy7idZz20EzrBe8xrw6UBfs1WvTGCBdT8e2A+cEurXbtUlzrofAWywrudZ4AvW8b8A37DufxP4i3X/C8Az1v1TrN/5KCDb+lsIC/a/C+A24EngZetxyF83UACkdjoW8N9zbaF0tAg4aIzJN8a0AE8Dlwa4Tv1mjHkfqOp0+FLgH9b9fwCXeR1/3LitB0aJSCZwPvCmMabKGHMceBO4wHouwRizzrh/8x73OldAGWOKjTFbrPt1wB5gLCF+7Vb9662HEdbNAMuBf1vHO1+359/j38AK6xvopcDTxphmY8wh4CDuv4mg/bsQkXHARcDD1mNhBFx3NwL+e64BpaOxwBGvx0etY6EgwxhTDO4PXiDdOt7dNfd0/KiP40HF6s6Yj/vbeshfu9Xtsw0ow/3BkAdUG2ParCLedW2/Puv5GiCF/v97BIM/AD8EXNbjFEbGdRtgjYhsFpEbrWMB/z3XLYA78tVPGOrzqru75v4eDxoiEgc8B9xqjKntofs3ZK7dGOME5onIKOAFYIavYtbP/l6fry+eAb9uEbkYKDPGbBaRcz2HfRQNqeu2nGWMKRKRdOBNEdnbQ9kh+z3XFkpHR4HxXo/HAUUBqou/lVpNWayfZdbx7q65p+PjfBwPCiISgTuYPGGMed46PCKuHcAYUw2sxd1XPkpEPF8avevafn3W84m4u0j7++8RaGcBl4hIAe7uqOW4Wyyhft0YY4qsn2W4v0AsIhh+zwM9uBRMN9wttnzcA3OeQbiZga7XAK9lIh0H5e+h44Ddb637F9FxwG6jdTwZOIR7sC7Jup9sPbfJKusZsLsw0Ndr1Utw9/f+odPxkL52IA0YZd2PAT4ALgZW03Fw+pvW/W/RcXD6Wev+TDoOTufjHpgO+r8L4FxODMqH9HUDdiDe6/7HwAXB8Hse8F+EYLvhnhGxH3cf9B2Brs8Ar+EpoBhoxf1t42u4+4rfBg5YPz2/OAI8YF3vTmCh13m+inuA8iDwFa/jC4Fd1mvux8q4EOgbcDbupvkOYJt1uzDUrx2YA2y1rnsXcKd1PAf3bJ2D1odslHU82np80Ho+x+tcd1jXtg+vmT3B/ndBx4AS0tdtXd9267bbU69g+D3X1CtKKaX8QsdQlFJK+YUGFKWUUn6hAUUppZRfaEBRSinlFxpQlFJK+YUGFBVSRMRpZWDdLiJbROTMXsqPEpFv9uG8a0Vkof9qOvyJyGMi8rlA10MFDw0oKtQ0GWPmGWPmArcDv+6l/CjcWWiDkteKb6WCngYUFcoSgOPgzu8lIm9brZadIuLJGvsbYJLVqrnHKvtDq8x2EfmN1/k+L+59R/aLyBKrbJiI3CMim6y9Jr5uHc8Ukfet8+7ylPdm7Wlxt3XOjSIy2Tr+mIj8TkTeBe629rl40Tr/ehGZ43VNj1p13SEin7WOnyci66xrXW3lNkNEfiMin1hl77WOfd6q33YReb+XaxIRud86xyucSD6oFKDJIVXoibGy7kbj3h9luXXcAVxu3MkiU4H1IvIf3CkqZhlj5gGIyKdxp+pebIxpFJFkr3OHG2MWiciFwM+AlbizENQYY04TkSjgIxFZA1wBvGGM+aWIhAGx3dS31jrntbjzUF1sHZ8KrDTGOEXkT8BWY8xlIrIcd3qZecBPrfeebdU9ybq2/7Fe2yAiPwJuE5H7gcuB6cYYYyWRBLgTON8Yc8zrWHfXNB+YBswGMoBPgEf69L+iRgQNKCrUNHkFhzOAx0VkFu70E78SkaW4U52Pxf2h2NlK4FFjTCOAMcZ7XxlPssnNuHOlAZwHzPEaS0gEpuDOhfSIlazyRWPMtm7q+5TXz997HV9t3BmEwZ1S5rNWfd4RkRQRSbTq+gXPC4wxx60MvKfgDgLgzkG1DqjFHVQftloXL1sv+wh4TESe9bq+7q5pKfCUVa8iEXmnm2tSI5QGFBWyjDHrrG/sabhzMqUBpxpjWq0MtdE+XiZ0n6q72frp5MTfjgDfNsa80eVE7uB1EfBPEbnHGPO4r2p2c7+hU518vc5XXQX3pklX+6jPImAF7iB0M7DcGHOTiCy26rlNROZ1d01Wy0xzNalu6RiKClkiMh131thK3N+yy6xgsgyYYBWrw71dsMca4KsiEmudw7vLy5c3gG9YLRFEZKqI2EVkgvV+fwP+jntLZl+u8vq5rpsy7wPXWOc/F6gwxtRadb3Z63qTgPXAWV7jMbFWneKARGPMq8CtuLvMEJFJxpgNxpg7gQrc6cx9XpNVjy9YYyyZwLJe/m3UCKMtFBVqPGMo4P6mvcoah3gC+K+I5OLOQrwXwBhTKSIficgu4DVjzA+sb+m5ItICvAr8pIf3exh399cWcfcxleMegzkX+IGItAL1wLXdvD5KRDbg/nLXpVVh+X/AoyKyA2gEVlnHfwE8YNXdCdxljHleRK4DnrLGP8A9plIHvCQi0da/y3et5+4RkSnWsbdxZ7Dd0c01vYB7TGon7gy87/Xw76JGIM02rFSAWN1uC40xFYGui1L+oF1eSiml/EJbKEoppfxCWyhKKaX8QgOKUkopv9CAopRSyi80oCillPKLEbUOJTU11UycODHQ1VBKqWFl8+bNFcaYtN7KjaiAMnHiRHJzcwNdDaWUGlZEpLAv5bTLSymllF9oQFFKKeUXGlCUUkr5hQYUpZRSfqEBRSmllF9oQFFDJr+8Hs0dp1To0oCihsTR442s/N17PLHhcKCropQaJBpQ1JDYW1yHy8DqzUcDXRWl1CDRgKKGRH5FPQDbj1STV14f4NoopQaDBhQ1JPLLG7BHhmETeHHrsUBXRyk1CDSgqCGRV17PKWMSOGtyKi9sPYbLpYPzSoUaDShqSOSXNzApLY7L54/l6PEmcguPB7pKSik/04CiBl1NYyuVDS3kpNk5f+ZoYiLCeGGrDs4rFWo0oKhBl2cNyOekxmGPCueCWaN5eUcxjlZngGumlPInDShq0OWXNwCQk2YH4PL5Y6lztPHO3rJAVksp5WcaUNSgyyuvJ9wmjE+OBeCsyamkx0fx/Bad7aVUKAm6gCIiPxeRHSKyTUTWiMiYbsq9LiLVIvLyUNdR9U9+eT0TUmKJCHP/uoXZhMvmj2XtvjKqGloCXDullL8EXUAB7jHGzDHGzANeBu7srhzw5aGrlhqo/PIGctLiOhy7fP5Y2lyGl3cUBahWSil/C7qAYoyp9XpoB3wuWDDGvA3UDUml1IA5XYbCysb28ROPGZkJTEyJ5cMDFQGqWfcOlNYx887XKahoCHRVlBpWgi6gAIjIL0XkCHAN3bdQ+nquG0UkV0Ryy8vL/VNB1WdHjzfS4nQxKTWuy3Pjk2MprWsOQK16tvNYDQ0tTvaX6vcVpfojIAFFRN4SkV0+bpcCGGPuMMaMB54Abj6Z9zLGPGSMWWiMWZiWluaP6qt+8MzwmpRu7/JcRkI0pTWOfp3vgwPllNX17zX9VWzVqVLHd5Tql4AEFGPMSmPMLB+3lzoVfRL4bCDqqPzDkwgyx0cLJSMhivL6Zpx9TMPiaHVy3aObePSjAn9WsYtj1U0AVNYHX+tJqWAWdF1eIjLF6+ElwN5A1UWdvLzyBpJiI0iyR3Z5bnRCNE6X6fMH9+GqRpwuQ7H1gT9YPOevqNcWilL9EXQBBfiN1f21AzgPuAVARBaKyMOeQiLyAbAaWCEiR0Xk/MBUV/Ukv7y+ywwvj/SEaABKa/sWUA5Zg+R9LT9Qni6vCm2hKNUv4YGuQGfGGJ9dXMaYXOB6r8dLhqxSasDyKxo4d6rvsavRVkApqXUwm8Rez+WZdVU6yGMoRe1dXtpCUao/grGFokJEraOV8rpmJqX7bqGMTjwRUPqioNIdUMoGsYVS39xGraMNgMoGbaEo1R8aUNSgac/hldp1hhdAij0Sm0BZHwOKp8urvrmN+uY2/1SyE8/4SXx0uLZQlOonDShq0OR7Znh1M4YSHmYjLT6Kkj5OHS6oaCQq3P0r29cg1F9FVl1mjUmkqrGlzzPQlFIaUNQgyi9vIMwmZFlJIX3JSIju0+LGphYnJbUOFmQlAYM3MO9pocwel4gxcLxRWylK9ZUGFDVo8ivqmZAcS2R4979mfV3cWFjl7u5anJMMMGiLG4tqHIjAzDEJgA7MK9UfGlDUoHEnhfQ9fuKRkRDVp1lbnhlei7NTACgdrC6v6iYy4qPJsGag6dRhpfpOA4rqk6aW/u2u6HQZ8iu6ZhnubHRCNNWNrb3u3nioohGAWWMTiI0MG7wur5omMkdFkxrnXoipAUWpvtOAonq19fBxZv7sdZ7eeLjPrymqbqKlzdXtDC8PT0ugt6nABRUNpMZFEh8d4e4mG6QWSnG1gzGJMaTYowDt8lKqPzSgqF7tKqrFZeD2F3by4ta+7bLoyeHV3RoUj4yEvq1FOVTZwMQUd3BKj48alLUoxhiKaprITIwmMSaCMJvoWhSl+kEDiurV4coGIsNtnJ6dwvdWb+f1XSW9vmZdXiUiMLm3Lq8+Lm4sqGhgotXacc8M838Lxd315iJzVAw2m5Bsj9QWilL9oAFF9aqwspGs5FgeXrWQueMS+fZTW3h3X1m35cvqHDy+rpBL547xmRTSW0a8p8ur+wDR0NxGWV0z2e0BJYrSWgfG+HeNiCfL8NhR7jql2CM1QaRS/aABRfWqsLKRiSmx2KPCeeyri5g+OoGb/rmZDfmVPsv/eW0eLU4Xt6yc2uu5E2LCiY6w9bi4sbDSPSDv6fLKSIjG0epqT5HiL56kkJmJMQCkxkVpl5dS/aABRfXIGMPhqkaykt0f5gnRETz+1UWMS4rhW09u7dKyKKpu4on1h/ncgnHtLYqeiEivixs9ObwmproXSKYn9N6qGYjiGncLJdPTQomL1FleSvWDBhTVo/K6ZppanUxIObHaPckeyZ+/dCoNzW185+mttDld7c/96Z2DAHxn5ZQu5+pOb4sbPTm82lso8e4ZWP6eOlxU7SAiTEi1ZnilxkXpGIpS/aABRfWosMrd3ZSV0jF9ytSMeH5+2SzW51fxf28fcJetbGB17hGuXjSesaNi+vweo3sZZC+oaCA9Pgp7lHu3hYz2fVT830IZnRiNzSaAu4XS2OKksWVwElEqFWqCbj8UFVw84xcTfOTj+typ49h0qIo/vXOQUyck8Z/tRYTZhG8tm9yv98hIcCeINMYgIl2eL/CaMgyQnmC1UPw806uouql9/ARob6lU1rcQm6x/Kkr1RlsoqkeHKxuwCYxL8p3g8a5LZzJ9dDy3PL2NF7ceY9WZE9vHOPoqIyGa5jYXtU2+WwKHKhrbx08AYiPDiY8O9/talKJqR4eWVYq1Wr6yQbu9lOoLDSiqRwWVjYwZFdNtgsfoiDAeuGYBbU4XMRFh3HTOpH6/R0+LG+scrVTUN7evQfF+jT+7vJwuQ2mtg8zEE8EwJc7TQtGBeaX6QgOK6lFhVWOHAXlfJqXF8eQNp/P3604juZd1J770tLjR0+WWndI5oET1KaA0tTj58XM72pNLdqeivpk2lyHTu4ViXYsOzCvVNxpQVI8OVza0Txnuydzxozg9J2VA7zG6h0H2E1OGOwWU+Og+zfJ6auNhnt50hDd297y637OP/JgOLRR3QCnXFopSfaIBRXWr1tHK8cbWXlsoJyvNMw3Yx9Thgk5Thj3SE6Ipq+t5tXxLm4u/fZAPnNiOuDtF1R0XNYJ7rCY2MkxbKEr1kQaUEHPXf3dzy9Nb/XKuwz3M8PKn6IgwkmIjfM7aOlTRyOiEaGIiwzocz0iIotVpON7Y2u15X9x6jOIaB/HR4eRX1PdYB8+ixs7TnXW1vFJ9F3QBRUR+LiI7RGSbiKwRkTE+yswTkXUistsqe1Ug6hqMNh6qYs3u0g6LDQeqfcpwSu9dXicrIyGakpquH9wFlQ0dZnh5l4fu16I4XYa/vJfHzDEJXDQ7s08aaqkVAAAgAElEQVQtlNjIMBJiOk4PTonTBJFK9VXQBRTgHmPMHGPMPOBl4E4fZRqBa40xM4ELgD+IyKihrGSwKqlx0NTqZF9p3UmfyzN+0XlR42DIsLqwutShoqFLd5e7vGe1vO+A8vquEvIrGvjWssnkpNmpbGihpofWTLGVtr7zOpgUe5SmX1Gqj4IuoBhjar0e2oEuneTGmP3GmAPW/SKgDEgbmhoGr+Y2Z/uaia2Hq0/6fIcrG0mNiyQuavAX9XkWN3qrdbRS2dDSZUAeID2++425jDE8uPYgOal2zp85mpxUdwr9vB66vYpqHIzxsbo/NS5S16Eo1UdBF1AAROSXInIEuAbfLRTvsouASCCvm+dvFJFcEcktLy/3f2WDiPeHqz8CSmFVA1mDPH7iMToh2j1116ur7lC57wF58Fot76OF8t7+cnYX1XLTuZMIs0n7vvY9dXu5V8l3XZCZEhdJVUMLLpd/U+UrFYoCElBE5C0R2eXjdimAMeYOY8x44Ang5h7Okwn8E/iKMcbnoIEx5iFjzEJjzMK0tNBuxHjSr8dHhbP1yPGTPt/hysYhGT8ByEiMxmXosP/IvzcfJdwmzBvftTczKrz7gfwH1+aRmRjNZfPGAjA+OZZwm5Bf7ruF0tLmoqK+ucMML48UexROl6GmqfvuMqWUW0ACijFmpTFmlo/bS52KPgl81tc5RCQBeAX4H2PM+sGu83DgWRi48pQM8ssbqG4ceFdNc5uT4lrHkLVQPBttea6hqLqJZzYd4fMLx7cvfOzymoSua1E2F1ax8VAVNy7NaV/dHxFmIyslttsWinuzrq4zvODEWhQdR1Gqd0HX5SUi3nnPLwH2+igTCbwAPG6MWT1UdQt2JdbU1wtmjQZg25GBd3sdqWrCGHzOsBoMnqDh6cJ64N2DGAw3L+8+0WR6QnSXPVEe/aiAxJgIrjptfIfjOalx3U4d9ixq9OyD4i3VSr+iOzcq1bugCyjAb6zurx3AecAtACKyUEQetspcCSwFrrOmF28TkXkBqm/QKK5xYI8M46zJqYic3DjK4SprhlcfVsn7g/eYyNHjjTybe4QrF/acBj8jPqpDC6WszsHru0r43KnjiI3sOJFgUpqdgspGnD7GQjrv1OjNE1B0LYpSvQu6nNzGGJ9dXMaYXOB66/6/gH8NZb2Gg9JaB6MTo4mLCmdaRjxbT6KFUlDhWYMyNC2UVHsUYTahtNbBA+/mIfSeBj8jIZry+macLkOYTXh20xHaXIZrFmd1KZuTZqelzcWx401dpkF79pIf46OF0p5xWFsoSvUqGFsoaoCKaxzt37LnZ41i+5HqAc9OOlzViD0yrD1B4mCz2YT0+Cg2Fx5nde4RrjptvM9pvN4yEtwD5pUN7qDy1MYjnDU5hZy0uC5lPcd8TR3eX1pHenxUl1YNQFJsJCKacVipvtCAEkJKahztK8jnj0+ipqmVQ5U9rxDvTmFlA1kpdp8bXg2WjIRo1udXYRPhm8t6T4N/Ym/5Zt7dW8ax6ia+fPoEn2VzUn1PHXa5DB8drODMSb4TW4bZhOTYSCp0LYpSvdKAEiKcLkNZXXP7Wor5We6ptgMdRymsamTiEHV3eXiyDn9xcZbP8YzOvNOv/GtDIRkJUayckeGzbLI9ksSYiC5Th/eW1FFR38LZU7qfUu5Ov6ItFKV6owElRFRYYwkZVkCZlBbnXo9yuP/rUZwuw9GqrmMNgy0rJZaocBvfOLdvm3R50q9sKjjOe/vL+cJpWYSH+f6VFhGyU+1dWigfHnQvdj17cmq37+NOv6ItFKV6owElRLTPVLK+tdtswrysUQNqoZTUOmhxupgwRDO8PL61bDKvfGdJe8ujN6lxUYjA4+sKsIlw9aKug/HectLsXaYOf3Cggsnpcd2udQFtoSjVVxpQQoQnD5b3B+P88aPYW1JLY4vvvdq7U2jtQTJUM7w8EmMimJzedUC9OxFhNlLsUTS2OFk5I73HoADuVltpbTP1ze5/D0erk42HqnpsnYCVwn4QWigPvHuQq/66rsc9XZQaTjSghAjPosYOASUrCZeBHUdr+nWuwir3lOGhWiV/MjzdXl8+fWKvZT0D854cYZsLj9Pc5mLJlN4CSiR1zW04Wp0nV1kvr+wo5p439rHhUBV7ik8+M7RSwUADSogoqW0mMsxGcuyJab6eHFj97fbaW1xLRJj0Om03GGSn2pmSHtftLC1vnqnDnm6vDw5UEG4TFveydXGKtbixyprp1ep08f3V27n9+R0DqvMnRbV8f/V2Zo5JAODdfWUDOo9SwUYDSogoqWkiIzEKm+3ENN8keyTZqfZ+DcwfqmjgqU1HuGBWJmG2oZsyPFC/vmI2q286o8N1d2dCSiwikGe1UD46WMGCrKRe0/N71uJU1rfgdBlue3Y7/958lOc2H+t3q6WqoYUbHs8lMSaCR687jdljE3l3rwYUFRo0oISI4hpH+7Rbb/PHj2LrkWqfKUc6M8bwPy/uJCrcxk8vmjEY1fS7+OgIRsX2bfFldEQY45JiyC+v53hDC7uKaji7l+4uONFCKa938OPndvDf7UUsn55Oi9PVr9Zfq9PFN5/YTHl9M3/98qmkJ0SzbHo6Ww4fP6lEnkoFCw0oIcKddqVrF9XSqWmU1zXzmT99SG5BVY/neHHbMT46WMkPL5jevmgw1OSkxpFf3sBHeRUYQ58CSqqVfuUXr+xh9eaj3LJiCr+/ah4isOFQZZ/f+1ev7mF9fhW/uWI2c63uyGXT0nAZeP9AxcAuSKkgogElBBhjrBZKVJfnLp03hgevWUB1Ywuf+8s6bntmW5cMvQDVjS384uU9zBs/imt6mX47nOWk2TlU0cAH+yuIjw5nztjEXl/jaaHklzfw9aU53LpyCokxEcwYncDGQz0HaY9dx2p47OMCrj1jAlcsGNd+fM64USTbI1mr3V4qBARdckjVf9WNrTS3uXy2UESEC2dncu60NB549yB/e/8Qaz4p5YuLs/jy6RMYb83k+s1re6luauVfV8zu03jEcJWTFkdTq5NXdxZz5uSUbhdCerNHhjF9dDxnTU7lx5+e3p6OZnFOMk9tPExLm6t97xVfjDHc9d/dJMdG8v3zp3V4LswmnDM1jbX7y3G5TEj/26vQpy2UEODZlMrXFrYesZHh/OD86az57lLOmZbG3z88xNJ73uX6f2zikQ8P8fSmI1y/JJsZmQlDVe2AmGRNHa5rbusx3Yo3EeG1W5bw04tP6ZDbbHF2Mo5WFzuP9TyO8srOYjYVHOf7508jITqiy/PnTkujqqGFHcf6N71bqWCjLZQQ4FnU2JcV5hNT7TzwxQUU1zTx5IbDPLnhMG/tKWNcUgy3rJjS6+uHO+9MxEt6WdDozVeSzEXZ7unG6/OrOHVCss/XOVqd/PrVvczITODKheN9llk6JQ2bwDt7y3xud+ytsr6Z5jbXsJjSrUYebaGEgBMbRPV9ID0zMYbvnTeNj29fzgNfXMAj153mM317qMlIiMIe6Z7tdbKZAJLtkUzNiGNDD+MoD72fz7HqJn72mVO6nYadZI9kflYSa/uwHuUH/97BZ//8MS1trgHXW6nBogElBJTUOrAJpMV3HZTvTVR4GBfNyWRqRvwg1Cz4iAiXzBvDNYsn+CU1/6LsZDYXVNHm7PoBX1Lj4M9r87hw9mhO72Xx5LJpaew4WkN5Xfc5w1raXKzLq6S4xsGL246ddN2V8jcNKCGgpKaJ1LgoIvowwKzg11fM6XNG494szk6hocXJ7qLaLs/d/fpenMZw+6d7X9OzbHo6AO/tL++2zPaj1TS1OokKt/HQ+/kD3jxNqcGin0AhoKS2uV/dXcp/Fme7x046r0fZeKiKF7Ye44Yl2e0z6XpySmYC6fFRPaZhWZdXiQj85MIZHCyr522daqyCjAaUEFBS09Rrpl01ONIToslOtXdYj9Lc5uT253cwLimGby2b3KfziAjLpqXz/v5yn91n4A4oM0YncM3iLMYlxfCX9/L8cg1K+YsGlBDQXdoVNTQWZyez8VBVe3qbP6/NI6+8gV9cNqtfEx2WTU+jztHmc7Gko9XJ5sPHOWOSe+3MDUty2Fx4vNfsB0oNJQ0ow1xDcxt1jjafixrV0FiUnUyto419JXUcLKvjwXfzuHTeGM6dlt6v85wzNZ24qHBe2Np1wH3r4Wpa2lycYQ3uf37hOJJiI7SVooKK3wKKiMSJyAIR6Xkife/n+bmI7BCRbSKyRkTG+CgzQUQ2W2V2i8hNJ/Oew5lnUePoxP7P8FL+4Ul/vz6/kp88v4uYyDB+evEp/T5PTGQYF84ezas7i2lq6ZjFeF1eBTaBRTnuMZvYyHBWnTmRt/aUsb9U91NRwWHAAUVEHvS6fzbwCXAfsFNELjyJOt1jjJljjJkHvAzc6aNMMXCmVWYx8GNfgcdfjlU38eYnpX49Z355PSvuW8vS377LlX9Zx81PbuEXL3/Crn6ulm7fqTFBWyiBMnZUDGNHxfCndw6wsaCKOy6aQWrcwAL8FQvG0dDiZM0nJR2Or8uvZNbYxA4r7a89YyLRETb++l7+SdVfKX85mRbK6V73fw5cZoxZBpwD/O9AT2qM8Z5/aQe6zI00xrQYYzwT9qMY5K67e9/Yx03/2sxrO4v9cr5Wp4tbn9lGRX0L87NGIeJOHviPdQX89KVd/TpXyQAWNSr/W5yTzPHGVk7PSebzp47r/QXdWDQxmbGjYnhuy4lur6YWJ9uOVHNGp03Eku2RfOG0LF7adozjDZr+XgWevz6IE4wxWwCMMflA2MmcTER+KSJHgGvw3UJBRMaLyA7gCHC3Maaom3I3ikiuiOSWl3c/x78nP79sFvPHj+LbT23l9V0lvZZ3uQy3PbuNrz22qX3/cm9/fPsAO47W8Osr/j979x0eV3UtfPi3Rr3LqlZzk+VuuSBsU0wz1YBNSS6QQEiAAAkkkHYJF24gX8olIbkhhCSEFkIuNQmmBmObajAGd1uucpcsq7mo2err+2OOZEmekWV5pBnJ632eeWZmn33O7GNLWrP7RH5/7RRevu00PvzRufzwwtGs2n2QHc6e7t1xpMnLAoo/XTguldjwYH555cQTmjDpcglXTc3gk4JySp3/2+W79tPYrG39J+1dPGEwTS3K6sLj25XTmN5wIgFljNPXsQ4YJSKDAETEBRy9Al47IrJIRPI9POYCqOp9qpoFPA/c6ekaqlqoqrnASOBGEUn1ku8JVc1T1bzk5O4tBthZdFgwf/3GqeRmxnHnCytZsL7roPLnj7bx6so9vLepjOuf+pzKw41tx1bs2s8fP9jK1VMzmT0xrcN5cydnIILHTllv9lYeJj4yhPCQE4rh5gRdPCGN1T+5sMNaYT115ZQMWhRed2bDf7ZtH8Eu4dRhR68XNjEjDpdgAcUEhBMJKGOBy4HLgAlAjZOegJdaRStVPV9VJ3h4vN4p6wvA1ce4VjGwHpjZk5vorpjwEP520zQmZMRxxwsrWeSlT+XjLeX8ZsFm5kxK5y83nMKG4iq+8uRS9tc2UFPfxPdeXkN6fAQPzjm603ZwXDhnZCcxb1URqt2bBV1SWW9DhgOEr5aeH5EczZQh8bzqNHst2baP3Mw4ojxsVRwVFkxOSgxriiygGP/rcUBR1V2tD6AMGOGkV6jqqz29roi0X/J2DrDJQ55MEYlwXg8CzgA29/QzuysmPITnbp7GuLRYvvX8Cn67YHOH0ThFBw5x10urGJUSw0NXT+Si8YN58sY8tpbVcM1fPuOef62l6MAhfnfNZGI8LGMO7m+nhfsPs3xX9/aBL6mySY0D0VVTM9lUUs2ynftZt6fyqP6T9iZnxbOm8GC3v4QY01tOuA9FRC4HVgPznfeTReSNE7jkQ07z11rgQuAu57p5IvKUk2cs8LmIrAE+An6jqutO4DO7LTY8hOduns7siWn84f2tzPrth7y9di91jc186/9W0tSsPH7DKW0T2s4elcyz35jGnoOHeXvtXr51TrbHpotWF08YTERIUNu30/Yqaur5xdsb+PlbG3j0vQL+tmQnhfsPW4f8AHR5bhohQcL98/JpblFOG+F9qf1JWfEcONTI7v2H+rCExhzNF+uVPwhMAz4EUNXVIjKspxdTVY9NXKq6HLjFeb0QyO3pZ5youIgQfn/tFL46fSgPvLGeO15YSWpsGKVV9Tz5tTyGO5s4tTotO5EXvzmDBRtKuGvWqC6vHRUWzEXjU3l7bTEPXD6urW+kpUX53surWbJtH2HBLg61qxll+6Dd3gSW+MhQZo1JZf76EkKDXJwydJDXvK17qKwuPMjQxCiv+Yzpbb4IKE2qWumLpcD7m2nDE3jrO2fywhe7+f2iLXz/glFcMM7j2AAmZcUz6RibJ7W6amomr60u5oNNZVzidNw/uXg7iwsq+OWVE/nK9CE0NrdQdbiRQw3NZNhmSwPSVVMzmL++hMlD4okI9T7oYlRqNOEhLlYXHmTu5Iw+LKExHfkioOSLyFeAIKf/47vAEh9ct18Icgk3zBjK9dOH+GR/DYAzRiaREhPGq6v2cMnENNYUHuThdzdzyYTBXDfNvetfSJCLxOgwut5lw/Rn54xOYURyFJd2Gg3YWXCQi4kZcayxkV7Gz3wxD+U7wHigHngRqALu9sF1+xVf1tCCXMLcyel8uLmMwv2H+O5Lq0iNDeehq3J9+jkmsIUGu3j/B+dw4+nDjpl3clY8+cVVtpOj8asTDiiqesiZN3KqM9/jPlWt80XhTmZXTsmksVn58uOfUbj/EL+/djJxkV1O7zEnsUlZ8TQ0tbC5xNb1Mv5zwk1eIpIM/CfuWkrbcCNVPe9Er30yG5cey5jBMWwqqeb7F4wir4uRYca0dcwXHWRiZpyfS2NOVr5o8noe91yR4cBPgZ3AMh9c96R39/mjuG5aVrc3aTInr4z4CJKiQ1m92/pRjP/4olM+UVWfFpG7VPUj4CMR+cgH1z3pXTxhMBdPGOzvYph+QESYlBlvM+aNX/mihtK6UNVeEblURKYAPV9u1RjTI5Oz4tlWXkNVXeOxMxvTC3wRUH4uInHAD4AfAk8B3/PBdY0xx2FSVjyqsK7o+PbUMcZXTrjJS1Xfcl5WAuee6PWMMT0zKfPIjPkzRnpfqsWY3uKLUV7Dcc9FGdb+eqo650SvbYzpvrjIEEYkRdlS9sZvfNEp/xrwNPAmYLOqjPGjSVnxfLK1AlW1SbCmz/kioNSp6qM+uI4x5gRNzopn3qo97K2sI93WeDN9zBed8r8XkQdE5DQRmdr68MF1jTHHqXUB0sUFPdvu2pgT4YsaykTgBuA8jjR5qfPeGNOHJqTHkpsZx8/f2kjesATb2sD0KV/UUK4ERqjq2ap6rvOwYGKMHwQHufjz9acQEuzitr+voKa+yd9FMicRXwSUNUD3NvowxvS6jPgIHvvKFHZU1PLDV9bY1sCmz/gioKQCm0TkXRF5o/Xhg+saY3ro9Owk7r1kDPPXl/CnD7f5uzjmJOGLPpQHfHANY4yP3XzmcNYWVfKbBZuZkBHH2aOS/V0kM8D5Yqa8LQRpTAASEX51dS7riyv55dsbOSsnyeammF7V4yYvEakWkSoPj2oRqfJlIY0xPRMRGsRNZw5nc2m1zaA3va7HAUVVY1Q11sMjRlVjfVlIY0zPzZmUTkRIEC8vK+zR+dV1jXxSUMHGvfY90XTNF30oPiUiPwPm4p7TUgZ8XVWLveSNBTYC81T1zr4rpTH9R0x4CJdPSuONNcXcf9k4osOO/Wv/waYyFmwoYeWug2wpq0YVYsKD+fCH55AYHdYHpTb9kS9Gefnaw6qaq6qTgbeAn3SR92eA9eEYcwzXnDqEQw3NvLXG43ezDubnl/CNZ5fx9tq9DI4L5+5Zo/jdNZM43NDMbxZs6YPSmv4q4Gooqtq+Xh2Fe9b9UUTkFNxDlucDeX1QNGP6ralD4hmVGs2Lywq5dtoQr/m2llXzg1dWMykrnpdvnUF4SFDbsXVFVfx1yQ6+On0IEzJs33pztECsoSAivxCRQuCreKihiIgL+C3wo25c61YRWS4iy8vLbX0jc3ISEa45dQhrCg967Qupqmvk1udWEBEaxOPXT+0QTADuOj+HhMhQfvrmepssaTzyS0ARkUUiku/hMRdAVe9T1SzgecBT38i3gX+r6jF7GVX1CVXNU9W85GQbh29OXldNySA0yOWxc76lRfn+y6vZvf8Qf/zKVNLijl6pOC4ihB9eNJplOw/w5tq9fVFk08/4JaCo6vmqOsHD4/VOWV8ArvZwidOAO0VkJ/Ab4Gsi8lAvF9uYfm1QVCgXTRjMqyuLqGts7nDs0fcLWLSxjP++bBzTRyR6vcZ/5GUxISOWX769kUMNtk6Y6SjgmrxEJKfd2znAps55VPWrqjpEVYfh3sf+OVX9cR8V0Zh+67pTs6iqa2J+fgkHaht44fPdXPfEUh5ZVMDVUzP52mlDuzw/yCU8ePl4Sqrq+LMt6WI6CbhOeeAhERmNe9jwLuB2ABHJA25X1Vv8WThj+rMZIxIZmhjJg2+up6auiaYWZURSFN87fxS3nT2iWzPp84YlMHdyOn/5eDtfP32YDSM2bQIuoKiqpyYuVHU5cFQwUdVngWd7t1TGDAwul/Dtc7J54uPtXHNqFnMmpTMuLfa4l2S55cwRvL66mA83l3P1KZm9VFrT3wRcQDHG9K5rTh3CNad6HzrcHePTY0mOCeP9zWUWUEybgOtDMcYEPpdLOHd0Mh9vKaexucVjHlW14cUnGauhGGN65NzRKbyyvIiVuw54HBn2tWe+YHFBBREhQUSEBhEREsSEjFgev/4UW/V4gLKAYozpkTNzkggJEt7fXHZUQFlfXMniggouHj+YrIQIDjc2s6WkhnfXl1JSVedxnovp/yygGGN6JCY8hFOHJfDBpjLuvWRsh2MvfrGb0GAXD109kfjIUABW7j7AVX9awtqiSgsoA5T1oRhjeuzc0SlsKa2h6MChtrRDDU28tqqYyyamtQUTgHFpsQS7hLVFti/LQGUBxRjTY+eOSQHgg81H1sl7c00xNfVNfGV6x5Fk4SFBjEqNYW1RZZ+W0fQdCyjGmB7LTo5iSEIkH2wqa0t74fPd5KREc8rQQUflz82MY92eShv9NUBZQDHG9JiIcN6YFJZsq6CusZn8PZWsKarkK9OHeBzJlZsZz8FDjRTuP+yzMhTuP0RziwWoQGABxRhzQs4ZnUxdYwufbd/Hi1/sJizYxVVTPE92zM1076Oyxkf9KKt2H+Dshz/gpWW7fXI9c2IsoBhjTsiMEYlEhATx9tq9vL66mEtz04iLDPGYd1RqDKHBLtbtOfF+lJYW5cE3N9Ci8NFm2+soENiwYWPMCQkPCeKMkYn8c0URAF+d7n1Zl9BgF2PTYllTeOI1lH+tLGJN4UHS48JZun0fzS1KkMsmTPqT1VCMMSfsnNHu0V6jU2OYOuTozvj2JmXGkb+nkpYT6PeormvkV/M3M2VIPP958Riq6ppYX2yjx/zNAoox5oTNGptCaLCLG08fdsxlVSZmxFHb0Mz2ipoef95j72+loqaeBy8fz+kj3bP0l2zb1+PrGd+wgGKMOWFpcREsvXcW103LOmbeSVnxAD2ej7K9vIZnPt3Bl0/JZFJWPCkx4eSkRFtACQAWUIwxPpEQFdqtRR+zk6OJDA3qcUD52VsbCAsO4kcXj25LO2NkEst27KehyfPKx6ZvWEAxxvSpIJcwIT2uR0uwvLexlA82l/PdWSNJiQlvSz8tO5HDjc2s9kFnv+k5CyjGmD6XmxnH+uIqr3upeFJd18j9r+WTkxLN108f3uHYjOGJiMCSbRW+Lqo5DhZQjDF9bmJmHPVNLRSUHumYr29q5r/mreOttcUez/nV/E2UVNXx6y/lEhrc8U9XXGQIE9LjrB/FzyygGGP63KTM1o55dxNVS4vy/VfW8MLnu/nui6uOCiqfb9/H/y3dzU1nDGeKl2HJp49MZNXuAxxqaOrdwhuvLKAYY/rc0MRIYsODWessFPmztzfw9tq9fP+CUeQNTeDul1azaEMpAHWNzfz41XUMSYjkBxeO8nrN07OTaGxWlu880Fe3YTqxgGKM6XMiQm5mPGuLDvLk4u389dOdfOOMYXznvJE8/fU8xqXH8u0XVvJJQQW/W7SFHRW1PHTVRCJDvS/uceqwQQS7xJq9/Cjgll4RkZ8Bc4EWoAz4uqoe1agqIs3AOuftblWd03elNMacqImZcfzlo23k76ni0olp/Pel4xARYsJDeO6maVz7xFJueW4ZDU0tXHtqFqePTOryepGhwUwZEs9n1jHvN4FYQ3lYVXNVdTLwFvATL/kOq+pk52HBxJh+ZlJmHC0K04cn8Nv/mISr3Tpc8ZGh/P3m6aTHR5ASE869s8d2caUjTstOYt2eSioPN/ZWsU0XAi6gqGpVu7dRgG10YMwAdM7oFP5r9hie+Foe4SFBRx1Pjgnj7e/MZP7dM4mL8Lx6cWdnZCfSou5OfNP3Ai6gAIjIL0SkEPgq3mso4SKyXESWisgVXVzrViff8vJyW+LamEARHhLErWdldxksIkKDOuxLfyyTh8QTHuKyfhQ/8UtAEZFFIpLv4TEXQFXvU9Us4HngTi+XGaKqecBXgEdEJNtTJlV9QlXzVDUvOTm5V+7HGBMYwoKDOCM7iXfXl5zQasamZ/wSUFT1fFWd4OHxeqesLwBXe7lGsfO8HfgQmNKrhTbG9Atzp2Swt7KOz3fs93dRTjoB1+QlIjnt3s4BNnnIM0hEwpzXScAZwIa+KaExJpBdMDaV6LBgXlu1x99FOekEXEABHnKav9YCFwJ3AYhInog85eQZCywXkTXAB8BDqmoBxRhDRGgQF08YzL/X7aWusdnfxTmpBNw8FFX11sS1HLjFeb0EmNiX5TLG9B9XTsngnyuKeG9jGZfmpvm7OCeNQKyhGGPMCVntRpAAACAASURBVJkxIpHU2DDmWbNXn7KAYowZcIJcwtzJGXy4uYz9tQ3+Ls5JwwKKMWZAumJyBk0tytvr9vq7KCcNCyjGmAFpbFoMo1NjbLRXH7KAYowZkESEK6ZksGLXAXbtq/V3cU4KFlCMMQPW3MnpALy2yvMukMa3LKAYYwas9PgIZoxI4OVlu1m12zbe6m0WUIwxA9p3Z+VwuLGZK/+0hJueXca6okp/F2nAsoBijBnQTs9OYvE95/Gji0azYtcBLn/sE259bjn7aur9XbQBxwKKMWbAiw4L5o5zR/LJPefy/QtG8dGWcr72zBe2EZePWUAxxpw0YsJD+O6sHP5ywylsKa3mpmeXcaihyd/FGjAsoBhjTjrnjE7h99dOYdXuA9z29xXUN9kikr5gAcUYc1KaPTGNh67OZXFBBd95YRVNzS3+LlK/ZwHFGHPS+o+8LB64fBwLNpQy+9HFPPvpDioPWb9KT4nqybNNZl5eni5fvtzfxTDGBJg31hTz1OLtrC2qJCzYxWW56cyeOJhhSVFkDoogLDjI30X0KxFZ4Wy53nU+CyjGGOOWv6eSF77Yzeur9lDb4O5XEYHBseGMSo3hrvNzmDpkULev19yiBLmkt4rbZyygeGABxRjTHbX1TWzcW8Xu/YfYvf8QhfsPs7ignLLqeq49NYv/vHgMCVGhXs/fUVHL/a+to6C0htfvPIO0uIg+LL3vWUDxwAKKMaanauqbePS9Ap75ZAfR4cH850VjuHJKBhGhR5rDGppaeHLxdn7/XgFhwS6aW5SJGXG88M0Z/bqmYgHFAwsoxpgTtaW0mvtfy+eLHftxCYxIjmZcWiyjB8fw+uo9bCmt4dKJaTxw+Tg+Lqjgh/9Yww8uGMV3ZuX4u+g91t2AEnB7yhtjTCAblRrDy7fO4KMt5azafZD1xVWs2HWAN9YUkx4XztM35jFrbCoAV0/NYHFBOY+8V8Bp2YnkDUvwc+l7l9VQjDHGByoPNxIREkRocMfZGNV1jVz66Cc0tyj//u5M4iJD/FTCnutuDSXg5qGIyM9EZK2IrBaRBSKS7iXfEOf4RhHZICLD+rakxhhzRFxEyFHBBNzLvTx63RRKq+r48atrqWscuLPyA66GIiKxqlrlvP4uME5Vb/eQ70PgF6q6UESigRZVPdTVta2GYozxl8c/2sZD72zCJTAsMYqc1GhGp8aQkxpDTmo0w5Oijprvoqo0tSghQf797t9v+1Bag4kjCjgq4onIOCBYVRc659T0UfGMMaZHbjtrBDkp0awtqmRLaTWbS6tZuKGUFucvnEtgaGIU8ZEhVB5upOpwI5WHG2lsViJCghgUGUJ8ZCgJUaGMz4hlxohE8oYOIiY8cJrQAq6GAiAivwC+BlQC56pqeafjVwC3AA3AcGAR8GNV7bIuaTUUY0wgqWtsZnt5LVvLa9haWk1BWQ1VdY3ER4QSGxFCfGQIESFBVB1u5MChRg4eaqC8pp6Ne6tobHZPmpyQHkt2SjSRoUFEhgYTERJEWIiLIBFcIohAkEu4YFwqmYMie1TOgK6hiMgiYLCHQ/ep6uuqeh9wn4jcC9wJPNApXzAwE5gC7AZeBr4OPO3hs24FbgUYMmSIr27BGGNOWHhIEOPSYxmXHntc5x1uaGbl7gN8vn0fS7fv5/Pt+znc2MyhhibqGj0vcjkiObrHAaW7ArKG0kpEhgJvq+qETukzgIdU9Rzn/Q3ADFW9o6vrWQ3FGDPQtbQo9U0tKEpzi9Ki7r6YyNBgj4MGuiOgayhdEZEcVS1w3s4BNnnItgwYJCLJTnPYeYBFCmPMSc/lkg6z9/v0s/3yqV17SETyRWQtcCFwF4CI5InIUwBOX8kPgfdEZB0gwJP+KrAxxpgArKGo6tVe0pfj7ohvfb8QyO2rchljjOlaINZQjDHG9EMWUIwxxviEBRRjjDE+YQHFGGOMTwT0PBRfE5FyYFc3sycBFb1YnEBl931ysfs+ufT0voeqavKxMp1UAeV4iMjy7kzkGWjsvk8udt8nl96+b2vyMsYY4xMWUIwxxviEBRTvnvB3AfzE7vvkYvd9cunV+7Y+FGOMMT5hNRRjjDE+YQHFGGOMT1hA6URELhaRzSKyVUR+7O/y9ISIPCMiZSKS3y4tQUQWikiB8zzISRcRedS537UiMrXdOTc6+QtE5MZ26aeIyDrnnEdFRPr2Dj0TkSwR+UBENorIehFpXal6QN+7iISLyBcissa575866cNF5HPnHl4WkVAnPcx5v9U5Pqzdte510jeLyEXt0gP290JEgkRklYi85bwf8PctIjudn8PVIrLcSfP/z7mq2sN5AEHANmAEEAqsAcb5u1w9uI+zgKlAfru0X+PeJhngx8CvnNezgXdwbwEwA/jcSU8AtjvPg5zXg5xjXwCnOee8A1zi73t2ypUGTHVexwBbgHED/d6dskQ7r0OAz537eQW41kl/HPiW8/rbwOPO62uBl53X45yf+TDcW2tvc34nAvr3Avg+8ALwlvN+wN83sBNI6pTm959zq6F0NA3YqqrbVbUBeAmY6+cyHTdV/RjY3yl5LvA35/XfgCvapT+nbkuBeBFJAy4CFqrqflU9ACwELnaOxarqZ+r+yXuu3bX8SlX3qupK53U1sBHIYIDfu1P+GudtiPNQ3BvP/dNJ73zfrf8e/wRmOd9A5wIvqWq9qu4AtuL+nQjY3wsRyQQuBZ5y3gsnwX174fefcwsoHWUAhe3eFzlpA0Gqqu4F9x9eIMVJ93bPXaUXeUgPKE5zxhTc39YH/L07zT6rgTLcfxi2AQdVtcnJ0r6sbffnHK8EEjn+f49A8Ajwn0DrRuqJnBz3rcACEVkhIrc6aX7/OQ+4Dbb8zFM74UAfV+3tno83PWCISDTwL+BuVa3qovl3wNy7uncxnSwi8cA8YKynbM7z8d6fpy+efr9vEbkMKFPVFSJyTmuyh6wD6r4dZ6hqsYikAAtFxNNW6a367OfcaigdFQFZ7d5nAsV+KouvlTpVWZznMifd2z13lZ7pIT0giEgI7mDyvKq+6iSfFPcOoKoHgQ9xt5XHi0jrl8b2ZW27P+d4HO4m0uP99/C3M4A5IrITd3PUebhrLAP9vlHVYue5DPcXiGkEws+5vzuXAumBu8a2HXfHXGsn3Hh/l6uH9zKMjp3yD9Oxw+7XzutL6dhh94WTngDswN1ZN8h5neAcW+bkbe2wm+3v+3XKJbjbex/plD6g7x1IBuKd1xHAYuAy4B907Jz+tvP6Djp2Tr/ivB5Px87p7bg7pgP+9wI4hyOd8gP6voEoIKbd6yXAxYHwc+73H4RAe+AeEbEFdxv0ff4uTw/v4UVgL9CI+9vGzbjbit8DCpzn1h8cAf7o3O86IK/ddW7C3UG5FfhGu/Q8IN855zGcFRf8/QDOxF01Xwusdh6zB/q9A7nAKue+84GfOOkjcI/W2er8kQ1z0sOd91ud4yPaXes+5942025kT6D/XtAxoAzo+3bub43zWN9arkD4ObelV4wxxviE9aEYY4zxCQsoxhhjfMICijHGGJ+wgGKMMcYnLKAYY4zxCQsoZkARkWZnBdY1IrJSRE4/Rv54Efl2N677oYjk+a6k/Z+IPCsiX/J3OUzgsIBiBprDqjpZVScB9wL/c4z88bhXoQ1I7WZ8GxPwLKCYgSwWOADu9b1E5D2n1rJORFpXjX0IyHZqNQ87ef/TybNGRB5qd70vi3vfkS0iMtPJGyQiD4vIMmeviduc9DQR+di5bn5r/vacPS1+5VzzCxEZ6aQ/KyL/KyIfAL9y9rl4zbn+UhHJbXdPf3XKulZErnbSLxSRz5x7/Yezthki8pCIbHDy/sZJ+7JTvjUi8vEx7klE5DHnGm9zZPFBYwBbHNIMPBHOqrvhuPdHOc9JrwOuVPdikUnAUhF5A/cSFRNUdTKAiFyCe6nu6ap6SEQS2l07WFWnichs4AHgfNyrEFSq6qkiEgZ8KiILgKuAd1X1FyISBER6KW+Vc82v4V6H6jInfRRwvqo2i8gfgFWqeoWInId7eZnJwH87nz3RKfsg597ud86tFZF7gO+LyGPAlcAYVVVnEUmAnwAXqeqedmne7mkKMBqYCKQCG4BnuvW/Yk4KFlDMQHO4XXA4DXhORCbgXn7ilyJyFu6lzjNw/1Hs7Hzgr6p6CEBV2+8r07rY5Arca6UBXAjktutLiANycK+F9IyzWOVrqrraS3lfbPf8u3bp/1D3CsLgXlLmaqc874tIoojEOWW9tvUEVT3grMA7DncQAPcaVJ8BVbiD6lNO7eIt57RPgWdF5JV29+ftns4CXnTKVSwi73u5J3OSsoBiBixV/cz5xp6Me02mZOAUVW10VqgN93Ca4H2p7nrnuZkjvzsCfEdV3z3qQu7gdSnwdxF5WFWf81RML69rO5XJ03meyiq4N026zkN5pgGzcAehO4HzVPV2EZnulHO1iEz2dk9OzczWajJeWR+KGbBEZAzuVWP34f6WXeYEk3OBoU62atzbBbdaANwkIpHONdo3eXnyLvAtpyaCiIwSkSgRGep83pPA07i3ZPbkmnbPn3nJ8zHwVef65wAVqlrllPXOdvc7CFgKnNGuPybSKVM0EKeq/wbuxt1khohkq+rnqvoToAL3cuYe78kpx7VOH0sacO4x/m3MScZqKGagae1DAfc37RudfojngTdFZDnuVYg3AajqPhH5VETygXdU9UfOt/TlItIA/Bv4ry4+7ynczV8rxd3GVI67D+Yc4Eci0gjUAF/zcn6YiHyO+8vdUbUKx4PAX0VkLXAIuNFJ/znwR6fszcBPVfVVEfk68KLT/wHuPpVq4HURCXf+Xb7nHHtYRHKctPdwr2C71ss9zcPdJ7UO9wq8H3Xx72JOQrbasDF+4jS75alqhb/LYowvWJOXMcYYn7AaijHGGJ+wGooxxhifsIBijDHGJyygGGOM8Qm/BBRnnaBNzjpB89ot+dA5305nnaLVznDP1vQHRWSPk77amXBljDHGj/zSKS8iFwLvq2qTiPwKQFXv8ZBvJx6GVYrIg0CNqv7meD43KSlJhw0b1tNiG2PMSWnFihUVqpp8rHx+mdioqgvavV0K9MmeCsOGDWP58uXHzmiMMaaNiOzqTr5A6EO5CXjHyzEFFojIChG5tdOxO50ms2ecJSc8EpFbRWS5iCwvLy/3VZmNMcZ00msBRUQWOfssdH7MbZfnPqAJeN7LZc5Q1anAJcAdzmJ7AH8GsnGvR7QX+K23cqjqE6qap6p5ycnHrLEZY4zpoV5r8lLV87s6LiI34t77YZZ66chR1WLnuUxE5gHTgI9VtbTddZ7kyFLcxhhj/MRfo7wuBu4B5rTuO+EhT5SIxLS+xr1HQ77zPq1d1itb040xxviPv1YbfgwIAxY6mwAtdfZlSAeeUtXZuDc/muccDwZeUNX5zvm/dlaEVWAncFsfl98YY0wn/hrlNdJLejHujZBQ1e3AJC/5bui90pnOKmrqeXj+Zu6/bCwx4SH+Lo4xJkAFwigvE+BeX13My8sLWbJtn7+LYowJYBZQzDF9vMU93LqgtNrPJTHGBDILKKZLdY3NfL7DXTPZUlrj59IYYwKZBRTTpRW7DlDX2EJUaBAFZRZQjDHeWUAxXfq4oJyQIOGKKRlsK6+hucV/G7KpKjc9u4z5+SV+K4MxxjsLKKZLi7dUcMrQQUzKiqehqYVd+2r9VpYdFbW8v6mMBRssoBgTiCygGK/Kq+vZsLeKmTnJjEqNAfzbj7Jq90EAtlnTmzEByQKK8erTre5dA87KSSYnJRrw70ivVYUHANhWXos/tl0wxnTNAorx6uOCcgZFhjA+PZaosGAy4iPY4sfaQWsNpaa+iZKqOr+VwxjjmQUU45GqsriggjNzknG5BIBRqdF+q6EcbmhmU0k1eUPdOxUU2BBmYwKOBRTj0aaSasqr6zkrJ6ktbVRqDNvLa2lqbunz8qzbU0lzi/LlvEwAtlo/ijEBxwJKP1K4/xBX/PFTyqp7v7lncYF7dvzMnCN7yOSkxtDQ3MLOfR4XiO5Vq3a7+0/OH5tKXEQIW8stoBgTaCyg9CPLdu5ndeFBPuuDNbUWF1QwKjWawXHhbWmjUv3XMb9q90GGJkaSGB3GyJRoq6EYE4AsoPQjrR3R64urevVz3Mut7O9QOwEY6Yz06uuhw6rKyt0HmJIVD0BOSrQNHTYmAPlrg62HRWSTsyf8PBGJ95IvXkT+6eTdKCKnOekJIrJQRAqcZ697yg8kpZXugJK/p7JXP+eLHftpaGphZrv+E4DI0GCyEiLYUta3NZS9lXWUVdczZYj7v3lkSjT7ahvYX9vQp+UwxnTNXzWUhcAEVc0FtgD3esn3e2C+qo7BvTfKRif9x8B7qpoDvOe8H/Da11B6cx7Goo2lhAa5mD488ahjo1JiPDZ5HahtYPnO/WwormLXvlrKq+upb2ru8nNUlZZuLOXSOlx4yhD3945sp6ZkzV7GBBZ/bbC1oN3bpcCXOucRkVjgLODrzjkNQOtX0rnAOc7rvwEf4t5SeEArqaoHoPJwI3sOHiZzUKTPP6PycCP/WlHEZblpRIQGHXU8JzWGjwvKaWxuISTI/X1EVfnGs8tYXXiwQ97wEBdfP304t589gvjI0Lb0+qZm/m/pbv74wVYuy03j/82d0GWZVu0+QFiwizGDYwEYmXwkoEwbnnBC92uM8R1/bQHc3k3Ayx7SRwDlwF9FZBKwArhLVWuBVFXdC6Cqe0UkxdvFReRW4FaAIUOG+Lrsfaqk8jA5KdEUlNWQv6eqVwLKi1/sprahmVtmjvB4fFRqNI3Nys6KWnKc5ViWbNvH6sKD3HFuNhMz4jnU0ERtfRPLdx3gLx9v4/nPd3HbWSO48fRhLFhfyv8u3MKeg4eJDQ/m9dXF/Pdl49qCkyerCg8yMSOO0GB3noz4CCJCgrpVQ6lvauaXb2/klpkjyErw/b+XMeaIXmvyEpFFIpLv4TG3XZ77gCbgeQ+XCAamAn9W1SlALT1o2lLVJ1Q1T1XzkpOTj31CgGpqbqG8up5zRicT5BI2FPu+H6WhqYVnP93JGSMTGZce6zGPpzW9/vjBVlJiwvjurBwunjCYq6ZmcsNpw/j9tVN4566ZTB+eyG8WbGHK/1vID/6xhoSoUP7v5un8+ku5VB5uZNnO/V2Wad2eSiZnHelmc7mE7JSobg0d/nBzOX/7bBcvLdvd3X8GY0wP9VoNRVXP7+q4iNwIXAbMUs8dAkVAkap+7rz/J0cCSqmIpDm1kzSgzFflDlQVNQ20KAxNjCI7OapXRnq9va6Ykqo6/ueqiV7zZCdHIwJbSqu5lDRWFx5kybZ9/NfsMYQFH91ENmZwLE/dmMfK3Qf4x/JCzhiZxOwJabhcQm19E6HBLhZtKOP07CQPnwYb91bR0NTS1iHfamRyNMt2HjjmPbUudf9JQQU/uuiY2Y0xJ8Bfo7wuxt3nMUdVPc6SU9USoFBERjtJs4ANzus3gBud1zcCr/dicQNCa4f84NhwxqfHke/jGoqq8uTHOxiZEs3Zo7zX5CJCgxiSEEmBM9LrTx9sJS4ihK9MH9rl9acOGcT/XJXLZbnpbUu5RIUFc0Z2Igs3lngdZNA6obG1Q77VyJRo9hw8TG19k9fPbGhqaRtgsHZPJQcP2agwY3qTv0Z5PQbEAAtFZLWIPA4gIuki8u92+b4DPC8ia4HJwC+d9IeAC0SkALjAeT+glThDhgfHhTM+PZbSqnoqaup9dv3Ptu1jw94qbjlzeNsffG9yUmLYUlpDQWk1CzaUcuPpw4gO61ll94Jxgyncf5jNXiZLri48SGpsGGntJljCkTkx27po9lqyrYLquia+edZwVN19PcaY3uOXgKKqI1U1S1UnO4/bnfRiVZ3dLt9qp/8jV1WvUNUDTvo+VZ2lqjnOs/dG+AGi1KmhpDo1FPDtBMcnF28nKTqUK6ZkHDPvqNRodlbU8uj7W4kICeIbpw/r8eeeP9Y9nmLRhlKPx1cVHmRK1iBEOga5kd0YOjw/v4TosGDuOHckMWHBLC6o6HE5jTHHZjPl+4m9lXWEBAmJUaFtHea+muC4tayaDzaXc8OMYYSHHN0P0tmo1BiaWpQ31xRz3bQhDIoKPeY53qTEhjMpK56FHgLKvpp6du07dFRzF7j7koJd4jWgNDW3sGBDKeeNSSEyNJgZ2Yl8srW8x+U0xhxbIAwbNt1QWlVHSkw4LpcQFxFCVkIEG3pYQ3lvYyklVXUcbmimrrGZxQUVhAW7uH5G94ZV5zhreoUECd88a3iPytDeheNSefjdzZRW1ZEae6Rp65XlRQDkDTt6IYSQIBfDkqK8BpQvdu5nf20Dl0wYDMCZI5NYuKGUXftqGZoY1WV5WlqUjwrKOWdU8lE1I2OMdxZQ+omSyroOCzVOSI9jfQ865vP3VHLz35Z3SAtyCbeeNYLE6LBuXSM7OZrwEBdXTM4gLS7iuMvQ2QVOQFm4oZTrZ7g79zcUV/G/Czdz0fhUpg7xvLLOyORotnjpe5mfX0J4iIuzR7sHGJzpLCOzuKDimAHlvU1lfPO55Tx30zTO6mKAgjGmIwso/URpVR1j047MDRmfHss7+SVU1TUSGx7S7eu09iPMv3smaXHuCYIhQXJc38TDQ4J4884zfTZRMCclmiEJkSza6A4odY3N3P3yKuIjQ/mfq3K9lm1kSjQLN5bS0NTSNukR3DWMd9eXcPaoZCJD3T/iI5KiSI8L55OCirag5c1KZ2TZmsKDFlCMOQ7Wh9IPqColnZqDWjvmNx5ns9eSbe5l6ccMjiUuIoTQYFePmnVyUmO61d/SHSLCBeNSWbJ1HzX1TTz87ma2lNbw6y/lktBF/8zIlGiaW5Sd+2o7pK8qPEhpVT2XTEjr8Bln5iSxZFsFzcdYP2yNs4TMul5ehNOYgcYCSj9QVdfEoYZmBscdaZIan+GurRzPSK+6xmaW7dzvdRKhP10wLpWG5hZ+9c4mnv5kBzfMGMq5o72uqAN4H+k1P38vIUHCeWM7nn9mTjJVdU2sLeq45lh7LS3K2iJ3ILGAYszxsYDSD7QOGR7crr8iJSac5Jiw45rguHL3AeoaWzhzZOAFlLyhg4iPDOHvS3cxIjmK/5o99pjntM7abx9QVJV38ks4c2TSUU2BZ2S7V0/+pIvhw9sraqipb2LM4Bj2VtZRXu27uT7GDHQWUPqBtkmNsR0n941Pjz2ukV5Ltu4jyCVMHxF4K/QGB7mYNSaVYJfwyDWTPa503FlEaBAZ8RF8vmMf8/NLeHVlEX/8YCtFBw53aO5qlRgdxvj0WBZv9R5QVhe6A/RXnX6W3t57xpiBxDrl+4H2y660NyE9jsUFFdQ1NnerP+PTbRVMyowj5jg68fvSfZeO5eunD2NiZly3zxmfHsu760v5dOuRWfAx4cGcPy7VY/4zc5J45pMd1NY3EeVhdv+awoNEhwUzZ1I6P3k9n3V7Kjl3TNdNb8YYNwso/UDrTo0psR2H9Y5Pj6W5RdlcUs2kLI+bXrapqmtkTeFB7jh3ZK+V80QlRIV22Qnvya+uzuWWmTVEhQYTFRZEZGgwsRHBHheqBJg5Mpm/fLSdz3fs47wxRwedNUXupfLjIkIYnhTV1p9ijDk2a/LqB0qq6hgUGXJULaR1pFd3Oo8/376fFiUgO+RPRHxkKKcOS2BceixDE6NIjgnzGkzAPUkyLNjFx1uObvaqa2xm496qtuCcmxFnTV7GHAcLKP2Ae1Lj0RMIsxIiGJEUxTOf7KChqaXLa3y6tYLwEBdTh3ZdkxnowkOCmJmTxDv5e48aPrxxbxWNzcrkLHegnpARR0lVHWXVdf4oqjH9jgWUfqCkqo7BsUfPYhcR/vvycWyvqOXpT3Z0eY1Pt1Zw6rCELr+9nyyunJJJaVU9S7Z1rKW0zj9pq6Fkup+tlmJM9/hrP5SHRWSTiKwVkXki4vFrs4jEi8g/nbwbReQ0J/1BEdnjLH2/WkRmezp/oCit6rjsSnvnjk7hgnGp/OH9AvZWHvaYp6yqjoKymoAcLuwPs8amEBMezKsr93RIX1NUSUpMWNvgh/HpsYhg/SjGdJO/aigLgQmqmgtsAe71ku/3wHxVHQNMAja2O/a7dsvf/9vz6f1fQ1MLFTUNHWbJd/aTy8bR3KL8/O2NHo+37gNyhgUUwN3sdVluOvPzSzps0LWm8CCTsuLbVg6ICgsmOznaaijGdJO/9kNZoKqtv8lLgczOeUQkFjgLeNo5p0FVvU9x7mf2Vh7m4kc+5o8fbO1yKZDW9vvOQ4bby0qI5FvnZPP22r186mGOxSdbK4iPDGFcmud94k9GV0/N4HBjc9sWwZWHGtleUdth73pwd8xbDcWY7gmEPpSbgHc8pI8AyoG/isgqEXlKRNovE3un02T2jIh4Xo4WEJFbRWS5iCwvL+/Zfhi19U1tW9H6ymPvb2VTSTUPv7uZa5/4jML9HndCPrKxlpcmr1a3n51NVkIED7yxvkMHvaqyZGsFp2cnHnMnxpPJKUMHMSQhkldXuZfIX7vH6T/J7BhQJmTEUVZd3/b/YIzxrtcCiogsEpF8D4+57fLcBzQBz3u4RDAwFfizqk4BaoEfO8f+DGTj3hZ4L/Bbb+VQ1SecXR/zkpN7tnLs/a/lc9Ozy6hrbO5W/qbmFr7/ymq+9/Jqj3ulF+4/xCvLC7l+xhB+d80kNu2tZvbvFzNvVdFR+fc6c1A6b4HbWXhIEA9cNp6tZTXc/9o6Xl+9hyVbK1hcUEFxZZ01d3UiIlw1NYMl2/ZRfPBwW4d850mVuc77dVZLMeaYem1io6qe39VxEbkRuAyYpZ7+6kIRUKSqnzvv/4kTUFS1bXs/EXkSeMsnhfbi6qmZzFu1h3fXlzB3ctdb5Koq//36+rYO3xkjErjm1I4bVz32/lZEhDvOHUlaXAR5QxOcALSG3fsOc9f5OW15gNJuTAAAIABJREFUvS274sn541KZOzmdV5YXtW1O1eqMATb/xBeumpLJI4sKeG31HlYXVjIiOYq4iI6rCIxLj8UlsHZPpdfZ98YYN7/MlBeRi4F7gLNV1WNbj6qWiEihiIxW1c3ALGCDc36aqu51sl4J5PdmeU/PTmRIQiQvfrH7mAHlD+9v5cUvdnP72dmsKTzIz9/ayMycZNLj3fNIdu2r5Z8ri7hhxtC2zamyEiJ56dbTuO3vK3j6k+3cetaItrWsSqvqCAt2HfWHzptHrpnM/5s7gfLqevejpp7wYPfuhqajIYmR5A0dxKsr91B5uJGZHmpxkaHBjEyxjnljusNffSiPATHAQmfY7+MAIpIuIu1HbH0HeF5E1uJu3vqlk/5rEVnnpJ8LfK83C+tyCddOy2Lp9v1sL/e85SzAK8sK+d+FW7hqagb3XDyaX38pl2ZVfvzquramrEff20qwS/j2Odkdzg1yCd+cOZyquibeXFPcll5SVc/guPBu71ki4t4ieGRKNKdlJzJnUjoXjh/cg7s+OVw1NZOtZTWUV9d7Xb5mgtMx77kibYxp5a9RXiNVNavdsN/bnfRiVZ3dLt9qp/8jV1WvUNUDTvoNqjrRSZ/TrrbSa750SibBLuGlZYUej3+wqYx7561jZk4Sv7ravctgVkIkP75kDB9vKeeV5YVsL69h3ip37STFQxPWtOEJjEqN5u9Ld7WllVbWdTlk2JyYSyemte326C2g5GbEUVFTT2mV96Xsq+oaO3wR8ORPH27l9dV7usxjTH8WCKO8+oWUmHAuGJfKP1cUUd/UsXN+a1k1d7ywkjGDY/jz9acQEnTkn/X66UOZMSKBn7+1kQff3EBYcBC3d6qdtBIRrp8xlHV7Kts6ifdWHe5W/4npmbjIEC4Ym0posIuxaTEe80x0Rn51tTHXo4sK+M6Lq47a7KtVS4vyx/e38vznu0+80MYEKAsox+G6aUPYX9vAgvVtYwI43NDMt59fSXhIEE/feCrRnZZEd7mEX189iWZVPt5Szo2nDyMp+uhlVFpdOSWDyNAg/m/pLlSV0qr6Y47wMifmgTnj+L+bp3tdlmZcWixBLuHzHfs9Hq9vauZfK92DINbt8Rx0tlfUUtvQzM6KWo/HjRkILKAchzNHJpE5KIIXvzjyLfOBN/IpKKvhkWsme10eZUhiJD+dM55RqdHcetaILj8jJjyEK6Zk8MaaYnbuO0RDU4s1efWylJhwpg33vulYRGgQF48fzD+WF1LTbmZ9qwXrSzlwqBHwvkxLa6d+WXV9h9n5xgwkFlCOg8slXHtqFku27WNnRS3/WuEennvHOSM5a1TXc1y+nJfFgu+d3a39Pq6fPpT6phb++MFWAK+ByvSdW5wBE6946EN78YvdZA6KYMqQeK/zVdpvMbDDailmgLKAcpy+nJdFkEt46J1N3P9aPtOHJ3B3u3kjvjAuPZapQ+J51WlGsRqK/00ZMohThw3imU930NR8ZCWCXftqWbJtH9fkZTEpM571xVUdjrdaV1RJbLi7OXTnPgsoZmCygHKcUmPDmTUmhfnrS4gMDeLR66YQHOT7f8brZwyldYkvq6EEhm/OHEHRgcPMX1/SlvbSskJc4v6iMSkrjsONzWwr7xgwWlqU9cWVXOQM395RbgHFDEwWUHrg5jOHMygyhEeundxrtYfZE9MYFBmCCKTEeO/EN33n/LGpDE+K4smPt6OqNDa38I/lRZw3JoXBceFMzPA8Gqy1Q37a8ATS4sLZYTUUM0BZQOmB6SMSWX7/BczM6dnaYN0RHhLEbWdnM21YQodhyMZ/XC7h5jOHs6aokmU7D/DexjIqauq51llaZ0RSFFGhQUdtydw68mtiZhzDEqNspJcZsOwvVQ8F9cHKvbefnc3Lt53W659juu/qqZkkRIXyxMfbeWnZblJjwzhntPuLhcslTMiIY02njvl1RVWEh7gYmRzN8OQo65Q3A5YFFGOOQ0RoENfPGMqijaV8tKWc/8jL6tCHlpsZx8a9VR22EMjfU8m4tFiCg1wMT4ziwKFGDh5q8EfxjelVFlCMOU5fO21o23It/5GX1eHYxMx4Gppa2FJaDUBzi5JfXMnEDPcy+K2LdFotxQxEFlCMOU5J0WF897yR3HjaMLISIjscy3UCR2s/yo6KGg41NLct3zLcCSg2dNgMRH5Zvt6Y/u7O8zzPPRqaGElMeDBriyq5btqRwNJaQxmSEIlLYEeF5x06jenP/FJDEZGHRWSTs4XvPBE5aplXERntLG3f+qgSkbudYwkislBECpxnr1sAG9OXRITczLi2kV2tHfLZye6aSWiwi8xBkdbkZQYkfzV5LQQmqGousAW4t3MGVd3curw9cApwCJjnHP4x8J6q5gDvcWRrYGP8Ljczns0l1dQ1NrNuz0HGp8d16LgflmRDh83A5K/9UBaoausKeUuBzGOcMgvYpqqtG4XMBf7mvP4bcIXvS2lMz+RmxNHYrGzYW8X64qq25q5WwxPdNRTbsMsMNIHQKX8T8M4x8lwLvNjufWrrplrOc0ovlc2Y4zYx0x1A3lhdzKGGZiZ0DihJUdTUN1FRY0OHzcDSawFFRBaJSL6Hx9x2ee4DmoDnu7hOKDAH+EcPy3GriCwXkeXl5eU9uYQxxyUjPoKEqNC2PVJyMzsGlGE20ssMUL02yktVz+/quIjcCFwGzNKu6/6XACtVtbRdWqmIpKnqXhFJA8q6KMcTwBMAeXl51sZgep2IMDEjjo+2lBMREkR2cnSH461Dh3eU13LqMO/7sBjT3/hrlNfFwD3AHFU91vjJ6+jY3AXwBnCj8/pG4HXfltCYE9NaKxmXHnvUMj0Z8RGEBIktEmkGHH/1oTwGxAALnSHBjwOISLqI/Ls1k4hEAhcAr3Y6/yHgAhEpcI4/1DfFNqZ7WjviO3fIAwQHuchKiLRl7M2A45eJjao60kt6MTC73ftDQKKHfPtwj/wyJiBNHTqIqNAgZuYkeTw+IinK+lDMgGMz5Y3pBUnRYax+4EKvWw8MS4zik60VtLQorj5YudqYvhAIw4aNGZC62sdmWFIUdY0tlFTVHdc1X1u1h6c/2XGiRTOmV1hAMcYPRrQOHT6OGfPNLcov/72RR98rsEmRJiBZQDHGD1rnomw/joDy2bZ9lFXXU3m4kV37bHFJE3i6DCgiEioi0u79uSLyAxG5pPeLZszANTg2nLBg13HVUOat2kNrd8uaTvvWGxMIjlVDWQbEA4jIj4BfABHA90Xkf3q5bMYMWC6XMPw4Rnodbmhmfv5erpiSQXiIizWFlcc+yZg+dqyAEqSqB5zX1+Ce1f5z3LPXL+3VkhkzwA1LjGJbefcWiVy0sZTahma+dEom49PjWGs1FBOAjhVQqkRkgvO6Agh3Xgd341xjTBemDo1nR0Utt/19Bftq6rvM+9qqPaTFhTNjeCK5mXHkF1fS1NzS5TnG9LVjBYXbgedF5Dnc62UtF5FngE+AX/Z24YwZyG45cwT3XzqWDzeXc9EjH/PexlKP+fbXNvDRlnLmTErH5RImZcZT19hCQVlNH5fYmK51GVBUdS0wFfdaWiuAPwHvAuer6gu9XzxjBi6XS7hl5gje+M4ZJMeEc/PflnPvq2upa2zukO/ttcU0tShXTMkAYFKWe4NTa/YygeaYzVaq2qyq7+BesfctVX1ZVe0n2RgfGTM4ltfuOJ3bz87mpWWFXPPEUsqrjzSBzVu1hzGDYxibFgvAsMRIYsODWW0d8ybAdKsfREQuB1YD8533k0Xkjd4smDEnk7DgIH58yRgev/4UtpRUc8UfP2VLaTW79tWycvfBttoJtO5bH281FBNwutux/iAwDTgIoKqrgWG9UyRjTl4XjR/MK7edRkNzC1f/aQk/e2sjIjBnUnqHfLmZcW371hsTKLobUJpU1erXxvSBiZlxvHbHGWQMimDRxlKmD08gPT6iQ57czHiaWtz71hsTKLobUPJF5CtAkIjkiMgfgCU9/VAReVhENonIWhGZJyLxHvKMdvZKaX1UicjdzrEHRWRPu2Ozj/4UY/qvjPgI/nH7aXz99GH88MLRRx2f3NoxX2jNXiZwdDegfAcYD9TjHvFVBdx9Ap+7EJigqrnAFuDezhlUdbOqTlbVycApwCFgXrssv2s9rqr/7ny+Mf1dTHgID84ZT56HbYIHx4WTEhPG2iJrODCBo1v7oTgbXd3nPE6Yqi5o93Yp8KVjnDIL2Kaqu3zx+cYMBLmZ8ay2jnkTQLo7yivZaab6t4i83/rwURluAt45Rp5rOXpf+TudJrNnRGSQtxNF5FYRWS4iy8vLy0+0rMYEjEmZcWwvr6WqrtHfRTEG6H6T1/PAJmA48FNgJ+6FI70SkUUiku/hMbddnvuAJuf63q4TCswB/tEu+c9ANjAZ2Av81tv5qvqEquapal5ycvIxbtOY/iPX6UfJt2YvEyC6uwVwoqo+LSJ3qepHwEci8lFXJ6jq+V0dF5EbgctwLzjZ1ep4lwArVbVtXYr2r0XkSeCt7tyEMQPJpMw4ANYUVXL6SM971xvTl7obUFrr1HtF5FKgGMjs6YeKyMXAPcDZTv9MV66jU3OXiKSp6l7n7ZVAfk/LYkx/FR8ZytDESJvgaAJGdwPKz0UkDvgB8AcgFvjeCXzuY0AYsNDZv2upqt4uIunAU6o6G0BEIoELgNs6nf9rEZkMKO7mt87HjTkp5GbGs2Lnfn8Xwxig+6O8WpuUKoFzT/RDVXWkl/RiYHa794eARA/5bjjRMhgzEEzOiufNNcW8vXYvl+ameczT3KKoKsFBtuOE6V3dCigiMhz3XJRh7c9R1Tm9UyxjTHf8R14m8/P3cueLKzl4eAJfnT60w/EPNpVxz7/WMnXIIB6/4RQ/ldKcLLrb5PUa8DTwJmC7+hgTIGLCQ3jupunc8cJK7puXz/6aBu48byS1Dc38/K0NvLSskPjIEOavL2FxQTkzc2yko+k93Q0odar6aK+WxBjTIxGhQfzlhlO4519r+e3CLWyvqGXZzv0UHzzMt87J5tvnZDP70cX84u2NvP3dJIJc4u8imwGqu42qvxeRB0TkNBGZ2vro1ZIZY7otJMjFb740iZvPHM68VXsIcgn/uP007rl4DDHhIdx7yVg2lVTzyvJCfxfVDGDdraFMBG4AzuNIk5c6740xAcDlEu6/dCyzJw5mbFoskaFHfr0vmTCYvKGD+O2CzVw+KZ3osO7+6hvTfd2toVwJjFDVs1X1XOdhwcSYACMinDI0oUMwaU2//7JxVNQ08OcPt/qpdGag625AWQMctcS8Mab/mJwVzxWT03ly8Q6KDhxrPrExx6+7ASUV2CQi74rIG62P3iyYMcb3fnTxGAT49fzN/i6KGYC625D6QK+WwhjTJzLiI7j5zOH86cNt3HV+DtnJ0d06r6y6joOHGhmVGtPLJTT9WbdqKKr6kadHbxfOGON7N505nLBgF08t3t6t/C0tyjf+uowLf/cxX3vmCz7fvq+XS2j6qy4DiohUO1vvdn5Ui4htZm1MP5QUHcbVp2Tyr5V7KK+uP2b+f+fvZX1xFXMmpbOhuJJrnljKlx9fwqdbK/qgtKY/6TKgqGqMqsZ6eMSoamxfFdIY41s3nzmcxuYW/v7Zzi7zNTW38L8LtjA6NYbfXTP5/7N35/Fx1fXi/1/vyb4vzdKkSbqmG6UtJbSsZWnZkargFTcWRURFUNSrfLmC/lwuV/Qiil4uoiJXFgFBENnaspRFStN9T9M9TdokbfY9mffvjzmTZplptklmmryfj8c8MnPO55x8DiR957O9P7z3vYv40dWncKiqiS/8YTV7KxtGpL7m5GDZ4owZg6amx7N0ViaPf7ifxtZ2v+X+tq6EPZUNfPuS6YS5hOiIMG44exLP3HoWboUV2474vdaMPUELKM6WwjucbXxfEBGf05JF5FsistXZ7fEpEYl2jk8WkdUisktE/urs7GiM6adbFk+hurGN59aW+Dzf3NbBr1bsYn5uMhfPzux2LicllpnjE1i+3QKKOS6YLZTlwBxVnQsUAXf1LCAiE4DbgQJVnQOE4dlfHuC/gAdUNR+oAr40IrU2ZpQomJjCaXnJPPruXjrcvTdNfWL1Acpqmvn3S2fg7FvUzdJZmazdX0V1Y+tIVNecBIIWUFT1DVX1trU/xP8OkOFAjIiEA7FAqXh+ui8CnnPK/Bn4+HDW15jRRkS45bwpHDjWyOtbD3c7V9/Szm/fKubcaWl+txdeOjuTDrfy9s6KkaiuOQmEyhjKF4FXex5U1UPAL4ADQBlQo6pv4Nl0q7pLQCoBJvi6sYjcIiKFIlJYUWE/+MZ0dckp48lLjeV/V+3haH0Lh2uaOXiskYfeLOZYQyvfuXSG32vnTkgiPSHKur1Mp2HNECciK4DxPk7draovOmXuBtqBJ3xcnwIsAyYD1cCzIvJ54HUf9+zdZgdU9RHgEYCCggKfZYwZq8Jcws3nTeaeF7dy+k9WdDt3yexM5uf6z7jkcglLZmbwz01ltLa7iQwPlb9PTbAMa0BR1aUnOi8iNwBXAUtU1dc/9kuBvapa4ZR/HjgbT/BJFpFwp5WSA5QGtPLGjBHXnZGHS4QOtxIR5iIy3EVUuIulszL7vHbJrEyeXnOQNfuOcY6frjEzdgQth7WIXAZ8Dzjf2TvelwPAmSISCzQBS4BCVVUReQu4FngauAF4cQSqbcyoExnu4vNnTuy7oA/nTksjKtzF8m1HLKCYoI6hPAQkAMtFZIOIPAwgItki8gqAqq7GM/C+DtiMp76PONd/D7hTRIrxjKn8YYTrb8yYFxMZxrnT0li54wi+OxnMWBK0FoqqTvNzvBS4osvne/GRnFJV9wALh62Cxph+WTIrk5U7yik6Us+M8ZY8ciyzUTRjzJAsmZUBwAqb7TXmWUAxxgxJZmI0c3OSWGkBZcyzgGKMGbIlMzNZf7Cayvq+sxeb0csCijFmyJbOzkAVXt1cFuyqmCCygGKMGbLZWYnMy03m9+/upb3DPej7lNU00dzWEcCamZFkAcUYM2Qiwm0XTuPAsUb+sWlwa4yb2zq49IFVPLCiKMC1MyPFAooxJiCWzMxg5vgEfvvWbtw+shf35f3iSmqb2/mg2LYYPllZQDHGBITLJXztwmkUl9f3yl7cH8udzbq2ldXS0OJ/0y8TuiygGGMC5spTs5icFsdDbxUPaOW8262s2F5OekIUHW5l48HqYaylGS4WUIwxARPmEr56/lS2ltbydtHx7SKa2zr47+VFfOuvG3x2h20s8Uw5/sZF0xCBNfuqRrLaJkAsoBhjAurjp00gOyma377paaWsKqrg0l+t4tcrd/HC+kO8tbO81zXLtx0hzCVcPS+bGZkJFO4/FoSam6GygGKMCajIcBdfOX8qhfur+Nyjq7n+jx8RJsLjX1xIdlI0v393T69rVmw/wsJJqSTHRnL6xBTWH6j2uS2xCW0WUIwxAffpM3JJT4iicH8V31o6nVe/eR6Lp6dz4zmT+HDPMbYcquksu/9oA0VH6lk627P/SsGkFOpb2tl5uC5Y1TeDZAHFGBNw0RFh/O3Ws3nz2+dzx9J8osLDALhuYR5xkWE82qWV4p3ddbGzoVfBxFQA1lq310knKAFFRO4XkR0isklEXhARn/uMisi3RGSriGwRkadEJNo5/piI7HX2UdkgIvNH9gmMMX3JGxdLTkpst2OJ0RF8+ow8Xt5URllNE+Dp7pqRmUDeOE/ZnJQYMhKibGD+JBSsFspyYI6qzgWKgLt6FhCRCcDtQIGqzgHCgOu6FPmuqs53XhtGotLGmKG76ZxJuFV57IN9VDe2smZfFUtnZ3SeFxEKJqWwdr8FlJNNUAKKqr7h7AUP8CGePeF9CQdiRCQciMX2jTfmpJebGsvlp2bx5OoD/GNTGR1u7bV/fcHEVA5VN3W2YszJIRTGUL4IvNrzoKoeAn6BZ1/5MqBGVd/oUuSnTpfZAyIS5e/mInKLiBSKSGFFRYW/YsaYEXTzuZOpa27nvle2k54Qxbyc7r3eBZNSACi0bq+TyrAFFBFZ4Yx99Hwt61LmbqAdeMLH9SnAMmAykA3EicjnndN3ATOBM4BUPPvL+6Sqj6hqgaoWpKenB+z5jDGDd1peCgUTU2ho7WDprAxcLul2flZWIjERYdbtdZIZtoCiqktVdY6P14sAInIDcBXwOfWdo2EpsFdVK1S1DXgeONu5d5l6tAB/wvaWN+akc8viKQBcNier17mIMBfzc5NtgeNJJlizvC7D06q4WlUb/RQ7AJwpIrEiIsASYLtzfZbzVYCPA1uGv9bGmEC65JTxrLjzfM6f7rvnoGBSCttKa6m3RJEnjWCNoTwEJADLnWm/DwOISLaIvAKgqquB54B1wGanro841z8hIpud42nAT0a4/saYAJiWEe/33OkTU3ArbDhgiSJPFuHB+KaqOs3P8VLgii6f7wXu9VHuouGrnTEmFCyYmIIIFO4/xrn5acGujumHUJjlZYwxvSRGRzAjM4EV249Q3dga7OqYfrCAYowJWdefNYltpbVc+Iu3eWL1fksYGeIsoBhjQtZnF+Xxz9vPIz8zgbtf2MKy375nU4lDmAUUY0xIm5WVyF9vOZMHr5tPRV0Ln3r4A1Y4CSVNaLGAYowJeSLCsvkTWPntC5gzIYnbn17fLQW+CQ0WUIwxJ434qHAevb6A5JgIvvTnNZbrK8RYQDHGnFQyEqP5401n0NDSwZceK7SFjyHEAoox5qQzc3wiD332NHYeqeP2p9bT3uEOdpUMFlCMMSepC2Zk8MOrT+HNHeVc+ev3eGVzGW4f04r3H22w8ZYREpSV8sYYEwhfOHMiKbERPLC8iK89sY4ZmQncsTSfzMRoVmw/woptR9hVXo8IPHjdaVw9LzvYVR7VxHei39GpoKBACwsLg10NY0yAdbiVlzeV8uuVu9hd0QBAuEtYODmVpbMyeW3rYdbtr+Lhz5/O0tmZfdzN9CQia1W1oM9yFlCMMaNFh1tZvu0IbR1uFk9PJykmAoC65jY+/+hqth+u4083nsE50yw32ED0N6DYGIoxZtQIcwmXzRnPx+ZldwYTgIToCP78xYVMSYvj5j8XUrjP9lkZDsHaD+V+EdnhbOH7gogk+yl3h7PL41YR+WaX46kislxEdjlfU0au9saYk1FybCT/96VFZCVFc9Of1rDrSF2wqzTqBKuFshyYo6pzgSI8W/p2IyJzgC/j2Y1xHnCViOQ7p78PrFTVfGCl89kYY04oPSGKv9y8iMhwF7c/vYGW9o5gV2lUCUpAUdU3VNW7GulDIMdHsVnAh6ra6JR9B/iEc24Z8Gfn/Z/x7NpojDF9yk6O4efXzmV7WS2/eH1nsKszqoTCGMoXgVd9HN8CLBaRcSISi2fjrVznXKaqloFnf3kgw9/NReQWESkUkcKKiooAV90YczJaMiuTL5w5kd+/u5f3dlUGuzqjxrAFFBFZ4Yx/9Hwt61LmbqAdeKLn9aq6HfgvPN1jrwEbnbIDoqqPqGqBqhakp/veu9oYM/b8vytmMS0jnjuf2UBVg23gFQjDFlBUdamqzvHxehFARG4ArgI+p37mLqvqH1R1gaouBo4Bu5xTR0Qky7lPFlA+XM9hjBmdYiLDePC6+VQ1tvL95zcxlpZQDJdgzfK6DPgecLWqNp6gXIbzNQ/4JPCUc+ol4Abn/Q3Ai8NXW2PMaHVKdhLfvXQGr289wn2v7aC5zQbphyJYYygPAQnAchHZICIPA4hItoi80qXc30RkG/AP4Ouq6t2q7T7gYhHZBVzsfDbGmAG7+dwpXLMgh/99Zw+XPLCKt3ZYh8dg2Up5Y4wB3i+u5J4Xt7C7ooGLZ2dyz1WzyU2NDXa1QoKtlDfGmAE4Z1oar96xmO9dNpP3dlVy2a9W8be1JTa2MgAWUIwxxhEZ7uKrF0xl+Z2LOWVCEt9+diPf/OsGapvbgl21k4IFFGOM6SEnJZanvnwm3754Oi9vKuOKB99l7f6qvi8c4yygGGOMD2Eu4RtL8nnmK2cB8KmHP+A/X9lOU+uJZ4I1tLTzxtbD3PX8Ju54ev2YmjlmG2wZY8wJnD4xhVfuOI//fGU7/7tqD69vPcx/fnIuZ00dB4CqsruinneKKnl7Zzmr9xyjtcNNfFQ49S3thInwy3+bh4gE+UmGnwUUY4zpQ2J0BP/5ybl8bF42dz2/mc/8/kOuPT2HMBHe3VVBaU0zAFPS47j+rIlcNDODgkmp/M/bu3lgRRGzsxO5+bwpQX6K4WcBxRhj+unsqWm8dsdiHlhRxKPv7iEuKpxzp6Vx20XpnJef1mua8Tcumsb2slp+9sp2ZoxP4Lz80Z3+ydahGGPMINQ0thEXFUZ42ImHohta2vnk7z7gcG0zL912DhPHxY1QDQPHtgD2wQKKMSYY9h9t4OqH3ic9IYqPz8/G5RJcIoS7hNS4SHJTY8lLjSU9PgqXK/TGWvobUKzLyxhjhtnEcXH89rMLuPUva/nFG0V+y0WGu8hIiCI+KpzYyDDiosKJCg+jpb2DljY3zc7XuKgwUuMiSY6NJDUukszEaCaNi2XiuDhyU2OICg8bwac7zlooxhgzQtxupd2tuNXzancr5bUtHKxqpORYIwermqioa6GhpZ3G1g4aWttpbnMTFe4iOsJFTEQYkeEuGlo6ONbQSlVjK8caWmlpd3d+D5dAalwULgEREDwtnv/+9DzOnpo2qHpbC8UYY0KMyyVE9ujSSoyOYFpG/KDvqapUNbaxt7KB/Ucb2He0kYq6ZlTB215QlNS4yKFUvV8soBhjzElMxDMOkxoXyekTU4Jal2Dth3K/iOwQkU0i8oKIJPspd4ezy+NWEflml+M/FJFDTur7DSJyxcjV3hhjjC/BSr2yHJijqnOBIuCungVEZA7wZWAhMA+4SkTyuxR5QFXnO69Xel5vjDFmZAUloKjqG6rpaawLAAAgAElEQVTq3R/+QyDHR7FZwIeq2uiUfQf4xEjV0RhjzMCEQnLILwKv+ji+BVgsIuNEJBa4Asjtcv42p8vsjyIS3I5DY4wxwxdQRGSFM/7R87WsS5m7gXbgiZ7Xq+p24L/wdI+9Bmx0ygL8DzAVmA+UAb88QT1uEZFCESmsqKgI1OMZY4zpIWjrUETkBuBWYImqNvaj/M+AElX9XY/jk4CXVXVOX/ewdSjGGDNwIb0ORUQuA74HnH+iYCIiGapaLiJ5wCeBs5zjWapa5hT7BJ7usT6tXbu2UkT297OaaUBlP8uOJvbcY4s999gy2Oee2J9CQWmhiEgxEAUcdQ59qKq3ikg28KiqXuGUexcYB7QBd6rqSuf4/+Hp7lJgH/CVLgEmUHUs7E9EHm3succWe+6xZbifOygtFFWd5ud4KZ7Bd+/n8/yU+8IwVc0YY8wghcIsL2OMMaOABRT/Hgl2BYLEnntsseceW4b1ucdUtmFjjDHDx1ooxhhjAsICijHGmICwgNKDiFwmIjtFpFhEvh/s+gyGk46mXES2dDmWKiLLRWSX8zXFOS4i8mvneTeJyIIu19zglN/lLET1Hj9dRDY71/xaREJiz1IRyRWRt0Rku5Oh+g7n+Kh+dhGJFpGPRGSj89w/co5PFpHVzjP8VUQineNRzudi5/ykLve6yzm+U0Qu7XI8ZH8vRCRMRNaLyMvO51H/3CKyz/k53CAihc6x4P+cq6q9nBcQBuwGpgCReNK9zA52vQbxHIuBBcCWLsd+Dnzfef994L+c91fgyaUmwJnAaud4KrDH+ZrivE9xzn2EZ5GpONdeHuxnduqVBSxw3ifgyWQ9e7Q/u1OXeOd9BLDaeZ5ngOuc4w8DX3Xefw142Hl/HfBX5/1s52c+Cpjs/C6EhfrvBXAn8CSejBmMhefGs/4urcexoP+cWwulu4VAsaruUdVW4GlgWR/XhBxVXQUc63F4GfBn5/2fgY93Of64enwIJItIFnApsFxVj6lqFZ6capc55xJV9V/q+cl7vMu9gkpVy1R1nfO+DtgOTGCUP7tT/3rnY4TzUuAi4DnneM/n9v73eA5Y4vwFugx4WlVbVHUvUIzndyJkfy9EJAe4EnjU+SyMgef2I+g/5xZQupsAHOzyucQ5NhpkqpNNwPma4Rz398wnOl7i43hIcbozTsPz1/qof3an22cDUI7nH4bdQLUe3yaia107n885X4MnI8VA/3uEgl8B/w54N1Ufx9h4bgXeEJG1InKLcyzoP+e2BXB3vvoJR/u8an/PPNDjIUNE4oG/Ad9U1doTdP+OmmdX1Q5gvnh2P30Bz35CvYo5Xwf6fL7+8Az6c4vIVUC5qq4VkQu8h30UHVXP7ThHVUtFJANYLiI7TlB2xH7OrYXSXQnd91zJAUqDVJdAO+I0ZXG+ljvH/T3ziY7n+DgeEkQkAk8weUJVn3cOj4lnB1DVauBtPH3lySLi/aOxa107n885n4Sni3Sg/z2C7RzgahHZh6c76iI8LZbR/tyoJ00VqlqO5w+IhYTCz3mwB5dC6YWnxbYHz8CcdxDulGDXa5DPMonug/L3033A7ufO+yvpPmD3kXM8FdiLZ7AuxXmf6pxb45T1DthdEezndeolePp7f9Xj+Kh+diAdSHbexwDvAlcBz9J9cPprzvuv031w+hnn/Sl0H5zeg2dgOuR/L4ALOD4oP6qfG4gDErq8/wC4LBR+zoP+gxBqLzwzIorw9EHfHez6DPIZnsKz8Vgbnr82voSnr3glsMv56v3BEeC3zvNuBgq63OeLeAYoi4GbuhwvwLNlwG7gIZyMC8F+AefiaZpvAjY4rytG+7MDc4H1znNvAe5xjk/BM1un2PlHNso5Hu18LnbOT+lyr7udZ9tJl5k9of57QfeAMqqf23m+jc5rq7deofBzbqlXjDHGBISNoRhjjAkICyjGGGMCwgKKMcaYgLCAYowxJiAsoBhjjAkICyhmVBGRDicD60YRWSciZ/dRPllEvtaP+74tIgWBq+nJT0QeE5Frg10PEzosoJjRpklV56vqPOAu4D/7KJ+MJwttSOqy4tuYkGcBxYxmiUAVePJ7ichKp9WyWUS8WWPvA6Y6rZr7nbL/7pTZKCL3dbnfp8Sz70iRiJznlA0TkftFZI2z18RXnONZIrLKue8Wb/munD0t/su550ciMs05/piI/LeIvAX8l7PPxd+d+38oInO7PNOfnLpuEpFrnOOXiMi/nGd91slthojcJyLbnLK/cI59yqnfRhFZ1ccziYg85NzjnxxPPmgMYMkhzegT42TdjcazP8pFzvFm4BPqSRaZBnwoIi/hSVExR1XnA4jI5XhSdS9S1UYRSe1y73BVXSgiVwD3AkvxZCGoUdUzRCQKeF9E3gA+Cbyuqj8VkTAg1k99a517Xo8nD9VVzvHpwFJV7RCR3wDrVfXjInIRnvQy84EfON/7VKfuKc6z/YdzbYOIfA+4U0QeAj4BzFRVdZJIAtwDXKqqh7oc8/dMpwEzgFOBTGAb8Md+/V8xY4IFFDPaNHUJDmcBj4vIHDzpJ34mIovxpDqfgOcfxZ6WAn9S1UYAVe26r4w32eRaPLnSAC4B5nYZS0gC8vHkQvqjk6zy76q6wU99n+ry9YEux59VTwZh8KSUucapz5siMk5Ekpy6Xue9QFWrnAy8s/EEAfDkoPoXUIsnqD7qtC5edi57H3hMRJ7p8nz+nmkx8JRTr1IRedPPM5kxygKKGbVU9V/OX+zpeHIypQOnq2qbk6E22sdlgv9U3S3O1w6O/+4I8A1Vfb3XjTzB60rg/0TkflV93Fc1/bxv6FEnX9f5qqvg2TTpMz7qsxBYgicI3QZcpKq3isgip54bRGS+v2dyWmaWq8n4ZWMoZtQSkZl4ssYexfNXdrkTTC4EJjrF6vBsF+z1BvBFEYl17tG1y8uX14GvOi0RRGS6iMSJyETn+/0e+AOeLZl9+XSXr//yU2YV8Dnn/hcAlapa69T1ti7PmwJ8CJzTZTwm1qlTPJCkqq8A38TTZYaITFXV1ap6D1CJJ525z2dy6nGdM8aSBVzYx38bM8ZYC8WMNt4xFPD8pX2DMw7xBPAPESnEk4V4B4CqHhWR90VkC/Cqqn7X+Su9UERagVeA/3eC7/conu6vdeLpY6rAMwZzAfBdEWkD6oHr/VwfJSKr8fxx16tV4fgh8CcR2QQ0Ajc4x38C/NapewfwI1V9XkRuBJ5yxj/AM6ZSB7woItHOf5dvOefuF5F859hKPBlsN/l5phfwjEltxpOB950T/HcxY5BlGzYmSJxutwJVrQx2XYwJBOvyMsYYExDWQjHGGBMQ1kIxxhgTEBZQjDHGBIQFFGOMMQFhAcUYY0xAjKl1KGlpaTpp0qRgV8MYY04qa9eurVTV9L7KjamAMmnSJAoLC4NdDWOMOamIyP7+lLMuL2OMMQFhAcUYY0xAhFxAEZEfO5v6bBCRN0Qk20+510SkWkRe9nXeGGPMyAq5gALcr6pznT0tXsazAZDPcsAXRq5axhhjTiTkAoqTltsrDj/7L6jqSjwZVI0xxoSAkJzlJSI/xZPuu4Yh7rkgIrcAtwDk5eUNvXIGgLrmNmIiwggPC7m/SYwxQRKUfw1EZIWIbPHxWgagqnerai7wBF02EBoMVX1EVQtUtSA9vc9p1KYf3G7lwl+8zROrDwS7KsaYEBKUFoqqLu1n0SeBfwL3DmN1zADVtbRTWd/KvqMNfRc2xowZIddf4ewe53U1zs56JnTUNrU5X9uDXBNjTCgJuYAC3Od0f20CLgHuABCRAhF51FtIRN4FngWWiEiJiFwanOqOPTVOQPF+NcYYCMFBeVW9xs/xQuDmLp/PG7FKmW46WyjNFlCMMceFYgvFhLiazi4vCyjGmOMsoJgB8waUumYbQzHGHGcBxQyYtVCMMb5YQDED1tlCaWmnw+0zkYExZgyygGIGrOvsrjobmDfGOCygmAHrGlBsLYoxxssCihmwbgHFWijGGIcFFDNgtU1tREe4Ot8bYwxYQDGDUNPURm5KLGAtFGPMcRZQzIDVNLWRm+oEFBtDMcY4LKCYAVFVapvbyU2JAayFYow5zgKKGZB6Z+1JdnIMLrEEkcaY4yygmAHxBpDk2AgSoiNsUN4Y08kCihkQb0BJiokgMSacWsvnZYxxWEAxA+INKIkxESTFWAvFGHOcBRQzILVdWyjRETYob4zpZAHFDEhNz4Bi04aNMQ4LKGZAeo6h2CwvY4yXBRQzIDVNbYS5hPiocOvyMsZ0YwHFDEhNUxuJ0eGICIkxETS2dtDW4Q52tYwxISDkAoqI/FhENonIBhF5Q0SyfZSZLyL/EpGtTtlPB6OuY1FNUztJMREAJEaHA7YVsDHGI+QCCnC/qs5V1fnAy8A9Pso0Ater6inAZcCvRCR5JCs5VtU0tXUGlKRYz1ebOmyMAQgPdgV6UtXaLh/jgF57zKpqUZf3pSJSDqQD1cNfw7GtpqmNxM4WihNQbBzFGEMIBhQAEfkpcD1QA1zYR9mFQCSw28/5W4BbAPLy8gJb0TGotqmtMzGkN7DYTC9jDASpy0tEVojIFh+vZQCqereq5gJPALed4D5ZwP8BN6mqz5FhVX1EVQtUtSA9PX04HmdMqe3S5dXZQrG1KMYYgtRCUdWl/Sz6JPBP4N6eJ0Qk0Tn3H6r6YQCrZ/xQ1W5jKIkxnh8f6/IyxkAIDsqLSH6Xj1cDO3yUiQReAB5X1WdHqm5jXWNrB+1u7T2GYl1exhhCMKAA9zndX5uAS4A7AESkQEQedcr8G7AYuNGZXrxBROYHqb5jRtdV8gCxkWGEu8RaKMYYIAQH5VX1Gj/HC4Gbnfd/Af4ykvUyvQOKd3GjrzEUVaWl3U10RNiI1tEYEzyh2EIxIapnQAHP4kZfLZRXNh9mzr2v85uVu2i3lfTGjAkWUEy/+QwoMRE+pw1vKqmm3a38cnkRn37kQw4cbRyxehpjgsMCiuk33y0U35tsHaxqZEpaHA9eN5+iI3Vc8et3eW5tCaq91qkaY0YJCyim32q77Nbo5W8b4JKqJnJSY1k2fwKv3nEes7MT+c6zG3lzR/mI1dcYM7IsoJh+q2lqQwQSoo7P5fC3DfDBY43kOCvqc1Ji+b8vLSTMJWw4aNlxjBmtLKCYfvOkro/A5ZLOY772RKlvaaeqsa0zoABEhYeRlxrL7or6EauvMWZkWUAx/dZ1lbxXYkwEzW1uWto7Oo8dqmoCIDcltlvZqelx7C5vGP6KGmOCwgKK6TefAcXZE6XrWpSDxzwzurq2UACmpsezt7KBDrcNzBszGllAMf3mr4UC3fN5lVR5Akpuas8WSjytHe7O86GirrmNPdYVZ8yQWUAx/ea7hdI7n1dJVRPRES7GxUV2Kzs1Iw4g5MZRfvvWbj7xuw9sSrMxQ2QBxfRbbZfNtbyOt1C6dHlVNZKTEouIdCs7JS0eIOTGUfZU1FPT1EZVo+UkM2YoLKCYfumZut4ryZvCvkcLJbfH+AlASlwk4+IiQ66FUlrjmURQWt0U5JoYc3KzgGL6pamtg7YO9d/l1dw9oOT0mOHlNTU9PvQCSnWz89UCijFDYQHF9IuvtCvQexvg2uY2apraes3w8pqaEcfuitDp8mpq7eBYQytgAcWYobKAYvrFX0CJCncRGebqnDZccsxZg5Lqv4VyrKG18x/xYPN2dwGU1TQHsSbGnPwsoJh+qWn0HVA8e6IcT2F/sMr3GhSvqemegflQmabbtVVyyFooxgyJBRTTL/5aKNA943CJs0r+RGMoEDpTh70BJTc1xlooxgyRBRTTLycMKDERndOGS6oaiYsMIyW2dzmACSkxRIa72BMi4yiHqpsRgQV5KTaGYswQWUAx/dJnQHHOHzzW5HMNileYS5iSFhdSLZTMhGgmpsZypLbZdpc0ZggsoJh+qfWmro8O73Wu6zbAJVWNfsdPvDxTh0OjhVJa3UR2cjRZyTG4FY7UtQS7SsactEIuoIjIj0Vkk4hsEJE3RCTbR5mJIrLWKbNVRG4NRl3HkpqmNhKiwrulrvfytlBUlUNVTX5neHlNTY/jwLHGbhmKg8UTUGLITo7p/GyMGZyQCyjA/ao6V1XnAy8D9/goUwac7ZRZBHzfV+AxgVPb3E6Sn3ERz6B8OzVNbdS1tPfdQsmIp8OtQd9n3u1WSmuamZAcw4TkaMACijFDEXIBRVVru3yMA3pl7FPVVlX19k1EEYLPMdp4N9fyJTEmnNYON8XlnnGR/nR5QfBneh1taKW13U12cgxZSd4Wis30MmaweneIhwAR+SlwPVADXOinTC7wT2Aa8F1VLfVT7hbgFoC8vLxhqe9Y4CuPl5f3+NZSz98C/qYMe01O82YdDu44irc1kp0cQ1xUOEkxEdZCMWYIgvKXvYisEJEtPl7LAFT1blXNBZ4AbvN1D1U9qKpz8QSUG0Qk00+5R1S1QFUL0tPTh+uRRr0TBRRvy2WbE1B67tTYU1xUONlJ0ewuD24L5XhA8XR3ZSVFU1ZjAcWYwQpKC0VVl/az6JN4WiH3nuBepSKyFTgPeC4A1TM+nDCgOMe3ldWSEBVOYkzfP1ZTM/pOEqmqbCqpYW5Okt9pyEPhXRk/wRmQn5AcwyHr8jJm0EJu7EFE8rt8vBrY4aNMjojEOO9TgHOAnSNTw7HpxC0UTwDZebiOnFT/a1C68k4dPtGmVk+vOciy377Pmn1Vg6t0H0qrm4mNDOt8ruzkGOvyMmYIQi6gAPc53V+bgEuAOwBEpEBEHnXKzAJWi8hG4B3gF6q6OTjVHf2a2zpobXf32lzLy3u8tcPd54C819T0OOpb2in3s+6jtd3NQ28WA7B2v/+A8uuVu7j7hcH9r/dOGfYGwKzkaGqa2mhoae/jSmOMLyE3KK+q1/g5Xgjc7LxfDswdyXqNZSdaJQ90m/3V1/iJV+dMr/J6MhOje51/fl0Jh6qbiAx3seGg/4DyTOFBDlU38fULp3WuJemv0pqmbtd4u77KapqYlpEwoHsZY0KzhWJCTJ8BpcuYSb9bKBn+pw63dbh56K1i5uYkcfmc8Ww8WOPzHkdqmympakIVXlh/qF/ft6vS6qbO9SeATR02ZogsoJg+9RVQosLDiI7w/Cj1N6BkJEQRHxXOlkO1vc69sO4QJVVN3H5RPvNzkzlc28xhH5mA1zldYWnxUTy3tuSE4zE9Nbd1UFnfSnbS8fpm2+JGY4bEAorp095Kz3qR8Um9u6a8vN1efaVd8RIRLpmdyV8LD/LIqt2dx72tkzkTElkyK4N5uckAbCyp7nWPtfuriAx3cefF09lb2cC6A/0fvPemqu/a5ZWZGI1LLKAYM1gWUEyf3tlZQUZCFPlON5Uv3oH5Cf1soQDcd81crpqbxc9e2cHPX9uBqvL39Yc4cKyR2y/KR0SYnZVIuEvYeNBHQDlQxbycJJbNzyY2Mozn1pb4/D77jzbgdndvvXRd1OgVEeYiIyGaUtsXxZhBsYBiTqi9w827uyo4f3r6CacDJ0Z7Vpr7S8/iS2S4iwevO43PLsrjd2/v5u6/b+G3bxUzOyuRi2d71qlGR4QxKyuxVwulua2DLYdqWDAxhbiocC6fk8XLG8toau2ecPKNrYc5//63eXrNwW7He65B8cpOjrYWijGDZAHFnNCGg9XUNrdzwYyME5bLSo5heqb/Fow/YS7hpx+fw1cvmMqTqw+w72gjty/J7xa85uUmselgTbdWxuZDNbR1KKfnpQBw7ek51LW088a2w51lSqoa+c6zGwH4+4bug/al1U2IQGZSVK/nsJ0bjRkcCyjmhN7eWUGYSzg3P+2E5X728VP5n8+fPqjvISJ877KZ/PBjs/nU6TlcMrt7Fp15OcnUtbSzp/J47i/v2pQFEz0BZdHkVHJSYjq7vdo63HzjqfWoeoLNmn3HuqVVKa1uIj0+iqjwsG7fy7NavmlAA/zGGA8LKOaE3imq4LTcZL8zvLySYiNIi486YZm+3HjOZO7/1Lxee67M9w7MdxlHWbu/islpcZ3f0+USPrkgh/eKKymtbuIXr+9k/YFq/vOaU/nqBVNRhX9uKuu8vrS62ee6leykaFrb3RxtaB3SsxgzFllAMX5V1LWw+VANF8wIblLNKenxxEeFd46jqCrr9lexwOnu8rpmwQRU4fvPb+Z/V+3hc4vyuGpuNlPT4zklO5GXuwWUpl7jJ+Dp8gIos7UoxgzYoAOKiNhK9VFuVVEFQJ/jJ8MtzCWcOiGps4Wy/2gjRxtaOX1i94AycVwcCyensqqogpnjE/jBVbM7z101N5sNB6s5eKzRs7Oks/VvT94gc8gG5o0ZsKG0UNaLSLGzZe/svoubk83bRRWkxUcyOysx2FVhXm4y28pqaWnv6Bw/6RlQAL583hRyU2P47ecWEB1xfHzkqrlZALy8qYxjDa20OBtr9ZTdJf2KMWZghhJQNgEfd+7xkohsFJHvi8ikQFTMBFeHW3l3VwWLp6f73Ed+pM3PTaKtQ9lWWkvh/ioSosJ9rou5eHYmq757YWeuMK/c1FhOy0vmHxtLO1Or+AooKbERRIW7bOqwMYMwlICiqrrF2QxrGvBlIAN4V0Q+CEz1TLBsLKmmurEt6N1dXvNzPa2RjQerWbe/itMmpvgNdP7Wy1w1N5ttZbW8V1wJ9F6D4r12QnKM5fMyZhCGElC6/daq6keqeieQB9w1pFqZoHt7ZwUugfOmnXi68EgZnxRNZmIU7+6qpKi8rnP9yUBceWoWIvCn9/cCvlso4EljX2pdXsYM2FACyv2+DqrHO0O4r+mnDrdy+YPv8sJ63ylHhuKdogrm5SaTEhcZ8HsP1rycZN7cWY6q7/GTvoxPimbhpFTK61qIjnCREut7KnR2km20ZcxgDDqgqOqT3vciEiMiMwJTJdNfJVWNbC+rZeX28oDe92h9C5tKqrlgemh0d3nNy01GFVwC8/OSB3WPq+ZlA3TbWKun7OQYyutaaOtwD7quxoxFQ16HIiIfAzYArzmf54vIS0O9r+nbriOevUS2lvZOAT8U7+6qRBXOD/L6k568Cxxnjk8kPmpwe8NdPmc8YS7xOX7ilZ0cjSo+U+YbY/wLxMLGHwILgWoAVd0ATArAfU0fip3NqfZWNlAfwG1r/7auhPSEKOZOSArYPQPh1JwkXDK47i6vtPgobr8on2tPz/FbZuK4OACu+Z8P+PlrO9h/tKHb+fYON/sqG6iy1fTGdBOILYDbVbXmRJlozfDwtlAAtpfVcsak1CHfc92BKt7dVcldl88MienCXSVGR/CHG85gdvbQ1sXcsTT/hOcXTU7l0esLeOqjAzz8zm5+9/ZuzpySSnxUOHsqGzhwtJF2tzIjM4HXvnneCbMwGzOWBCKgbBGRzwJhIpIP3A4MetqwiPwYWAa4gXLgRlUt9VM2EdgOvKCqtw32e56siivqmZYRT3F5PVsO1QQkoPxm5S5SYiP4/JkTA1DDwLtw5vCP64gIS2dnsnR2Jodrmnlu7UH+vqGU6sY2ZmQmcNkp46lqbOWpjw6ytbSWOSHWkjMmWAIRUL4B3A20AE8BrwM/HsL97lfVHwCIyO3APcCtfsr+GBiTM8pUleIjdVx7eg7VjW0BGUfZVFLNWzsr+O6lM4gb5BjFaDM+KZrbLsrntou6t2qqGlp5bm0JL20stYBijGPIYyiq2ugsbjxDVQuc94MezVTVrv8yxgE+84iLyOlAJvDGYL/XyaysppmG1g6mZSZwSnYiWw7VDPmev15ZTFJMBNefFZqtk1CSEhfJ+dPTeWlDaa/dII0ZqwIxyytdRO4XkVdE5E3va4j3/KmIHAQ+h6eF0vO8C/gl8N2hfJ+TWXG5Z/wkPyOeORMSKS6vp6W9o4+r/NtyqIYV24/wpXMnkzCAXRfHsqvnT+BwbTMf7TsW7KoYExICMcvrCWAHMBn4EbAPWHOiC0RkhYhs8fFaBuC0cnKde/saG/ka8IqqHvRxruf3ukVECkWksKKiYmBPFsJ2OQFlWkY8p2Qn0e5Wig7X93GVfw+9WUxCdDg3nD0pQDUc/ZbOyiA2MowXN/gc4jNmzAlEQBmnqn8A2lT1HVX9InDmiS5Q1aWqOsfH68UeRZ8ErvFxi7OA20RkH/AL4HoRuc/P93rE6YorSE8PrXUVQ1FcXk9KbATj4iI5xZn1tKV0cN1eOw7X8trWw9x0zuQ+N9Iyx8VGhnPJ7Exe2VxGa3v3RZCqyuo9Ryncd4za5rYg1dCYkRWIkVfvb0uZiFwJlAL+J/n3QUTyVXWX8/FqPK2fblT1c13K3wgUqOr3B/s9T0bF5XXkZyQgIuSlxpIQHc7WAQaUDreyeu9RHlheRHxUOF88Z9LwVHYUWzZ/An/fUMqqogqWdtm6+M8f7OOH/9jW+Tk7KZoZ4xP4yvlTOXPKuGBU1ZhhF4iA8hMRSQK+DfwGSAS+NYT73eekcXED+3FmeIlIAXCrqt48xPqe9FSVXeX1XD7Hs8eHiDA7K5Eth/o302vt/ipe3HCIVzYfprK+hZiIMO6+chbJsaGTt+tkcW5+GimxEby4sbQzoPxr91F+/M/tLJ2VwWcX5bHjcB1Fh+t4r/godz2/mTe/fb6tXTGj0pADiqq+7LytAS4MwP18dXGhqoVAr2Ciqo8Bjw31+55Mjja0Ut3Y1m0/kDkTknhi9X463ErYCRYkPrPmIP/+t01ER7hYMjOTK+dmceGMDGIiw/xeY/yLCHNx5dwsnltbQkNLO9VNbXz9yXVMGhfLA5+eT0J0BBfN9ASaZwsP8t3nNrHuQBWnTxz6miFjQs2QA4qITMazFmVS1/up6tVDvbfxzbtCPj/zeEA5JTuR5jY3eyrqyc9M8HndWzvLueuFzSyens7/fG6BrTUJkGXzJ/CXDw/wj42l/GX1ftra3TxyfUGv2XKXn5rFPS9u5bm1hyygmFEpEP+i/B34A6N4+Q4AACAASURBVPAPPN1UZpA2l9SQHBtBbmrsCct5c3hNy+gaUDyL67aU1vgMKJtLavj6E+uYOT6B31kwCajT81KYkBzDD17cQrtbefT6gl47RgLER4Vz+ZzxvLyxlHs/NrvbFsXGjAaBmOXVrKq/VtW3nFle79h+KINz8+Nr+OFLW/ssV3ykjviocMYnRncem5oeR1S4i60+xlEOHmvkpsfWkBIbyZ9uPGPQmXqNby6X8LF52bR1KHcunc6SWZl+y15zeg51Le28se3ICNbQmJERiH9ZHhSRe/GsWG/xHlTVdQG495hRWd/CkdoWGluO9TkOsqu8nqkZ8d0GdsPDXMzMSuw1dbiqoZUb/vQRbR1unr5lERldgpAJnG9cNI15OUlcesr4E5Y7a8o4spOi+dvaEq529mYZrA638qsVRbyx9QjP3HqWTfk2QReIgHIq8AXgIo53eanz2fTTzsN1ANS1tLOttJZTc/znhyour2fx9N5rauZkJ/KPjaWoKiJCfUs7Nz62hpKqJp64eRHTMnyPrZihi4sK5/JTs/os53IJn1yQw+/eLuZIbTOZgwzwxxpauePp9by7qxKAd3dVcNXcoQUoY4YqEF1enwCmqOr5qnqh87JgMkDby453Va3ee9RvuZqmNsrrWrrN8PI6JTuJ2uZ2SqqaaG7r4JbHC9lyqIbffnZBQDIRm8D45IIJuBX+vv7QoK7feLCaj/3mPVbvPcbPPnEqidHhrCoaPVkgzMkrEAFlIzC4/VhNpx2H60hPiGJyWhwf7vEfUIrLew/Ie3lXzG84WM3tT63ng91Huf/auVw823+fvhl5U9LjWZCXzN/WlaDa/8SSqsoTq/fzqYf/BcDfbj2bzy7K45xpaawqqhzQvYwZDoEIKJnADhF5XURe8r4CcN8xZefhOmaOT2DR5FQ+2usZR/GluNzTNZbvo/tqxvgEwlzCD17cwhvbjvDDj83mkwsGnbTADKNrTs+h6Eh9vxejNra2c+czG7n7hS2cOXUcL3/j3M5u0cXT0zlc20zRkcHncjMmEAIxhnJvAO4xprV3uCk6Usf1Z03klOwknl5zkO1lvjdu2nWknugIFxNSeu+JHh0RRn5GPDsO1/GtpdO58ZzJI1F9MwhXzc3mR//Yxp/e38uSWZnsOFzL9rJaSqqaOGdaGsvmZ3PqhCREhOLyer72xFp2lddz58XTue3Cad120/SOp60qqmDGeBsnM8ETiJXyNkV4iPYdbaSl3c3M8YksmuIZ61i995jPgFJcUc+UtHi/s8BuPX8qh2ub+criKcNaZzM0STERXDw7k+fXH+L59YcIcwlT0uJIT4ji8X/t4w/v7WVKWhyLp6fzbOFBoiLCePyLCzkvv/dkjAnJMUzLiGfVrgq+bP/fTRANOqCISB2+N78SQFV1aBt/jyE7Dnu6PWZmJZCVFMPEcbF8uOcoXzq3dwtj15F6Cial+L3Xx0+bMGz1NIF1z1WzWTIzg/yMBPIz4zsXOtY0tvHqljJe3FDKn/+1jwV5KTz02dPISurdKvVanJ/OX1bvp6m1w9LomKAZdEBRVWtbB8iOsjrCXNI50L5ociqvbz2C263dujYaW9s5VN3Edem5waqqCaDMxGifY1xJsRFctzCP6xbmUd/STlxkWJ/JJBdPT+OP7+9l9d6jXDAjY7iqbMwJBWJQ3gzRjsN1TEmLIyrc85flmVPGUdPUxg5nbYqXrxxeZnSLjwrvV2biRZPHERnuYlVR5QjUyhjfLKCEgB2Ha5mZdbyHcJGzX0bX9Siqym/e3EV0hIvT8vx3eZmxKSYyjEWTU1m1y9ajmOCxgBJktc1tlFQ1MbPL7JwJyTHkpsZ0W4/y0sZSVmwv5zuXzBj06mozup0/PZ3i8noOVTcFuypmjLKAEmRFTrfWrKzuQ1KLJo/jo73HcLuViroW7n1pKwvykrnJpgIbP7pOHzYmGCygBNl2J6DMGN99UtyZU8ZR1dhGUXkd97y4hcbWDn5+7bwTJo00Y1t+RjzjE6MtoJigsYASZDsP15IQHU52UvdurEWTPetRfvLydl7dcphvLs33mW7FGC8RYfH0NN4rrqS9w7YmMiPPNsYIsh1ldcwan9hrJk9uaiwTkmN4r7iSUyckcct5tmDN9G3x9HSeKSzh3pe2MiElhrjIcOKiwrlwRjrj4qOCXT0zyllACSJVZcfhOj65wPdixLOnjuPvGw5x/6fmEh5mjUnTt/Py05k4LpanPjpA13Rws7IS+ec3zu22rsmYQAu5gCIiPwaW4dlbpRy4UVVLfZTrADY7Hw+cjHvYl1Q1Ud/SzszxvpMKfO/ymXzhrIl+zxvTU1JMBO9890JUlZZ2N/Ut7by6uYwfvLiVN7Yd5rI5fe/Z4ktZTRONrR0+tzY2xisU/+y9X1Xnqup84GXgHj/lmlR1vvM66YIJ0Llw0V9Cv7T4KObm2M4AZuBEhOiIMNLio/jMwjympMXxqxW7cPvJYn0ire1uPvfoaq7/w0eWIt+cUMgFFFXtms87Dt/5wkbU7op6nltbEvD77nRyeFmGWDOcwsNc3L4knx2H63h1y+EBX/+H9/ayp6KBQ9VN7DxS1/cFfqzcfoQ7/7ph0Neb0BdyAQVARH4qIgeBz+G/hRItIoUi8qGIfPwE97rFKVdYUTG46ZQPv72b7z63cdA77Pmz/XAdeamxxEeFXM+jGWU+Ni+bqelxPLiyaECtlNLqJn69chcLnR0/39ox+CnJzxQe5Pn1h6hqaB30PUxoC0pAEZEVIrLFx2sZgKreraq5wBPAbX5uk6eqBcBngV+JyFRfhVT1EVUtUNWC9PTeqb/748cfn8Oiyal8+9mNrNh2ZFD38GVHWW23FfLGDJcwl3DH0ukUHannn5vL+n3dT/65DUX55b/NY3ZWIm/tKB/U91dV1h2oBmBPpW0ENloFJaCo6lJVnePj9WKPok8C1/i5R6nzdQ/wNnDacNU3OiKMR284gznZiXztyXV8sLvvBHwt7R00trb7PX/wWCN7KxssoJgRc+WpWeRnxPPgyl1+dwTtalVRBa9sPsxtF04jNzWWC2ems/ZAFTWNbQP+3oeqm6ioawFgd0XDgK83J4eQ6/ISkfwuH68GdvgokyIiUc77NOAcYNtw1is+KpzHblrIxNRYvvznQjYcrPZZbl9lAz95eRtn/GQFc3/4Bv/28L/4zcpdrD9QRXltM3/+YB/X/s8HnPfztwA4b/rgWk3GDJSnlZJPcXk9L2/qNXGym5b2Dn740lYmjYvt3LTrwhkZdLiVd4sH3u21/sDx35c9FlBGrVDsvL9PRGbgmTa8H7gVQEQKgFtV9WZgFvC/IuLGExTvU9VhDSgAKXGR/OXmRVz78Ad8/tHVLJiYwoTkaCYkx5AaF8Ub2w7zTlEFYSJcNmc8OSmxvF9cyS+XF/HL5UWd95k5PoHvXjqDq+ZmMXFc3HBX25hOV8zJYkZmMf+9vIgLZmSQFBPhs9yj7+5lT2UDj910Rue2CqflpZAcG8GbO8q5am72gL7vugNVREe4yEqKYXeFdXmNViEXUFTVXxdXIXCz8/4D4NSRrJdXZmI0T958Jj9/fSf7jzawrbSGynrPIGN6QhR3LMnnMwvzumUEPlrfwvu7j3KoqoklszKYnmndXCY4XC7hh1efwvV/XM2XHlvD/31pUa8dHl/bcpgHV+zi0lMyu23WFeYSFuen887Oil6bv/Vl/YFq5k5IJjk2wgLKKBZyAeVkkJsay28+c3zIprmtgyO1zWQlxRAZ3rsXcVx8FFfPG9hfdMYMl7OmjuNXnz6N255ax1efWMsjXyjo/Ll9Zs1Bvv/8JubnJvPza+b1uvbCmem8tLGUzYdqmJfbvzVSLe0dbCut5aZzJ+ES4a2d5bR3uC37wyhk/0cDIDoijInj4nwGE2NC0ZVzs/jZJ07l7Z0VfPvZjXS4ld+v2sO//20T50xL4y83LyIptnd32PnTMxCBNwcw22vLoVpaO9yclpvClLQ42jqUg1W2Z8toZC0UY8aozyzMo6apjfte3cHeynq2HKrlylOz+O9Pz+scN+kpNS6S+bnJvL2znG9dPL1f32f9gSoAFuQldwaS3eX1TE6z8cPRxv6kNmYMu/X8qdx6/lS2HKrlMwtz+fVnTvMbTLwunJHBxpKazmnAfVl/oJoJyTFkJEYzNd0TRGwtyuhkAcWYMe57l81gxZ3n87NPnNqvDdwumukZqH+nnxt5rT9QxWl5nvGW5NhIxsVFDnrq8Id7jnLRL96mtnnga2HM8LOAYswYJyJMy4jvtSePP7OzEklPiOKtnX2Poxyuaaa0ppkFeSmdx6akxw06oKzYdoQ9lQ1sOVQzqOvN8LIxFGPMgLhcwgXT03llcxnfeXYjre1u2jrcuFzCHUvyu02L946feFsoAFPS4lmxfXApjDaVeALJriP1nD01bQhPYYaDtVCMMQN2zek5JMZE8EFxJRtLqikur2fVzgq+/sQ6mts6OsutP1hNZJiL2dnH9/SZkh7H0YbWAadw6XArW0o9AWUoWY/N8LEWijFmwM6cMo5/3bWk27G3d5Zz45/W8MDyIu66YhbgaaHMmZDYbaDfu0nX7sr6bl1hfSkur6ex1ROsdllACUnWQjHGBMQFMzL4zMI8Hnl3D2v3H6O13c2mkhpO6xE0pnhneg1wHGVjiScf2KLJqRQdqbfNvkKQBRRjTMDcfeUsJiTH8O1nNrLuQBUt7e5erZDc1FjCXTLgFCybS2qIjwrn0lPGU9PURnk/py2bkWMBxRgTMPFR4dx/7Tz2HW3kW87ujF0H5AEiwlzkjYtlzwADyqaSauZMSGRmlmfQv8i6vUKOBRRjTECdNXUcN50zibKaZjITo8hKiu5VZmp6/IC6vFrb3Wwvq2NeTnLnLLKiI7Y4MtRYQDHGBNy/XzqT6ZnxnJef7nN9y5T0OPYfbaS9w92v++047MkHNjcnmbT4KFLjIm1gPgTZLC9jTMDFRIbxj2+cS7jL99+sU9Piae1wU1LVxKR+5PTyrj+Zm5MEwPTMeJs6HIKshWKMGRZR4WF+U7lMGWBOr00l1aTERpCTEgPA9MwEim2mV8ixgGKMGXHetSj9HUfZVFLD3Jzkzu6z/MwE6lraKatpHrY6moGzgGKMGXEpcZGkxEawux8BpbG1naIjdcxzursAZjgD89btFVosoBhjgmJKeny/1qJsK63FrTA35/j04+mZnhaODcyHFgsoxpigmNrPrMMbewzIgycNfnpClE0dDjEhF1BE5McisklENojIGyLiczN2Eclzzm8XkW0iMmlka2qMGYop6fFU1rf0ubfJppJqxidGk5HYfT3LjMwEa6GEmFCcNny/qv4AQERuB+4BbvVR7nHgp6q6XETigf5NaDfGhIQpznThKx58l5TYSBKiw0mIDudTp+eydHZmZznPgHxSr+vzM+N5+qODuN2Kqx8bg5nhF3ItFFWt7fIxDug1L1BEZgPhqrrcuaZeVRtHqIrGmAA4Nz+Nm86ZxGl5KaTFR9LW4WZzSQ03P17IA8uLcLuVmqY29lY2MC83udf10zMTaGrr4FB1UxBqb3wJxRYKIvJT4HqgBrjQR5HpQLWIPA9MBlYA31fVjp4FReQW4BaAvLy8YauzMWZgYiPDufdjp3Q71tzWwX/8fQsPrtzFtrJarj09B8BnC+V4CpY6clNjh7/Cpk9BaaGIyAoR2eLjtQxAVe9W1VzgCeA2H7cIB84DvgOcAUwBbvT1vVT1EVUtUNWC9PT0YXkeY0xgREeEcf+1c7n3Y7N5c0c533hyPQBzJ/RuoeQ7M71s6nDoCEoLRVWX9rPok8A/gXt7HC8B1qvqHgAR+TtwJvCHgFXSGBMUIsJN50xmRmYCX39yHRPjY0mKjehVLjE6gqykaHbZTK+QEXJdXiKSr6q7nI9XAzt8FFsDpIhIuqpWABcBhSNVR2PM8Dt7Whorv30Bre3+59tMz0ywNPYhJOQG5YH7nO6vTcAlwB0AIlIgIo8COGMl3wFWishmQIDfB6vCxpjhkRoXyXgf6e+9pmfGU1xeT4e7/zm9qhpaefqjAzS39RpyNUMUci0UVb3Gz/FC4OYun5cDc0eqXsaY0JOfmUBLu5sDxxqZ3EfWYlXlxQ2l/H8vb+NYQyu1zW3csnjqCNV0bAjFFooxxvSLN6fXJme/eX8OHmvkhj+t4Zt/3UBeaixzc5L443v7TtidZgbOAoox5qQ1Y3wCuakx3P3CFj7YXdnrvKry2Pt7ueSBVazdd4wfXX0Kf/vq2Xzr4ukcrm3mpY2lQaj16GUBxRhz0oqOCOPZr5xNVlI0N/5xDa9tKes8V1HXwk2PreGH/9jGoimpLL/zfP7/9u49SMrqzOP49zc9d5gbV7kJYgQUMagTFDQsooWIJmg0FVK7BYakjBqzxuwm0XVXNxddDVuJZUwqa1xRsy7eYtxoZJWgBhe5ZECuIjAQjQjOjIjcL8PMs3+8p6UHumeANHTT/Xyquvrt06dfzjO8PU+/7+l5zpRRA4gViDGDujO4ZwUPzlnna6qkkScU59wJ7aSqUp6+fiRn9qnkxscXM2PhX3jl7QbG3zeHees284OJQ5l+7WfoXV32yWskcd3ogaxp2MFrq5syOPrc4gnFOXfCqy4v5r++dh6jB3XntmeXM/WROrpXlPD8Ny9k8sgBSde1/9yne9OrqpT/mLMuAyPOTVn3LS/nnDsa5cWF/GpyLXe/uIrSohg3X3wapUWxlP2LCwuYesEp3PXiKpa+93HSemHuyPgZinMuZxTFCrjzc0P53vgh7SaTuEkj+lFRWsiDc9Yfh9HlPk8ozrm8VVFaxN+d35+ZKzbx7ubDW9/epeYJxTmX174yagCFBQXc8uQSNu/Ym+nhnNA8oTjn8lqPylLumzSclRu3ceUv5lLf6LXBjpYnFOdc3pswrBdPXHc+u/e1ctUv3uD/1h76R5KuY55QnHMOOPvkGp77xih6V5UxZfpC7p+9ltUfbPc/fDwCyqcfVm1trdXVeZV751xq2/c0c8uTS/jDqkYAunUuZuSp3fjsp7px6dCTkq7NkuskLTKz2g77eUJxzrlDbdiyizfqNzN33Ye8sW4zTdv3Uhwr4OLTe3DV2X0YM7gHAPWNO1i1aRurG7bTrXMxVw7vQ4/K5CX3m1taKYqdeBeGPKEk4QnFOXc0zIzl72/l2cXv8/zSjWzeuY/OJYXsaW5hf1iLpThWwL6WVmIF4m8Gdeeac/syvF81de9uYV5IShu27OYfxw3mhjEnVtl8TyhJeEJxzv21mltaeX1tE7PeaqCmvJjTe1Vyeq8KBnTtxF8+2sUzizbwm8UbaNh24CvIFSWFnDewKy2trby6uomvjx7IrZcNSVoSJht5QknCE4pz7nhoaTVeX9vEuqad1PavYWjvSgpjBbS2Gnf+biW/nv8uX6rtx91fGEasIPuTyuEmFK/l5ZxzaRYrEGMG92DM4LbtBQXiBxOHUl1exM9eqWfr7mYmj+rP2oYdrGnYztqGHexq3k+/mnJO7lJO3y7l9Kspo3d1Gb2qSqkoze4vBHhCcc6540gS/zBuMNXlxfzwhbf435UfAFBZWsjgkyro2qmENQ3bmf124yErSnYuKaRnZQmdSgopjhVQXFhAUSy6FRaIWEwUFYiSwhiVZYVUlRVRVVZEZVkRIwd2TfllgXTJuoQi6YfARKAVaASuNbONB/W5CPhpQtMQYJKZPXfcBuqcc3+Fr154CuecXM2OvfsZ1LOCHhUlbeZUWluNxu17eW/LLjZt3cMHW3ezaeseGrbtYde+FppbWtm3v5Wde/fT3GLsb21lf6uxv8XYu7+Frbub2dN8ICE9OnXEMU8oWTeHIqnSzLaF7b8HzjCz69vp3wWoB/qa2a729u1zKM65fBJPLNt2N3NSVRmdS47uHOKEnUOJJ5OgE9BRxrsGmNlRMnHOuXxTUhijR0WMHhXH9swkLusSCoCku4DJwFbgog66TwJ+cswH5Zxzrl0Z+ZNNSX+QtCLJbSKAmd1uZv2Ax4Gb2tlPL2AY8FI7fa6TVCeprqnJ1452zrljJevmUBJJ6g/83szOTPH8zcBQM7vucPbncyjOOXfkDncOJeuKykg6LeHh54G32+n+ZWDGsR2Rc865w5F1CQW4J1z+WgaMA24GkFQr6aF4J0kDgH7AHzMxSOecc21l3aS8mV2dor0O+FrC43eAPsdpWM455zqQ1XMo6SapCXj3MLt3A/Jx2TaPO7943PnlaOPub2bdO+qUVwnlSEiqO5xJqFzjcecXjzu/HOu4s3EOxTnn3AnIE4pzzrm08ISS2oOZHkCGeNz5xePOL8c0bp9Dcc45lxZ+huKccy4tPKE455xLC08oB5E0XtJqSfWSbs30eI6GpIclNUpakdDWRdIsSWvDfU1ol6T7Q7zLJJ2T8Jopof9aSVMS2s+VtDy85n4lrgqUQZL6SXpV0ipJK0Ott5yPXVKppIWSloa4vx/aT5G0IMTwpKTi0F4SHteH5wck7Ou20L5a0qUJ7Vn7vpAUk/SmpBfC45yPW9I74ThcIqkutGX+ODczv4UbEAPWAQOBYmAp0QJfGR/bEcYxGjgHWJHQ9mPg1rB9K3Bv2J4AzAQEnA8sCO1dgPXhviZs14TnFgIjw2tmApdlOuYwrl7AOWG7AlgDnJHrsYexdA7bRcCCEM9TRCuZAvwSuCFs3wj8MmxPAp4M22eEY74EOCW8F2LZ/r4Avg38N/BCeJzzcQPvAN0Oasv4ce5nKG2NAOrNbL2Z7QOeIFqO+IRiZnOAjw5qngg8GrYfBa5MaH/MIvOBakXLAlwKzDKzj8xsCzALGB+eqzSzeRYdeY8l7CujzGyTmS0O29uBVUTleXI69jD+HeFhUbgZMBZ4JrQfHHf85/EMcHH4BDoReMLM9prZn4lWQh1BFr8vJPUFLgceCo9FHsSdQsaPc08obfUB3kt4vIHcqRfW08w2QfSLF+gR2lPF3F77hiTtWSVczjib6NN6zsceLvssARqJfjGsAz42s/2hS+JYP4kvPL8V6MqR/zyywX3Ad4H44uldyY+4DXhZ0iJJ8eU7Mn6cZ11xyAxLdp0w179XnSrmI23PGpI6A78BvmVm29q5/JszsZtZCzBcUjXwW+D0ZN3C/ZHGl+yDZ8bjlnQF0GhmiySNiTcn6ZpTcQcXmNlGST2AWZLaW+bjuB3nfobS1gaikvhxfYGNGRpLujWEU9n4SpeNoT1VzO21903SnhUkFRElk8fN7NnQnBexA5jZx8BrRNfKqyXFPzQmjvWT+MLzVUSXSI/055FpFwCfl/QO0eWosURnLLkeN2a2Mdw3En2AGEE2HOeZnlzKphvRGdt6oom5+CTc0EyP6yhjGUDbSflptJ2w+3HYvpy2E3YLQ3sX4M9Ek3U1YbtLeO5PoW98wm5CpuMN4xLR9d77DmrP6diB7kB12C4DXgeuAJ6m7eT0jWH7G7SdnH4qbA+l7eT0eqKJ6ax/XwBjODApn9NxA52AioTtN4Dx2XCcZ/xAyLYb0Tci1hBdg7490+M5yhhmAJuAZqJPG18lulY8G1gb7uMHjoCfh3iXA7UJ+5lKNEFZD3wlob0WWBFe8wCh4kKmb8CFRKfmy4Al4TYh12MHzgLeDHGvAO4I7QOJvq1TH37JloT20vC4Pjw/MGFft4fYVpPwzZ5sf1/QNqHkdNwhvqXhtjI+rmw4zr30inPOubTwORTnnHNp4QnFOedcWnhCcc45lxaeUJxzzqWFJxTnnHNp4QnF5RRJLaEC61JJiyWN6qB/taQbD2O/r0mqTd9IT3ySHpF0TabH4bKHJxSXa3ab2XAz+zRwG/BvHfSvJqpCm5US/uLbuaznCcXlskpgC0T1vSTNDmctyyXFq8beA5wazmqmhb7fDX2WSronYX9fVLTuyBpJnw19Y5KmSfpTWGvi66G9l6Q5Yb8r4v0ThTUt7g37XCjpU6H9EUk/kfQqcG9Y5+K5sP/5ks5KiGl6GOsySVeH9nGS5oVYnw61zZB0j6S3Qt9/D21fDONbKmlOBzFJ0gNhH7/nQPFB5wAvDulyT1moultKtD7K2NC+B7jKomKR3YD5kn5HVKLiTDMbDiDpMqJS3eeZ2S5JXRL2XWhmIyRNAO4ELiGqQrDVzD4jqQSYK+ll4AvAS2Z2l6QYUJ5ivNvCPicT1aG6IrQPAi4xsxZJPwPeNLMrJY0lKi8zHPiX8G8PC2OvCbH9c3jtTknfA74t6QHgKmCImVkoIglwB3Cpmb2f0JYqprOBwcAwoCfwFvDwYf2vuLzgCcXlmt0JyWEk8JikM4nKT9wtaTRRqfM+RL8UD3YJMN3MdgGYWeK6MvFik4uIaqUBjAPOSphLqAJOI6qF9HAoVvmcmS1JMd4ZCfc/TWh/2qIKwhCVlLk6jOcVSV0lVYWxToq/wMy2hAq8ZxAlAYhqUM0DthEl1YfC2cUL4WVzgUckPZUQX6qYRgMzwrg2SnolRUwuT3lCcTnLzOaFT+zdiWoydQfONbPmUKG2NMnLROpS3XvDfQsH3jsCvmlmLx2yoyh5XQ78WtI0M3ss2TBTbO88aEzJXpdsrCJaNOnLScYzAriYKAndBIw1s+slnRfGuUTS8FQxhTMzr9XkUvI5FJezJA0hqhq7mehTdmNIJhcB/UO37UTLBce9DEyVVB72kXjJK5mXgBvCmQiSBknqJKl/+Pd+Bfwn0ZLMyXwp4X5eij5zgL8N+x8DfGhm28JYb0qItwaYD1yQMB9THsbUGagysxeBbxFdMkPSqWa2wMzuAD4kKmeeNKYwjklhjqUXcFEHPxuXZ/wMxeWa+BwKRJ+0p4R5iMeB5yXVEVUhfhvAzDZLmitpBTDTzL4TPqXXSdoHvAj8Uzv/3kNEl78WK7rG1EQ0BzMG+I6kZmAHMDnF60skLSD6cHfIWUXwr8B0ScuAXcCULYowfAAAAIZJREFU0P4j4Odh7C3A983sWUnXAjPC/AdEcyrbgf+RVBp+LreE56ZJOi20zSaqYLssRUy/JZqTWk5UgfeP7fxcXB7yasPOZUi47FZrZh9meizOpYNf8nLOOZcWfobinHMuLfwMxTnnXFp4QnHOOZcWnlCcc86lhScU55xzaeEJxTnnXFr8Pxhd4JbkSLXGAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2d63d43208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "learner.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner.fit_one_cycle(500, 2e-2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-15T11:15:02.866857Z",
     "start_time": "2019-08-15T11:15:02.810887Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.recorder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optional Fine tune regular fit "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:26:43.041291Z",
     "start_time": "2019-08-21T03:26:43.034173Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.callbacks.append(\n",
    "    ReduceLROnPlateauCallback(learner, monitor='train_loss', mode='min', factor=0.2, patience=0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T06:19:15.311297Z",
     "start_time": "2019-08-21T03:27:35.491980Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#learner.opt_func = RAdam\n",
    "learner.fit(50, 1e-3)# , moms=(0.75,0.70))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner.fit_one_cycle(300, 5e-4)#, moms=(0.75,0.70))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-10T11:24:45.530739Z",
     "start_time": "2019-08-10T11:24:44.737186Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.recorder.plot_lr(show_moms=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-16T16:57:30.084191Z",
     "start_time": "2019-08-16T16:57:29.686756Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.recorder.plot_losses()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-16T16:57:43.396559Z",
     "start_time": "2019-08-16T16:57:41.386234Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.recorder.plot_metrics()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner.fit(10,1e-4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Validate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:22:55.624674Z",
     "start_time": "2019-08-21T02:21:08.404032Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Better model found at epoch 0 with 👉🏻LMAE👈🏻 value: -3.0636985301971436.\n",
      "[-3.1482856, tensor(-3.0637, device='cuda:0'), tensor(-2.0012, device='cuda:0'), tensor(-3.5223, device='cuda:0'), tensor(-2.1895, device='cuda:0'), tensor(-3.4989, device='cuda:0'), tensor(-3.0694, device='cuda:0'), tensor(-3.6023, device='cuda:0'), tensor(-2.9372, device='cuda:0'), tensor(-3.6889, device='cuda:0')]\n"
     ]
    }
   ],
   "source": [
    "learner.to_fp32()\n",
    "\n",
    "val = learner.validate()\n",
    "print(val)\n",
    "val = val[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:22:56.164688Z",
     "start_time": "2019-08-21T02:22:55.627089Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss-3.5933val-3.0637\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    sub_fname = f'loss{learner.recorder.losses[-1]:.04f}val{val:.04f}'\n",
    "except:\n",
    "    sub_fname = f'val{val:.04f}'\n",
    "learner.save(sub_fname)\n",
    "print(sub_fname)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-15T11:17:05.906378Z",
     "start_time": "2019-08-15T11:17:04.743742Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:22:59.667874Z",
     "start_time": "2019-08-21T02:22:56.166810Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "30629"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import gc\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Inference\n",
    "\n",
    "Make sure `tranforms` are activated to test set otherwise TTA > 1 will be as TTA =1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:02.472345Z",
     "start_time": "2019-08-21T02:22:59.671773Z"
    }
   },
   "outputs": [],
   "source": [
    "test_fname = Path('test.npz')\n",
    "try:\n",
    "    npzfile  = np.load(fname_ext(test_fname, ext))\n",
    "    xt_xyz   = npzfile['x_xyz']\n",
    "    xt_type  = npzfile['x_type']\n",
    "    xt_ext   = npzfile['x_ext']\n",
    "    xt_atom  = npzfile['x_atom']\n",
    "    mt = npzfile['m']\n",
    "    xt_ids = npzfile['x_ids']\n",
    "except:\n",
    "    xt_xyz,xt_type,xt_ext,xt_atom,mt,xt_ids = \\\n",
    "        preprocess(test_fname.with_suffix('.csv'), type_index=types,ext=ext)\n",
    "    np.savez(fname_ext('_'+test_fname, ext), \n",
    "             x_xyz  = xt_xyz,\n",
    "             x_type = xt_type,\n",
    "             x_ext  = xt_ext,\n",
    "             x_atom = xt_atom,\n",
    "             m=mt,\n",
    "             x_ids=xt_ids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:02.480589Z",
     "start_time": "2019-08-21T02:23:02.474709Z"
    }
   },
   "outputs": [],
   "source": [
    "if True:\n",
    "    try:\n",
    "        xt_coulombmat = load_fn(f'xt_coulombmat32{ext}.npy')\n",
    "    except:\n",
    "        xt_coulombmat = np.load(f'xt_coulombmat{ext}.npy', allow_pickle=True)\n",
    "        xt_coulombmat = np.array(xt_coulombmat.tolist()).astype(np.float32)\n",
    "        np.save(f'xt_coulombmat32{ext}.npy', xt_coulombmat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:02.487954Z",
     "start_time": "2019-08-21T02:23:02.482416Z"
    }
   },
   "outputs": [],
   "source": [
    "xt_qm9_mulliken = load_fn(f'xt_qm9_mulliken{ext}.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:02.493831Z",
     "start_time": "2019-08-21T02:23:02.490149Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'xt_qm9_mulliken_ext.npy'"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "f'xt_qm9_mulliken{ext}.npy'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:02.501444Z",
     "start_time": "2019-08-21T02:23:02.495993Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(756113, 3, 29),\n",
       " (756113, 1, 29),\n",
       " (756113, 1, 29),\n",
       " (756113, 1, 29),\n",
       " (756113, 1, 29),\n",
       " (756113, 29),\n",
       " (756113,)]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[v.shape for v in [xt_xyz,xt_type,xt_ext,xt_atom, xt_qm9_mulliken,xt_ids, mt]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner.data.add_test(ItemList(items=(MoleculeItem(i,*v) for i,v in \n",
    "                              enumerate(zip(xt_xyz,xt_type,xt_ext,xt_atom,xt_qm9_mulliken,xt_coulombmat)))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:23:50.865687Z",
     "start_time": "2019-08-21T02:23:46.400798Z"
    }
   },
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'xt_xyz' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-75-5d0cf10742b2>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mgc\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0;32mdel\u001b[0m \u001b[0mxt_xyz\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mxt_type\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mxt_ext\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mxt_atom\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mxt_qm9_mulliken\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmt\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mgc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcollect\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'xt_xyz' is not defined"
     ]
    }
   ],
   "source": [
    "import gc\n",
    "del xt_xyz,xt_type,xt_ext,xt_atom, xt_qm9_mulliken, mt\n",
    "gc.collect()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "TTA_N = 10\n",
    "learner.data.test_ds.tfms = tta_tfms if TTA_N > 1 else tfms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-07-16T06:05:22.050993Z",
     "start_time": "2019-07-16T06:05:22.048033Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.model = nn.DataParallel(learner.model)\n",
    "#data.batch_size = int(4096*2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "learner.model = learner.model.module"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:33:49.692070Z",
     "start_time": "2019-08-21T02:23:50.867913Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "Total time: 52:41 <p>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>scalar_coupling_constant</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>4658147</td>\n",
       "      <td>17.557838</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>4658148</td>\n",
       "      <td>196.185962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4658149</td>\n",
       "      <td>5.034210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4658150</td>\n",
       "      <td>194.207101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4658151</td>\n",
       "      <td>17.426578</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        id  scalar_coupling_constant\n",
       "0  4658147                 17.557838\n",
       "1  4658148                196.185962\n",
       "2  4658149                  5.034210\n",
       "3  4658150                194.207101\n",
       "4  4658151                 17.426578"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sub = defaultdict(int)\n",
    "xt_ids_not_extended = (xt_ids!=0) & (xt_ids<=7163688) # TODO\n",
    "ids = xt_ids[xt_ids_not_extended]\n",
    "\n",
    "mb = master_bar(range(TTA_N))\n",
    "for tta in mb:\n",
    "    test_preds = np.zeros((0, 29), dtype=np.float32)\n",
    "\n",
    "    for batch_idx, batch in progress_bar(\n",
    "        enumerate(learner.dl(DatasetType.Test)), total=len(learner.dl(DatasetType.Test)), parent=mb):\n",
    "        _, _, preds_,_,_,_ = learner.pred_batch(ds_type=DatasetType.Test, batch=batch)\n",
    "        preds_ = preds_.sum(dim=1)\n",
    "        test_preds = np.concatenate([test_preds, preds_.data.cpu().numpy()], axis = 0)\n",
    "\n",
    "    preds = test_preds[xt_ids_not_extended]\n",
    "    for k in range(len(ids)):\n",
    "        sub[int(ids[k])] += preds[k]\n",
    "    \n",
    "for k in range(len(ids)):\n",
    "    sub[int(ids[k])] = sub[int(ids[k])]/TTA_N\n",
    "\n",
    "sub_df = pd.DataFrame(sub.items(), columns=['id', 'scalar_coupling_constant'])\n",
    "sub_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Submit to Kaggle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:33:49.698278Z",
     "start_time": "2019-08-21T02:33:49.694339Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'loss-3.5933val-3.0637'"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#sub_fname = 'loss-4.9044val-2.5880'\n",
    "#sub_fname='loss-4.9516val-2.8229'\n",
    "sub_fname"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:34:01.489116Z",
     "start_time": "2019-08-21T02:33:49.700328Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_df.to_csv(sub_fname, index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:34:01.493376Z",
     "start_time": "2019-08-21T02:34:01.490900Z"
    }
   },
   "outputs": [],
   "source": [
    "comp = 'champs-scalar-coupling'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:34:01.503188Z",
     "start_time": "2019-08-21T02:34:01.495166Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>scalar_coupling_constant</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2505537</th>\n",
       "      <td>7163684</td>\n",
       "      <td>0.730870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2505538</th>\n",
       "      <td>7163685</td>\n",
       "      <td>4.156758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2505539</th>\n",
       "      <td>7163686</td>\n",
       "      <td>1.872070</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2505540</th>\n",
       "      <td>7163687</td>\n",
       "      <td>4.130894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2505541</th>\n",
       "      <td>7163688</td>\n",
       "      <td>120.472697</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              id  scalar_coupling_constant\n",
       "2505537  7163684                  0.730870\n",
       "2505538  7163685                  4.156758\n",
       "2505539  7163686                  1.872070\n",
       "2505540  7163687                  4.130894\n",
       "2505541  7163688                120.472697"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sub_df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T02:34:01.507189Z",
     "start_time": "2019-08-21T02:34:01.504925Z"
    }
   },
   "outputs": [],
   "source": [
    "ext=''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:20:28.337421Z",
     "start_time": "2019-08-21T03:20:22.710973Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████| 64.0M/64.0M [00:04<00:00, 13.6MB/s]\n",
      "Successfully submitted to Predicting Molecular Properties"
     ]
    }
   ],
   "source": [
    "!kaggle competitions submit -c {comp} -f {sub_fname} -m 'knockout no QM9 no ext no max_atoms tta {TTA_N} {ext} transformer + attn no scale + 1024'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:21:29.884652Z",
     "start_time": "2019-08-21T03:20:28.342603Z"
    }
   },
   "outputs": [],
   "source": [
    "import time\n",
    "time.sleep(60)\n",
    "!kaggle competitions submissions -c {comp} -v > submissions-{comp}.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:21:29.911396Z",
     "start_time": "2019-08-21T03:21:29.890254Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-3.051"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submissions = pd.read_csv(f'submissions-{comp}.csv')\n",
    "submissions.iloc[0].publicScore"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# L2 data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:09.681791Z",
     "start_time": "2019-08-21T03:22:20.434837Z"
    }
   },
   "outputs": [],
   "source": [
    "learner.data.valid_dl = data.valid_dl.new(shuffle=False)\n",
    "TTA_N =1                                          \n",
    "sub = defaultdict(int)\n",
    "targets = defaultdict(int)\n",
    "targets_types = defaultdict(int)\n",
    "\n",
    "mb = master_bar(range(TTA_N))\n",
    "for tta in mb:\n",
    "    test_preds = np.zeros((0, 29), dtype=np.float32)\n",
    "    test_exts = np.zeros((0, 29), dtype=np.float32)\n",
    "    test_types = np.zeros((0, 29), dtype=np.float32)\n",
    "    test_targets = np.zeros((0, 29), dtype=np.float32)\n",
    "\n",
    "    for batch_idx, batch in progress_bar(\n",
    "        enumerate(learner.dl(DatasetType.Valid)), total=len(learner.dl(DatasetType.Valid)), parent=mb):\n",
    "\n",
    "        types_, ext_, preds_,_,_,_ = learner.pred_batch(ds_type=DatasetType.Valid, batch=batch)\n",
    "        preds_ = preds_.sum(dim=1)\n",
    "        test_preds = np.concatenate([test_preds, preds_.data.cpu().numpy()], axis = 0)\n",
    "\n",
    "        targets_ = batch[1][0].sum(dim=1)\n",
    "        test_targets = np.concatenate([test_targets, targets_.data.cpu().numpy()], axis = 0)\n",
    "        \n",
    "        test_types = np.concatenate([test_types, types_.data.cpu().numpy().squeeze(1)], axis = 0)\n",
    "        test_exts = np.concatenate([test_exts, ext_.data.cpu().numpy().squeeze(1)], axis = 0)\n",
    "    \n",
    "    test_preds = test_preds.flatten()\n",
    "    test_types = test_types.flatten()\n",
    "    test_targets = test_targets.flatten()\n",
    "    test_exts = test_exts.flatten()\n",
    "    \n",
    "    mask = (test_types != -1) & (test_exts == 0) #.squeeze(1)\n",
    "    preds = test_preds[mask]\n",
    "    test_targets = test_targets[mask]\n",
    "    test_types = test_types[mask]\n",
    "    for k in range(len(preds)):\n",
    "        sub[k] += preds[k]\n",
    "        targets[k] = test_targets[k]\n",
    "        targets_types[k] = test_types[k]\n",
    "    \n",
    "for k in range(len(sub.keys())):\n",
    "    sub[k] = sub[k]/TTA_N\n",
    "\n",
    "sub_df = pd.DataFrame(sub.items(), columns=['id', 'scalar_coupling_constant'])\n",
    "sub_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:10.021713Z",
     "start_time": "2019-08-21T03:24:09.684318Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets = pd.DataFrame(targets.items(), columns=['id', 'scalar_coupling_constant'])\n",
    "sub_targets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:10.360866Z",
     "start_time": "2019-08-21T03:24:10.024618Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets = pd.DataFrame(targets.items(), columns=['id', 'scalar_coupling_constant'])\n",
    "sub_targets.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:10.685807Z",
     "start_time": "2019-08-21T03:24:10.363096Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets_types = pd.DataFrame(targets_types.items(), columns=['id', 'type'])\n",
    "sub_targets_types.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-05T21:42:00.805651Z",
     "start_time": "2019-08-05T21:41:59.063328Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets_types.to_csv('validation_types', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:10.695479Z",
     "start_time": "2019-08-21T03:24:10.687958Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_df.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:10.704696Z",
     "start_time": "2019-08-21T03:24:10.697572Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets.tail()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-03T20:07:20.048256Z",
     "start_time": "2019-08-03T20:07:20.041023Z"
    }
   },
   "outputs": [],
   "source": [
    "mask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-03T20:09:32.533038Z",
     "start_time": "2019-08-03T20:09:30.294063Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_targets.to_csv('validation_targets', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:12.898543Z",
     "start_time": "2019-08-21T03:24:10.706332Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_df.to_csv(sub_fname + '_val', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-21T03:24:12.906168Z",
     "start_time": "2019-08-21T03:24:12.901116Z"
    }
   },
   "outputs": [],
   "source": [
    "sub_fname + '_val'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-05T21:00:08.361139Z",
     "start_time": "2019-08-05T21:00:07.258433Z"
    }
   },
   "outputs": [],
   "source": [
    "!head loss-4.9516val-3.0042_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-03T20:11:05.245077Z",
     "start_time": "2019-08-03T20:11:04.340447Z"
    }
   },
   "outputs": [],
   "source": [
    "!head loss-5.7552val-2.9950_val"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-08-03T20:11:18.206277Z",
     "start_time": "2019-08-03T20:11:17.307404Z"
    }
   },
   "outputs": [],
   "source": [
    "!head validation_targets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}