a2c_example.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# A2C (Advantage Actor Critic)\n",
    "\n",
    "\n",
    "For ReLAx & Mujoco installation see: https://github.com/nslyubaykin/relax/blob/master/README.md\n",
    "\n",
    "Useful readings regarding A2C:\n",
    "\n",
    "    A2C description: https://julien-vitay.net/deeprl/ActorCritic.html#sec:advantage-actor-critic-a2c\n",
    "\n",
    "!Note: For examining trained actor you may just load trained policies from \\trained_models directory. To do that instead of running 'Run Training and Validation loop' cell, just run 'Load Models' cell\n",
    "\n",
    "!Note: Trained policies, logs and videos are in the main repository of that notebook: https://github.com/nslyubaykin/relax_a2c_example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlyubaykin/anaconda3/envs/relax/lib/python3.6/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import gym\n",
    "import torch\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from tqdm import tqdm\n",
    "from inspect import getsource\n",
    "from torch.utils.tensorboard import SummaryWriter\n",
    "\n",
    "from relax.rl.actors import VPG\n",
    "from relax.rl.critics import BootstrappedContinuous\n",
    "\n",
    "from relax.zoo.policies import CategoricalMLP, NormalMLP\n",
    "from relax.zoo.critics import VMLP\n",
    "\n",
    "from relax.data.sampling import Sampler\n",
    "\n",
    "from relax.gym.utils import visualize_actor"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Constructing Actor's stochastic policy\n",
    "\n",
    "A2C is an RL algorithm that utilizes stochastic policy. On practice, that means that policy is a torch.nn.Module that takes a batch of observations with a shape (BATCH_SIZE, \\*obs.dim) and outputs:\n",
    "\n",
    "1) For discrete control case - batch of logits with a shape of (BATCH_SIZE, ACS_DIM), which then is passed to SoftMax function over the actions dimension (-1). Resulting probs parametrize torch.distributions.Categorical distribution\n",
    "\n",
    "2) For continuous control case - batch of actions' means with a shape of (BATCH_SIZE, ACS_DIM). After that these means together with state independent actions log(stds) (implemented as trainable torch.nn.Parameter) parameterize torch.distributions.MultivariateNormal distribution\n",
    "\n",
    "To sum up, A2C policy is a network that takes a batch of observations with a shape (BATCH_SIZE, \\*obs.dim) and outputs a parameterized torch.distributions.Distribution object\n",
    "\n",
    "Therefore any custom user defined networks meeting this requirements should work well with A2C actor\n",
    "\n",
    "So let's examine the default implementation of stochastic policy offered by ReLAx:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "class CategoricalMLP(nn.Module):\n",
      "    \n",
      "    def __init__(self, obs_dim, acs_dim, nlayers,\n",
      "                 nunits, activation=nn.Tanh(),\n",
      "                 out_activation=nn.Identity()):\n",
      "        \n",
      "        super(CategoricalMLP, self).__init__()\n",
      "        \n",
      "        layers = []\n",
      "        in_size = obs_dim\n",
      "        for _ in range(nlayers):\n",
      "            layers.append(nn.Linear(in_size, nunits))\n",
      "            layers.append(activation)\n",
      "            in_size = nunits\n",
      "        layers.append(nn.Linear(in_size, acs_dim))\n",
      "        layers.append(out_activation)\n",
      "        \n",
      "        self.layers = nn.Sequential(*layers)\n",
      "        \n",
      "    def forward(self, x):\n",
      "        \n",
      "        logits = self.layers(x)\n",
      "        y = Categorical(probs=F.softmax(logits, dim=-1))\n",
      "        \n",
      "        return y\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Discrete Case\n",
    "print(getsource(CategoricalMLP))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "class NormalMLP(nn.Module):\n",
      "    \n",
      "    def __init__(self, obs_dim, acs_dim, nlayers,\n",
      "                 nunits, activation=nn.Tanh(),\n",
      "                 out_activation=nn.Identity(),\n",
      "                 acs_scale=1, acs_bias=0,\n",
      "                 init_log_std=0.0):\n",
      "        \n",
      "        super(NormalMLP, self).__init__()\n",
      "        \n",
      "        layers = []\n",
      "        in_size = obs_dim\n",
      "        for _ in range(nlayers):\n",
      "            layers.append(nn.Linear(in_size, nunits))\n",
      "            layers.append(activation)\n",
      "            in_size = nunits\n",
      "        \n",
      "        self.layers = nn.Sequential(*layers)\n",
      "        \n",
      "        self.mean_layer = nn.Sequential(\n",
      "            nn.Linear(nunits, acs_dim),\n",
      "            out_activation\n",
      "        )\n",
      "        \n",
      "        self.log_stds = nn.Parameter(\n",
      "            torch.ones(acs_dim) * init_log_std\n",
      "        )\n",
      "        \n",
      "        self.acs_scale = acs_scale\n",
      "        self.acs_bias = acs_bias\n",
      "\n",
      "    def forward(self, x):\n",
      "        \n",
      "        x_seq = self.layers(x)\n",
      "        means = self.mean_layer(x_seq)\n",
      "        means = means * self.acs_scale + self.acs_bias\n",
      "        cov = torch.diag_embed(torch.exp(self.log_stds))\n",
      "            \n",
      "        y = MultivariateNormal(loc=means,\n",
      "                               covariance_matrix=cov)\n",
      "        return y\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Continuous Case\n",
    "print(getsource(NormalMLP))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Constructing Bootstrapped-Continuous Critic\n",
    "\n",
    "A2C is closely related to VPG in terms of architecture, the only difference is the way critic is updated and advantages are estimated. While Baseline critic subtracts aproximated Value function from reward-to-go Monte-Carlo Q-function estimation to caclucate an Advantage, Bootstrapped-Continuous critic uses a bootstrapped estimate which can be expressed as: \n",
    "\n",
    "$ \\hat{A}^{\\pi}(a_{t}, s_{t}) \\approx r(a_{t}, s_{t}) + \\gamma*\\hat{V}^{\\pi}_{\\phi}(S_{t+1}) - \\hat{V}^{\\pi}_{\\phi}(S_{t})$\n",
    "\n",
    "Actor's policy is trained in a pair with Bootstrapped-Continuous critic network which approximates value function. Critic network takes a batch of observations with a shape (BATCH_SIZE, \\*obs.dim) and outputs a batch of Value function estimations with a shape of (BATCH_SIZE, 1)\n",
    "\n",
    "Therefore any custom user defined networks meeting this requirements should work well with BootstrappedContinuous critic\n",
    "\n",
    "So let's examine the default implementation of V-Network Critic offered by ReLAx:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "class VMLP(nn.Module):\n",
      "    \n",
      "    def __init__(self, obs_dim, nlayers,\n",
      "                 nunits, activation=nn.Tanh(),\n",
      "                 out_activation=nn.Identity()):\n",
      "        \n",
      "        super(VMLP, self).__init__()\n",
      "        \n",
      "        layers = []\n",
      "        in_size = obs_dim\n",
      "        for _ in range(nlayers):\n",
      "            layers.append(nn.Linear(in_size, nunits))\n",
      "            layers.append(activation)\n",
      "            in_size = nunits\n",
      "        layers.append(nn.Linear(in_size, 1))\n",
      "        layers.append(out_activation)\n",
      "        \n",
      "        self.layers = nn.Sequential(*layers)\n",
      "        \n",
      "    def forward(self, x):\n",
      "        return self.layers(x)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(getsource(VMLP))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create training and eval environments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[1]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env_name = 'LunarLander-v2'\n",
    "env = gym.make(env_name)\n",
    "eval_env = gym.make(env_name)\n",
    "eval_env.seed(np.random.randint(100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Discrete(4)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.action_space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Box(8,)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "env.observation_space"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "acs_dim = 4\n",
    "obs_dim = 8"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define VPG actor:\n",
    "\n",
    "!Note: A2C uses the same actor class as VPG"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "actor = VPG(\n",
    "    policy_net=CategoricalMLP(obs_dim, acs_dim, \n",
    "                              nunits=64, nlayers=2),\n",
    "    device=torch.device('cuda'),\n",
    "    learning_rate=0.01,\n",
    "    gamma=0.99\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define BootstrappedContinuous critic:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "critic = BootstrappedContinuous(\n",
    "    critic_net=VMLP(obs_dim, nlayers=2, nunits=64),\n",
    "    device=torch.device('cuda'),\n",
    "    learning_rate=0.01,\n",
    "    gamma=0.99,\n",
    "    n_target_updates=1, \n",
    "    n_steps_per_update=100,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Provide actor with critic:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "actor.set_critic(critic)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Wrap envs into Sampler:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "sampler = Sampler(env=env)\n",
    "eval_sampler = Sampler(env=eval_env)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualize untrained actor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Simulation done, total reward: -223.47360955979602\n",
      "Loading video...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<video alt=\"test\" autoplay \n",
       "                    loop controls style=\"height: 400px;\">\n",
       "                    <source src=\"data:video/mp4;base64,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\" type=\"video/mp4\" />\n",
       "                 </video>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_actor(env=eval_env, actor=actor, nsteps=200, train_sampling=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Run Training and Validation loop:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 100/100 [1:36:23<00:00, 57.83s/it] \n"
     ]
    }
   ],
   "source": [
    "n_steps = 100\n",
    "logdir = '/home/nlyubaykin/Desktop/RelaxExamples/A2C/tensorboard_logs'\n",
    "run_name = 'a2c_lander'\n",
    "logger = SummaryWriter(log_dir=f'{logdir}/{run_name}')\n",
    "all_logs = []\n",
    "\n",
    "for step in tqdm(range(n_steps)): # \n",
    "    \n",
    "    # sample training batch\n",
    "    train_batch = sampler.sample(actor=actor,\n",
    "                                 n_transitions=40000,\n",
    "                                 max_path_length=None, \n",
    "                                 reset_when_not_done=False,\n",
    "                                 train_sampling=True)\n",
    "    \n",
    "    # update critic\n",
    "    critic_logs = critic.update(train_batch)\n",
    "        \n",
    "    # update actor\n",
    "    actor_logs = actor.update(train_batch)\n",
    "     \n",
    "    # evaluating the policy\n",
    "    eval_batch = eval_sampler.sample_n_episodes(actor=actor,\n",
    "                                                n_episodes=5,\n",
    "                                                max_path_length=None, \n",
    "                                                train_sampling=False)\n",
    "    # Logging\n",
    "    train_logs = train_batch.create_logs(prefix='train')\n",
    "    eval_logs = eval_batch.create_logs(prefix='eval')\n",
    "    \n",
    "    logs = {**critic_logs,\n",
    "            **actor_logs,\n",
    "            **train_logs,\n",
    "            **eval_logs}\n",
    "    all_logs.append(logs)\n",
    "\n",
    "    for key, value in logs.items():\n",
    "        logger.add_scalar(key, value, step)\n",
    "    \n",
    "logger.flush()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The tensorboard extension is already loaded. To reload it, use:\n",
      "  %reload_ext tensorboard\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Reusing TensorBoard on port 6006 (pid 6546), started 2 days, 3:11:01 ago. (Use '!kill 6546' to kill it.)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "\n",
       "      <iframe id=\"tensorboard-frame-a375119d8376959e\" width=\"100%\" height=\"800\" frameborder=\"0\">\n",
       "      </iframe>\n",
       "      <script>\n",
       "        (function() {\n",
       "          const frame = document.getElementById(\"tensorboard-frame-a375119d8376959e\");\n",
       "          const url = new URL(\"/\", window.location);\n",
       "          const port = 6006;\n",
       "          if (port) {\n",
       "            url.port = port;\n",
       "          }\n",
       "          frame.src = url;\n",
       "        })();\n",
       "      </script>\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%load_ext tensorboard\n",
    "%tensorboard --logdir /home/nlyubaykin/Desktop/RelaxExamples/A2C/tensorboard_logs"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Plot learning curve"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEWCAYAAABIVsEJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4XNWZ+PHvO6My6r1ZluUmN2xjjDElBExwwKQRUgikAEl2SSG/3WSTTd0NKZtsNmWTTSEbQlgSAoEkmwQSyNIJzcY24F5l2ZatZvWukUZzfn/ce0cjaUYaSTMaWXo/zzOPR/feuffMSJ73nvOeIsYYlFJKqYlyxbsASimlzk4aQJRSSk2KBhCllFKTogFEKaXUpGgAUUopNSkaQJRSSk2KBhAVcyLyVxG5Od7lmC4iUiQiz4lIp4h8L97lORuJiBGRpfEuhxqbBpBZTEROiMjmeJfDGHONMeaX8S7HNLoVaAIyjTGfnurJROQWEXlh6sWautn4xS4i3xWRo3bAPyQiN8W7TGcLDSBqSkQkId5lmKoYvIdy4ICZxCjdmfp5ztRyjTTJcnYDbwWygJuB/xKRS6JasFlKA8gcJSJvEZFdItImIi+JyNqgfZ8XkWP2HdkBEbkuaN8tIvKiiHxfRJqBrzh3yPadXKuIHBeRa4Je86yI/F3Q68c6dlFQ88+TIvITEfn1GO/jWvt9dNhl3mJvH1b7EpGvOOcRkYX2nfSHRaQaeNpuZvvEiHPvFpF32M9XiMgTItIiIodF5Pow5bkH60vosyLSJSKbRSRZRH4gIrX24wcikmwfv0lETovI50SkHvif8X97w64Xyfu8WUSqRaRJRL4UdOxGEdlq/w3UiciPRSQpaL8RkdtE5ChwdJxyLBGRp0Wk2b7OfSKSPaKcnxGRPSLSLiIPiognaP8/22WoFZEPjTh3sv33Ui0iDSLy3yKSEunnZ+/7/Yht/yUiPwQwxtxujDlkjPEbY14GngcuHvuTV6ABZE4SkfOAu4GPAHnAz4CHnS814Bjweqw7sq8CvxaRkqBTXAhUAUXAN4K2HQbygW8DvxARCVOEsY69H9hul+srwAfGeB8bgV8B/wxkA5cBJ8Z7/0EuB1YCVwO/AW4MOvcqrJrEIyKSBjxhl60QuAG4wz5mGGPMLcB9wLeNMenGmCeBLwEXAeuAc4GNwL8EvawYyLWvd+sEyh+pS4HlwJXAl0Vkpb19EPgU1u/hYnv/x0e89u1Yv69R73UEAf4dmIf1mZZh/f6CXQ9sARYBa4FbAOyg/xngjUAFMLLZ9VvAMqzPbylQCnw5aP94n98DwJtEJMO+ntsuy/2j3oQVmC4A9o/zfhWAMUYfs/SB9WW6OcT2nwJfH7HtMHB5mPPsAq61n98CVI/YfwtQGfRzKmCAYvvnZ4G/G+9YYAHgA1KD9v8a+HWYcv0M+H4k7x3ry+zX9vOF9jUXB+3PwGrKKLd//gZwt/38PcDzIa59e5hr3wP8W9DPx4A3Bf18NXDCfr4J6Ac8Y/webwFemML7nB+0fztwQ5hzfRL4Y9DPBnjDiGMMsDSCv723A6+NKOf7g37+NvDf9vO7gW8F7VvmXAcrMHUDS4L2Xwwcj/Tzs497AbjJfv5G4FiY434J/B8gk/k/N9ceWgOZm8qBT9tNF20i0oZ1xzgPQERukqHmrTZgNdZdquNUiHPWO0+MMT320/Qw1w937DygJWhbuGs5yrC+nCcrcG5jTCfwCFbtAqzayH3283LgwhGf1/uwgl4k5gEng34+aW9zNBpj+iZR/kjVBz3vwf69iMgyEfmLiNSLSAfwTYb/nmHszz9ArJ5nD4hIjX2uX4c4V8hyYH0WwdcJ/qwKsG4yXgn67P/P3u4Y9vnZzZFd9uN99ub7GaphvpfQtY/vYP2tX2/saKLGpgFkbjoFfMMYkx30SDXG/EZEyoGfA58A8owx2cA+rDtBR6z+c9UBuSKSGrStbIzjTwFLwuzrxvricYT6sh/5Pn4D3CgiFwMe4Jmg6/xtxOeVboz52BhlC1aLFYQcC+xt4coxEZG8z3B+ChwCKowxmcAXGf57nkjZvmkfu8Y+1/tDnCucOob/nhcEPW8CeoFzgj77LGNM8M3JsDIaq9dfuv1wbgJ+B2wSkfnAdYwIICLyVeAa4CpjTEeE5Z7zNIDMfoki4gl6JGAFiI+KyIViSRORN9ttxGlY/yEbAUTkg1h3ZTFnjDkJ7MRKzCfZX+RvHeMlvwA+KCJXiohLREpFZIW9bxdwg4gkisgG4F0RFOFRrC/6rwEPGmP89va/AMtE5AP2+RJF5IKgXMJ4fgP8i4gUiEg+Vvt92I4BYciI36OTgJ7M+3RkAB1Al/25RRoQk0aUxW2fqwtoF5FSrLxUpH4L3CIiq+ybh9udHfbv4OfA90WkEMD+PV89gfNjjGnEakr9H6zmr4POPhH5AlatZLMxpnki553rNIDMfo9i3cE5j68YY3YCfw/8GGgFKrETmsaYA8D3gK1AA7AGeHEay/s+rDbuZuDfgAcBb6gDjTHbgQ8C3wfagb8xdKf/r1i1k1asjgCjmixCnM8L/AEriXt/0PZO4Cqs5q1arKaY/wCSQ5wmlH/DCox7gL3Aq/a2ibiE4b/HXvtmYMLvM8hnsL44O7G+pB+M8HX7R5Tlg/a112P9Hh7B+hwjYoz5K/AD4Gmsv8WnRxzyOXv7Nrt57EmsTgETdT8jfre2b2LVeiqDmr6+OInzzzmiTX1qJhORB4FDxpjbxz1YKTWttAaiZhS7aWiJ3SS1BbgW+FO8y6WUGu2sGF2q5pRirOaPPOA08DFjzGvxLZJSKhRtwlJKKTUp2oSllFJqUmZ1E1Z+fr5ZuHBhvIuhlFJnlVdeeaXJGFMw3nGzOoAsXLiQnTt3xrsYSil1VhGRk+MfpU1YSimlJkkDiFJKqUnRAKKUUmpSNIAopZSaFA0gSimlJkUDiFJKqUnRAKKUUmpSNIAopdQE7atp5+UqXTpEA4hSSk3Q5/53D1/44954FyPuNIAopdQEnG7tYX9tB3VtfcRyMtoznbE9fzRoAFFKqQl44kADAL0Dg3T0+aJ6bt+gn0f31nH9z7ay8RtPcedzVVE9f7RpAFFKqQl4fH9D4Hl9e1/UztveO8Abv/8cH7/vVWrbellRnMFPnqmkvWcgateINg0gSikVobaefrafaOHCRbkA1HdEL4BsP97C8aZuvnndGv72z1fw/feso9Pr46d/Oxa1a0SbBhCllIrQ04fOMOg33HTxQgDq23ujdu69p9twCVx3Xilul7CyJJO3ryvlf148HtWaTjRpAFFKqQg9vr+B4kwPm1cVAlDf7g17bGffwIS6+u6taWdZUQYpSe7Atk9tXobfGH749NHJFzqGNIAopVQE+gYG+duRRt64qojkBDf56UnUd4Svgdz9wgnee9fLdHnHT7QbY9hb087q0qxh2xfkpXLjxgU8uOMUzx9tpN/nB6Cn38eDO6q5/mdbue/liJbuiIlZvaCUUkpFywtHm+gdGOSqc4oAKM7yUDdG09K+2nYG/YbmLi/pyWN/1dZ39NHU1c+aEQEE4BNvWMrDu2v5wC+240l0sbY0m4N1HXR6fbjECj7vu7B8am9ukrQGopSadU639kTtXH6/4dXqVu56oYoMTwIXLsoDoDjTM2Zu4mBdBwDN3f3jXmPP6XYA1swfHUAKMzw8+5lN/Pf713PjxgV4B/1sXlXE7z56Me+7sJyDdZ34/cPHi/z7owf50jQMdNQaiFIqLvx+w/3bqynLTeXyZeMuvx2xV0628M6fbuWv//h6VpZkTulc97x4nJ89V0Vdex+JbuGTm5eRlGDddxdnedh5sjXk6zr6BjjdajVvNXeNH0D21bTjdgmrwpQ3OzWJLatL2LK6ZNj2yjNd3LvtJKdaeyjPSwtsf+JgA4vz0yN6j1OhAUQpNe26vD4+/dtdPLa/gWVF6Vy+7PKonbu6xap9HKrvmFIA2VfTzlf/coALynP57JblvGFFEVkpiYH9JVkptPUM0DcwiCfRPey1h+o6A89busMn2h17TrdTUZg+6jzjcQLOwbqOQABp7xmgqrGbd66fP6FzTUbcmrBEpExEnhGRAyKyX0T+0d6eKyJPiMhR+98ce7uIyA9FpFJE9ojI+niVXam5rKW7n6/9+QAdfWMPcNt9qi1kAvlkczfvuONFnjjQwPnlORxp6OJMZ/S6qbZ0D9jXmXwzljGGr/55P7mpSdx1ywauO2/+sOABUJTpAUIPJnSarwCaxqmBGGPYV9MeMv8xnuXFGbgEDtQOXW/36TYA1pVlT/h8ExXPHIgP+LQxZhVwEXCbiKwCPg88ZYypAJ6yfwa4BqiwH7cCP53+Iiulvvbn/dz94nGePngm7DFdXh/v/OlL/Hrb6B5CH/v1qzR0ePnVhy7k9reuAmDrsejNbNtq5xyqxwkgTx9qCJsreWRvHTtOtPKZq5eT6UkMeUxJlh1AQgwmPFTfQXZqIqlJblrGyYHUtvfR3N3P2hD5j/F4Et0sLkjnQFDA2nWqDZHQ+ZRoi1sAMcbUGWNetZ93AgeBUuBa4Jf2Yb8E3m4/vxb4lbFsA7JFpASl1LR54WgTf9pVC8Br1aHb/wFONHXj8xvq2kZ3cz3e1M31G+ZzaUU+58zLItOTwIuVTRMuS1OXl9/uODVqe0uP9YV9ork77GurGrv40D07uer7z3HPi8eHJaH7Bgb590cPsbIkk+s3lIU9x1g1kAN1nawsziQvPYnmrrGbsPbaCfSRXXgjtaokk4NBTWa7TrWxpCA9bOCLphmRAxGRhcB5wMtAkTGmzt5VDxTZz0uB4L+W0/a2uqBtiMitWDUUFixYELMyKzXX9A0M8i9/2svCvFRy0pLYdaot7LHHm6wv76YRd9/dXh+9A4PkpScD4HYJFy/J48XKZowxiEjE5bnzuSrufK6KTSsKKMzwBLYHaiAt4WsgzxxuBGBNaRZf+fMB/rKnjresLSE1OYHXqtuoaevle9efi9sVvjzFdg1kZFfeQb/hcH0H791YTs/A4Li9sPbWtJFgjzyfjFXzMnl4dy1tPf1kpSSy+1QbV6wonNS5Jiru3XhFJB34X+CTxpiO4H3Gmst4QvMZG2PuNMZsMMZsKCiIXs8Opea6nzxTyYnmHv7t7Wu4cFEeB+o66BsYDHnsCSeAdA6/+3Z6JOXbAQTgdUvzqWnrHfMLP5RnD5+xrzH8C9ppMmrq6g87iO/Zw2dYWpjOA7dexH9efy7HGrv4yp8P8Nnf7+E326t585oSLlqcN+b105MTyEhOoGFEE9aJ5m76BvysLMkgPy1p3F5Ye2s6qCjKmHAC3eEk0g/UdXC6tZfm7v5pyX9AnGsgIpKIFTzuM8b8wd7cICIlxpg6u4nKaWitAYLrk/PtbUqpGKhq7OK7jx+mb8CPz2/YeqyJ684r5dKKfLq8PgYGDQfqOli/IGfUa4/bzUcj776b7B5JeelJgW2vW5oPwIuVzcO6oo6ltq2XIw1d1jlHNBG19vST4BJ8fkN1cw+r5g2/s+/2+ni5qoWbLylHRHjH+vm89dx5dPb5AjWkBbmpEZXDGkw4vJnOSaCvLMlk+/EW9td2hHopMJRAf+PKorDHjMepuRyo7Qgk7KcrgMSzF5YAvwAOGmP+M2jXw8DN9vObgYeCtt9k98a6CGgPaupSSkXZb3ee5rH9DZzp7KO9d4BNywv50ptXAnDeAusLald16GYspwYysv3fqZHkpw3VQBbnp1Gc6eHFY5HnQZ61m6AAmrtHBpCBQNCobhmdB3npWDP9g36uWD7UzJPodpGblkRZbirLJlAbKM7yUN8x/PqH6jpxu4SlhenkpSfT3O0NuTCUMYYXKpto6e5n9RQS3gUZyRRkJHOwrpNd1W0kJ7hYXpwx6fNNRDxrIK8DPgDsFZFd9rYvAt8CfisiHwZOAtfb+x4F3gRUAj3AB6e3uErNLduqmjmvLJvff+ySUfuKMj2UZHl4LUwe5ITdA6q1ZwDfoJ8Et3Wv6tRI8jOGaiAiwiVL83j2cCN+v8E1Rt7B8ezhM+SlJdHc3T+sCcsYQ2t3P1vOKWbP6faQXXmfOXyGtCQ3Gxbmjnud8RRnejjS0Dhs28G6DpYUpOFJdJOXlsTAoKHT6wsktY0xPLjjFL/cepKDdR1kpSSyaYoDKVeVZHKgroOURBdrSrNIdE9P3SCevbBeMMaIMWatMWad/XjUGNNsjLnSGFNhjNlsjGmxjzfGmNuMMUuMMWuMMTvjVXalZrsur4+9Ne1j5gHWlWWz69TonljtvQO0dPcHmoGCu7E6NZDctKRhr3ndknxauvs5WB++ucfR7/PzYmUTW1YXk5TgGtaE1en14fMbyvNSyUlNDAQyhzGGZw+d4dKK/MCI8qkoyfLQ2OnFN+gPbDtYNzSA0WmqC86DPHP4DJ//w16MMXzjutW89Pk3UBZhk1k4K0syqTzTyb7ajmlrvoIZkERXSs08O060MOg34waQUy29o5qpnOarDeVWbiR4IF1zdz+ZngSSE4Y3ETl5kJcqxx8PsvNEC939g2xaXkhBevKw8zs9sHJSk1iQlzaqCetIQxe17X3Dmq+moijLg99Ao/0ZtPX0U9veFwggTqAMHo3u5G5+a89llTbORIuRWDUvk4FBQ7/Pz7oFGkCUUnG0raqZRLdwfvnoBLnjPDt5PrI7rzP+wmkiCs5RNHV5h/XAchRneVhckMZLEeRBnj3SSJLbxSVL8shLTxpWA3FqO7lpSSzMSx3VhPWM3XNrU5QCSMmIrrzOeIwVdg7Cea/BQa66pYfs1MSojtMInkPr3PkaQJRScbStqoV1ZdnDFjcaaU1pFm6X8NqIRPrxpm5EhhLtwV/wTV3eYT2wgl1QnsuuU20hE87Bnj18hgsW5ZCWnEC+naR2tNqDCHPSkijPTaW2rTewhgbAM4fOsLIkMzCGY6qcwYQNgQBiNcGtGlUDGQogp1p6Iu7lFalF+Wl4El3kpycxPyclquceiwYQpdQwnX0D7Bsn/wGQkuRmeVHG6BpIUzfzslKYl219kQW3/zd39YesgQCsLcuitWdoFttQnO67m5ZZNYi8tKRhSXRnHqxcuwnLb6DGHg3f3jPAzpOtXLE8euPDSrKs91jX3ke318fdLx5naWE6BRnWewwXQMpyohtA3C7hwkV5XFZRMKHBmFOlAUSpOa63fzDwJQuw80TruPkPx7oF2ew+1TZsKpDjzT0syk8j05NAoluGNd+MVQNZW2rVWJzJAEN56pDTBGUFgfyM4d1kAzmQtETK86wv6ZN2k9rDe2oZ9Bu2rC4e931FKic1kaQEFw0dfXzrr4eoaevlW+9YE/gS9yS6SU9OCNTCBv2GmrbeKSfNQ/nFzRv4zrvPjfp5x6IBRKk57sfPHOWK7z7LHvuL28l/hBogONK6smw6vT6qmqzEsDGG441dLMxPRUTIS0sOJNl9g35aewbC1kCWF2eQ5HYF5oYayTfo539eOM6K4gyWFlprXTjdZDt6rRHnLfYgwvTkBMpznQBi5UF+t/MUK4ozJjXrbTgiQnGmh8cPNHDvtpN86HWLRnUPzktPCtRA6jv6GBg0lOVGv5kpwe0ac+qVWNAAotQcd6C2g36fn4/c+wqNnV62VTWPm/9wXGB/WT5lz8zb2jNAR5+PhfaI8vyMoSS3M8lhXpgAkpTgYmVJRtgayEO7aqlq6uaTm5cF7vCdpiJnhHtbTz85aUmICAUZyaQkujnZ3MPBug72nG7n+g1lUW/iKc7ycLypm4V5qXzmquWj9ucGTWdyyp6uJdo5kHjRAKLUHFfV1M2qkkxae/r5yL072VvTzsURNF+BlbzduDCXX798kkG/CUyiuCjfCiB5acmBwYNOriI/LXQTFsDa+dnsq+kYtUSrb9DPj54+yqqSTK4+Z2jajzx7RLszvqSlu5/cVOv8IsKC3FSqW7p5cMcpktwurjuvNKL3NRElWR5E4NvvOjdk0A3+DJwAEu0cSLxoAFFqFnt4dy1/2VMbdr/XN8iplh42ryzkP965ller2/AbIsp/OG6+ZCGnWnp59vCZwBiQhU4ASR+6+3Z6S+VnhK6BgLWGRZfXR1XT8PEbf3ythhPNPXxyc8WwGoQzot3Js7R2D5CTNtQ9tjwvlaNnuvjTrhreeE4ROWMEr8n62KYl3PHe9WxcFHpke17a0JTup1p6cAmBDgZnOw0gSs1idzxTyZcf2j+sK2uwUy09+A0sKkjj2nWlfHzTEvLSkgJjPCJx1TlFFGUm88utJznR3I1Lhu6w89OTaeyyktxOU1beGF/izhiGvTVDzVgDg35+9HQlq0szeeOq4ZMOOjUQJzi19PQPG+Vebo8FaesZ4D1jrO0xFSuKM7lmTfiliZwciDGGU629lGSlRGUU/EwwO96FUiqkuvY+Wrr7AwPoRjrWaN3pL863ktKf3bKCl794ZUT5D0ei28X7LiznuSONPHu4kfk5qYEvyPz0JPp9frq8vqGp3MeogSwtTCcl0c3uU0OJ9D++WkN1Sw+fCsp9OHLTknDJUBNWa3c/OalDAWSBnYuZl+UJjHafbrlpSfj8VqK/uqVnWsdpxJoGEKVmqd7+Qdp7rXERv3/ldMhjqpwAUjA0jXrCJCbiu2FjGYluYW9Ne6D5CoJqCF39NHX1k+R2kTHG1B1ul7C6NJO9NVYA6RsY5AdPHuHcsmzeEGKRJLdLyE1Loqm7H7/f0DqyBmInq9+1oWzaeyg5AvNhdXtjMogwnjSAKDVLOWt1z8vy8MyhMyGXVq1q7KIgI5mMKU6rUZjh4U12M86ivKEvyOAvT2cMyHi9oNbOz2Z/bTu+QT/3bj1JbXsfn9uyPOzr8tKSaer00tE3gN9AdlANZOOiXP7+9Yu45ZKFU3p/U+EE0Zq2Xs50emMyBiReNIAoNUs5Cx3detlifH7DQ7tGJ9OrmrpZnB/ZIk7juenihQCBMRowNBdUY2c/zWHmwRpp7fws+gb8vFrdxk+ereSyZQVcsiR881N+hjWt+9A8WEPB0JPo5ktvXjVq9t/p5Fx7jz2+RWsgSqkZr96en+ny5YWsKc0K2Yx1vKmbxQXpo7ZPxvnlOfzm7y/iXecPJaudgNHc7aW5uz/sKPRga+1E+md/v5u2ngE+e/XosRXB8tKSaeryDs2DlRq/YBGK8xk4c4bFYhBhvGgAUWqWcpqwijM9vOv8+Ryo6+BA0PKqbT3WXXu0aiAAFy/JG5aAd+6+m7v6aeqMrAZSnptKhieBE809vO3ceaweZ+R4frrVhNXqzIMVx9pGKE63YmfOMG3CUkrNePXtfWSlJJKS5OZt584j0S3876tDtZBjIRLo0ZaU4CLTY80F1RRhDcTlEtbOzyLBJXz6qmXjHp+XnkR3/yC1dpPdTKuBJCe4ybA/A0+ii4IIgujZQgOIUrNUXXtfYL2KnLQkrlxRxEO7agKr51U1WvNXRasJK5z8jGRONPfQ7/MPWwt9LP989Qp+/N7zKM8bP7g5X8hH7YWaZloNBIbGvpTlpE7rbLmxpgFEqVmqvr1v2LoX160vpamrn+ePWos2VTV1k+gWymI8LiE/LZnD9lK1wWuhj2VdWTZbVocfnBfMqdUcPdNJUoKL1AmMYZkuzvxfs6n5CuIcQETkbhE5IyL7grblisgTInLU/jfH3i4i8kMRqRSRPSKyPn4lV2rmq2vvozhzKIBcsbyQ7NRE/vhaDQDHG7tZkJs6qXEfE5GXnkRDhzMKPfrNN/lBNZCc1MQZeYfv1IpmUw8siH8N5B5gy4htnweeMsZUAE/ZPwNcA1TYj1uBn05TGZU66/T7/DR3e4fVQJISXLxlbQmPH6i355vqinnzFTAscR5JEn3C589wenr1z7j8hyPfriXNplHoEOcAYox5DmgZsfla4Jf2818Cbw/a/itj2QZki0hkdVyl5pgznX0YM7Rmt+O680rpG/DzyJ5aTjT3RLUHVjjBifP8CJLoEz5/UM5jJuY/QGsg06nIGFNnP68HnNnTSoFTQcedtrcNIyK3ishOEdnZ2NgY25IqNUM5Y0CKs4bf8a5fkEN5Xio/ffYY/T5/THtgOYLX/4jFbLieRHdgepRYnD8anKY7zYFMI2OtU2nGPXD4a+40xmwwxmwoKIje2sdKnU3q7AAysgYiIrx9XSkn7FX6pqUJy/5Sz0lNJDFG+RanlpM7Q5uwNq8s4gMXlQ8bpT8bzMQA0uA0Tdn/OtOI1gDB8zHPt7cppUZwaiBFmZ5R+4IXVZqOJiwnRxFuJcKoXMM+90ytgSzIS+Xrb18dswAaLzPx3TwM3Gw/vxl4KGj7TXZvrIuA9qCmLqVUkPqOPlKT3GR6Rs98uzA/jfULsslKSZyWnIGTo4hF/iNwjUANZGqTQqqJCT+v8jQQkd8Am4B8ETkN3A58C/itiHwYOAlcbx/+KPAmoBLoAT447QVW6izhjAEJ16X1G9etobatd1q6vDo1j7lcA5mt4hpAjDE3htl1ZYhjDXBbbEuk1OxQ1947Kv8RbGVJJitLMqelLJmehJhP4eEEp5najXe2imsAUUrFRn17HxePMQX6dBIRfvaBDTFNIBc4TVhaA5lWGkCUmmUG/YaGTi/FWTNn0r7Ll8W2R+TFS/J4fUX+tHRLVkM0gCg1yzR3eRn0m1FjQGazpYUZ3PvhC+NdjDlnJvbCUkpNQWAMSIguvEpFkwYQpWJo0G/442unuev5qmm7Zl1gFLoGEBVb2oSlVAwYY3jy4Bm+89ghjtjrVNy4cQFpybH/L1dvL6w0Vi8spaJBayBKxcAdzx7j73+1E9+g4foN8wGoaeudlmvXdfSR5HZpjyQVcxpAlIqBR/fWcX55Do9/6jJu2LgAgNOtPdNy7Yb2PoqykmfkuhhqdtEAolSUtfcMcKCug8uXFZDgdgXWgDjdOj01kNq2Pkoy504PLBU/GkCUirLtJ1owBi5anAdYa3YnJ7hiGkCMMbx0rIkb79zG9hMtrCzJiNm1lHJoEl2pKNt6rJnkBBfnlmUB1kjs0pyUmDVh9fv8fOTenTxzuJHCjGT+9S2reN+FC2JyLaWCaQBRKsq2VTVzfnkOyQnuwLb5Oakxq4F857FDPHPaL4wrAAAgAElEQVS4kc9fs4JbLlmIJ9E9/ouUigJtwlIqitp6+jlY3xFovnKUZqfEJIA8eaCBnz9/nJsuLuejly/R4KGmlQYQpaJo+/Hh+Q/H/JwUWrr76fb6onatmrZePv273awuzeSLb1oZtfMqFSkNIEpF0baqFjyJQ/kPh9MTK1pjQfx+wz/85jUG/YYf37heax4qLjSAKBVFW0PkP8DKgQDURKkZ6/evnuaVk6187dpzWDgNy9IqFYoGEKWipK2nn0P1HVy0KG/UvrLAWJCp98Tq8vr4zmOHWb8ge9j65kpNNw0gSkXJy07+Y8noAJKfnkxSlMaC3PFMJY2dXr781nN0tLmKq7MugIjIFhE5LCKVIvL5eJdHKcfWY814El2snZ81ap/LJcyPQk+sUy093PXCcd5xXinryrKndC6lpuqsCiAi4gZ+AlwDrAJuFJFV8S2VUtA3MMjDu2t5fUXBqPyHY6qDCY0xfPPRg7hF+Octyyd9HqWi5awKIMBGoNIYU2WM6QceAK6Nc5mU4s+7a2np7ueWSxaGPWYqgwlPNnfzgV9s56/76vn4piWUzKHVBtXMdbaNRC8FTgX9fBrQdSzVtDre1E1ygot52daXuDGGX249QUVhOpeEyH845uek0NzdT0+/j9SkBIwxtHT3k5c+9trldz1fxXcfP0yiy8XX376a923UaUrUzHC21UDGJSK3ishOEdnZ2NgY7+KoWaZvYJB3//dWrv3Ji9TbK/+9Wt3KvpoObr5k4ZhJ7cBYELsW8vtXTnPhN5/iQG1H2NfsPtXGvz1ykNctyeeJf7qcD1xUjsuliXM1M5xtAaQGKAv6eb69LcAYc6cxZoMxZkNBQcG0Fk7Nfg/vqqWpy0tbTz8f+fUr9A0M8j8vniDDkzBul1pnLMjp1l6MMdz1/HF8fsN//+1Y2Nf87UgjIvDtd63VJWrVjDNuABGRAhH5oojcKSJ3O4/pKFwIO4AKEVkkIknADcDDcSqLOouc6ejjpru3c6ajb9LnMMZw1wtVrCzJ5Ec3nsfuU2384wOv8X/76nnPhrJxl6sNHguytaqZww2dLM5P4y97aqluDp1cf+5II6vnZY3bzKVUPERSA3kIyAKeBB4Jekw7Y4wP+ATwGHAQ+K0xZn88yqLOLi8da+a5I408urdu2HbfoJ8fPnWUow2d457juaNNHGno4u8uXcSW1SX8wxuW8tj+BgaN4aaLF477+sBYkLZe7nnxBLlpSdzzwY0kuFz8/PmqUcd39A3w2qk2LluWH/H7VGo6RZJETzXGfC7mJYmQMeZR4NF4l2OuOtPRx/3bq7n1ssWkJp09fTCONXYB8PzRJm553aLA9m1VLfznE0e4d9tJ/vCxSyjLTQ17jruer6IwI5m3njsPgE9uXkZ9Rx8piW4W5IV/ncMZC/JyVQt7Trfx0cuXsCAvlXesL+W3O0/xD1dWUJAxVNN4qbKJQb/hsgptilUzUyQ1kL+IyJtiXhJ1Vnh4dy0/ePIot9y9I6ozy07F0YZOfvTUUYwxYY+pauwGrLU6+n3+wPYnDzaQnODCOzDITXdvp7nLG/L1h+s7ef5oEzdfspCkBOu/jcslfPtd5/LVa1dHXNbSnBR2nWpDRHj/ReUA3HrZYvoH/dzz0vFhx/7tSBPpyQmsL8+J+PxKTadIAsg/YgWRXhHpEJFOEQnfbUTNaofrO0lOcPFKdSs3372drhkQRO57uZrvPXGE403dYY851thFSqKb7v5BXqtuBaycxpMHG7h0aT5333IBtW29fPCeHXT0DQx77aDf8ONnKklJdE95pT+nJ9bV5xQFugEvLkhnyznF/GrrSdp7BwJle+5IIxcvySPRfbb1dVFzxZh/mWL1STzHGOMyxqQYYzKNMRnGmMxpKp+aYY40dLJhYQ4/vOE8XjvVxvt+vo17t57gxcomznROPkE9FQfrrPuZbVUtIfcP+g3Hm7q5dt08XAIvVDYBcKShi9OtvWxeVcSGhbnc8b717K/t4PJvP8Mdz1bS5fXxUmUTb/nRC/x5dy0fuLic7NSkKZXV6Yl1yyWLhm2/7Yql9PQPcvtD+wCoauqmpq2Xy5Zp85WaucZsxDbGGBF5BFgzTeWZM9p6+jnS0MXGRbnxLkrE/H7DkYYubthYxpvXluB2wWd/v4d/fcjqxyACf/z466Z1jiZjTFAAaea9IWoItW29eH1+1pVlc6Shk+eONvHpq5bz5MEGAK5cUWj9u7KIP3zsEr7/5BG+/X+H+fHTlfT0D1KancKPbjyPt6wtmXJ533NBGaXZKVywcHiz1OrSLP7hDRV8/8kjbFpeSFtPPwCXa/5DzWCRZEFfFZELjDE7Yl6aOeS+l6v53uOH2XX7VWR6EuNdnIicbu2ld2CQ5UUZAGxZXcLV5xTT0OFlf207H/7lTl452TqtAaS2vY+OPh9Jbhdbq5oxxowazOck0BcXpHNpRQE/fvoo7T0DPHmwgbXzsyjMHBpfcW5ZNvd8cCOvVbdy79aTLClM58OXLoragk356cm8Pcx4kduuWMLzRxv51z/tY1FBGgvzUiNKzisVL5E0rl4IbBWRYyKyR0T2isieWBdstmvt7sdv4Ej9+N1HZ4rDdlfXZcUZgW0iQnGWhytXFpGfnsTh+ulNjx20R3G/bd08Gju9VIXIgxyzE+hLCtJ4fUU+fgMP76ll16k2Nq8sCnne8xbk8J/vWcdtVyydttX+Etwuvv+edQDsOd2uzVdqxoskgFwNLAHeALwVeIv9r5oCJ/l8OILxBzPFEbusFYXpIfcvL87g0DQHxEN2wHImMdxW1TzqmKrGLrJSEslNS2JdWTbpyQl8/4kjGANXriyczuKOqyw3lX+7zurVFS64KTVTRBJATJiHmoJOJ4CcTTWQ+k5Ks1PICNPktqI4kyMNnQz6p+/P42BdJ+V5qZwzL5OizOSQifRjjV0sKUhDREh0u7hocR4t3f3My/KwqmTm9Qe5dl0p275wpdZA1IwXSQB5BPiL/e9TQBXw11gWai7o6rMCyHTfsU/FkYZOVgQ1X420vDiDvgE/J5vDd6eNtoN1HawozkBEuGhxHluPNY8aD1LV2M3igqFakzOy+8qVRTN2RT+d90qdDcYNIMaYNcaYtfa/FVhrcmyNfdFmN6cJ60hD55gD4GaKgUE/xxq7huU/RlpZbN3Nx6pW9eSBBu4KmvKjp9/H8eZuVtq1iIsX59HU5Q3kPAA6+wY40+llSVAAuXJlEQUZ4ZPZSqnITHiEkjHmVXQNjilzRnG39VhfcDPdiaZuBgZNoAdWKBVF6bgEDsYggLR29/Pp3+3mm48epK7dmg79cH0nxhAIIBctttbiCM6DOCPQFxekBbaVZqew40ubOV9HeCs1JZHMxvtPQY/PiMj9QO00lG1W6+zzUWqPRD4bmrECPbDGCCCeRDcL89Ni0hPrP584QmffAH4Dv9t5Ghj63Jw8RnleKiVZHrYGBRCnC29wDUQpFR2R1EAygh7JWLkQXUZ2irq8vsAd8NnQlfdIfSdulwy7kw9lRQx6Yh2q7+C+l0/y/ovKuXRpPg/uOIXfbw0gTE9OCEwP4uRBXq4ayoNUNXbjdgkLxpgkUSk1OZEEkAPGmK/aj28YY+5Du/FOiTGGLq+PBbmpFGQknxU1kEP1nSzMSx13TMSK4kyqW3qiNtGiMYavPnyAzJRE/umNy3jPBWXUtPXyQmXTsAS6Y9PyApq6+vn9K1Yt5VhjF+W5qYEJEJVS0RPJSPQvAL+LYJuKUN+An0G/IS05gRXFGRxumPlzUx5p6GTVvPG7vC4vzsAY6/jzFkwux/Cdxw7xtyONLClIJy05ga1VzXz92nPITk3iqnOKyElN5IEd1Ryq6xyVCH/r2nnc93I1X/vzAV63NN/ugTV2rUkpNTlhb8tE5BoR+RFQKiI/DHrcA8R/CtazmNMDK92TwLKiDI42dE3r2IlIOc1Avf2DnGzpGTP/4XC6+U6lJ9afXqulocPLzhOt3P9yNatKMrlxozXHVXKCm3esn89f99XT6fUFEugOl0v47rvOZdAYPvv7PRxv7tb8h1IxMlYNpBbYCbwNeCVoeyfwqVgWarZzAkhGcgLLizPw+qyxE4tnyBedMYavPLyfP7xawzvPn895C7IxhjF7YDnKclJJTXIHmuW8vkG+9/gRatp6reGnYnW3fef6+aQkjW4O6xsYpKatl09tXsY/bq6gp9+H2yUkBE1pfsMFZfziBWvtjJUlo8u0IC+VL75pJf/yJ2tmW62BKBUbYQOIMWY3sNvudZUALDDGHJ62ks1iziDC9OQECjOtFegO13dOWwDZfrwF36CfS5aGXir1B08e5ZdbT3Legmzuf7mae146AVjNU+NxuYRlRRmBKUZuf2g/D+w4xeKCNASr+e6RPXV89/HDvHfjAj66acmwySRP2muDL7K/9EOtelhRlMH55Tm8Wt0atkzvu3ABj+2v5/mjTTMmMCs120SSA9kCfBdIAhaJyDrga8aYt8W0ZLNYp9daNCjdk0BFYQYiVjfZa9ZMfbrwSHzpj3upaevl8U9dFlifwnHvtpP811NHuX7DfP7jnWtp7u7nwR2nqGnrpTwvsjv5lSUZ/HVfPfe/XM0DO05x2xVL+OerVwBW7WbnyVbuer6Kn/7tGD39g3zlbecEXnu8yep2u2ica33pzSt59WRr2GV1RYTvvftc7t9ePa2zAys1l0TSNeUrWKPP2wCMMbuARWO9YDwi8m4R2S8ifhHZMGLfF0SkUkQOi8jVQdu32NsqReTzU7l+vAXXQFKS3JTnpk7bnFgt3f0cPdNFT/8gX/zjvmGj4B/eXcuXH9rH5pVFfPO6NYgI+enJ3HbFUr553Rrcrsim/VhelEFbzwBffmgfly0r4J/euDywT0S4YGEuP/vABi5YmMue023DXuvMprswf+xut+sX5PB3r1885jGFmR4+uXmZruinVIxE8j9rwBjTPmLbVDO++4B3AM8FbxSRVcANwDlYNZ87RMQtIm7gJ8A1wCrgRvvYs1IgiZ5s3T0vL86YtgCy44Q12eDbzp3Hc0ca+cOrNQA8uKOaTz7wGhcszOXH7z1vWM5holbYie2SbA8/vGFd2MCzsjiDIw1d+IM6EJxo6qYgIznshI1KqZkjkias/SLyXsAtIhXAPwAvTeWixpiDQKiJ7K4FHjDGeIHjIlKJVfsBqDTGVNmve8A+9sBUyhEv3UG9sMC6Y3/iQAN9A4MxX3tix/EWkhJcfPtda6lp6+XrjxyguqWH/3rqKJctK+Bn7z9/ymVYV5bNu8+fz9+9fvGYS8AuL86ky3uSmrZeyuyBfsebusdtvlJKzQyR3Gb+P6wagRe4H2gHPhmj8pQCp4J+Pm1vC7d9FBG5VUR2isjOxsbGGBVzajpH1ECWFWfgN0PTbsTSjhMtrCvLxpPo5j/euYYe7yD/9dRRrlldzM9vOj9kz6iJ8iS6+c67zx036e7sDx5Iebyph0X5GkCUOhuMGUDspqOvGWO+ZIy5wH78izGmb7wTi8iTIrIvxCOm06AYY+40xmwwxmwoKJiZ6yl09flIdAvJ9ujoikLri7TyTGwDSLfXx77aDjYutNZhX1qYwb+/Yw0f37SEH914HskJ07PynmN5YMyI1WOrs2+Api5voAeWUmpmG7MJyxgzKCKXTubExpjNk3hZDVAW9PN8extjbD/rdHl9pCUnBJrwFuWn4XYJRxtiG0BerW5l0G+4YFFuYNs7z58f02uOJT05gbLclMDsvSearC68C7UJS6mzQiQ5kNdE5GGsqUsCCy0YY/4Qg/I8DNwvIv8JzAMqgO2AABUisggrcNwAvDcG158WXX2+QPMVQFKCi4V5qRw9E9tE+o7jLbiEGTWN+fKizEAHgiq7C68O/FPq7BBJAPEAzVhrojsMMOkAIiLXAT8CCoBHRGSXMeZqY8x+EfktVnLcB9xmjBm0X/MJ4DHADdxtjNk/2evHW5d3eAABqxnrSIwDyPYTLZwzL2vUteNpZUkGzxw+Q9/AIMebuhFBZ85V6iwx7jeJMeaD0b6oMeaPwB/D7PsG8I0Q2x8FHo12WeKhy+sjwzMigBSl88TBBry+wZjkIvp9fl6rbuN9F5ZH/dxTsbw4g0G/ofJMFyeaupmXlRLznmhKqejQEVZxEKoGsrQwnUG/CeQBom1vTTten5+Ni2ZO8xUMn3zxeJPOnKvU2UQDSBx09VlJ9GBOT6wjDbFpxtp+3BpAeMHC3HGOnF4L89JISnBxqL6D403dmkBX6iyiASQOOkM0YS0uSMMlcDRGXXl3nGhhSUEaeenJMTn/ZCW4XVQUpvPSsWY6+nw6BkSps8ikAoiIrI92QeaSkb2wwBp8tyA3lcpJJNLPdPbhG/SPuf/5o41cvqxwwueeDsuLM9hfa40F0TEgSp09JlsD+VhUSzGHDPoNvQODpCePnutpaWHGhMeCnGjq5tL/eIbf7DgV9pjfvHyKgUHD+y9aMOHyToeVxUOLQuk0JkqdPSYVQIwxfx/tgswVXSPmwQpWUZTO8aZuBsaoTYz0gyeP0O/zcyxM01e/z8+vXz7JpuUFM3ZdDGdEeoJLmJ+TEufSKKUiNW4AEZH1IR5LRGTmDCaIk+rmHlq6+yf0mqGZeEd3Va0oTMfnN5xs7h61r62nn5vu3s5r1a2BbYfqO3hody0A9e2hZ5f56746Gju93HzJwgmVczqtsFcVXJCbOqVZgJVS0yuS/613ANuAO4GfA1uxRqUfFpGrYli2Gc3vN1z/s618+/8OTeh1Q2uBjG7CctYcD9WMta2qmeeONPKhe3ZQZU+6+N3HjpCenMCa0izqO0IHkHteOsGi/DQur5iZ84IBFKQnk5eWpAl0pc4ykQSQWuA8e4LC84HzgCrgjcC3Y1m4mWxfbTv1HX209ky0BjK0GuFISwrSkTA9sfbVdOB2CS4Rbv6f7Ty+v54nDzbw0cuXsKwoI2QNZNepNl6rbuPmi8txRbgYVDyICN+9/lw+9cZl8S6KUmoCIgkgy4KnDTHGHABWOGtzzFVPHzoDWGt8T0SXdxAg5HQiKUlu5uekhA4gte1UFKbzi1suoLHTy633vkJ+ehK3XLKQkiwPjV3eUT2xfvnSCdKS3HGdMDFSVywvZHVpVryLoZSagEgCyH4R+amIXG4/7gAOiEgyMBDj8s1Yzxy21hrpGxic0OucJqyR40AcFYUZHB0xmNAYw76adlaXZrGuLJufvHc9iW7hk5uXkZacQFGWh0G/oalreG3oqYMNvGlNia7up5SKiUgS4bcAH2doEakXgc9gBY8rYlOsma2pyxtYy7vPN9EaiN2EFWZCw4rCdF6obMI36A8klBs6vDR19bN6ntXd9cqVRez68lWB0ewlmR4A6jv6KM6ynrf19NPR5wvkVZRSKtoiCSDXAD82xnwvxL7YL6E3A/3tcCPGQGl2Ct4J1kA67RrIyKlMHEsL0+n3+alq6g58+e+tsZakXzN/qIkn+PVO0Khv7wusmlLdYs2pVaYz2yqlYiSSJqy3AkdE5F4ReYt234VnDp+hICOZ9eU5E2/CGrGc7UiXLM0HhnIsAPtq2nEJrCzJDPmaoQDSG9jmBJDyPA0gSqnYGDeA2NO5L8XqunsjcExE7op1wWYq36Cf5440smlZAamJbnonkQNJTXLjDtMrqjQ7hdWlmTy+vz6wbX9tO0sK0klNCh10clOTSHQL9R3ewDatgSilYi2iUVvGmAHgr8ADwCvA22NZqJns1eo2Ovp8XLGiEE+ia8K9sLr7R8+DNdIbVxbz2qk2znRaXXP32gn0cFwuoSjTM7wG0txDXlrSjFo8Sik1u0QyEv0aEbkHOAq8E7gLKI5xuWaspw+dIcElXFqRjyfRPeEmrM4+X8gxIMGuOqcIY+Cpg2c409lHQ4eXc+aFbr5yFGd6hg0mrG7pYYE2XymlYiiS29ObgAeBjxhjvOMdPNs9e/gMGxbmkOlJJDnRjdfnxxiDSGQD9UItJjXSiuIM5uek8MSBhkB+Y804YySKszzss5PtYAWQmbT2uVJq9okkB3KjMeZPTvAQkUtF5CexL9rM09DRx6H6TjYtt6ZF9yRaH593Al15Q03lPpKIcNWqYl6obOLlKmshqFUR1kCMMQwM+qlt69W1xZVSMRVRDkREzhOR74jICeDrwMQmgBp9vu+IyCER2SMifxSR7KB9XxCRShE5LCJXB23fYm+rFJHPT+X6k/XC0SYALrV7Snnstcsn0owVSQ0ErGasfp+f+7adZHF+2riDAYuzPPQN+GnvHaCmtRe/QQOIUiqmwgYQEVkmIreLyCHgR0A1IMaYK4wxP5ridZ8AVhtj1gJHgC/Y11wF3ACcA2wB7hARt4i4gZ9gjUlZBdxoHzutXqhsIi8tiVV2d1pPohNAJlAD8Y6fAwHYUJ5DdmoinV4f50QwxUegK29HX6AHlgYQpVQsjVUDOQS8AXiLMeZSO2hMLGMchjHmcWOMz/5xG+BM1nQt8IAxxmuMOQ5UAhvtR6UxpsoY04/VG+zaaJRlAmXmhcomLlmaH5iY0GnCmmgNJCOCGkiC28WVK4oAAiPQx1ISNJgwEEA0ia6UiqGxAsg7gDrgGRH5uYhcCcRiStcPYXURBigFgpfWO21vC7d9FBG5VUR2isjOxsbGqBXycEMnjZ1eXm83X0FQDcQXWQAxxtDV5ws7Cn2kN62xOrutjyAZXpQ5FEBOtfSQlOCiKMMT0XWUUmoywgYQO3F+A7ACeAZrLqxCe2LFcdcBEZEnRWRfiMe1Qcd8CfAB9039rQTKfac99fyGgoLJrYHh9Q1yx7OV1AWNqwjkPyqCA4hTA4msCcvr8+Pzm4iasADesKKQR//h9VywMHfcYwszhpqwTjb3UJaTMqOncFdKnf3G/SYzxnQD9wP3i0gO8G7gc8Dj47xu81j7ReQW4C3AlcYYY2+uITCbE2A1bdXYz8Ntj7ozHV7+68mj7K/t4CfvXQ9Y+Y/FBWnMyx5acnUoBxJZDcSZByuSJiywemON1/vKkZTgIj89OdCEpfkPpVSsTWj9UGNMq32Hf+VULioiW4DPAm8zxvQE7XoYuEFEkkVkEVABbAd2ABUiskhEkrAS7Q9PpQxjKctN5eOblvLInjpeONqE1zfIy1Utw5qvYCiARDqdyVjroUdDcVYydXYTlgYQpVSsxWsB6h8DGcATIrJLRP4bwF646rfAAeD/gNuMMYN2wv0TwGPAQeC3wYtcxcJHLl/MgtxUbn94Hy9XtdA7MMilI5aFdbrxRjojb7c3/HK20VCcmcKh+g46vT4W5OnysEqp2IrLREnGmKVj7PsG8I0Q2x8FHo1luYJ5Et185W2r+NA9O/n8/+7B7RIuWpw74piJ5UCGpnJ3R7ewtuKsZJ48aE0WoDUQpVSsxasGclZ4w4oiNq8spLa9j3Vl2aMG8000B+I0YWXEqAZSkjWUn9EAopSKNQ0g47j9reeQmuRm88qiUfsmHkDs1QhjlANxuvKCBhClVOzpXN/jKMtN5cXPvSHkGuaBJqwI58Jy1kOP1RTrzmDCgoxkUpJi00ymlFIOrYFEICctKbA+ebBI5sKqbu7hr3vr6OwboMtrHRcqGEWDUwPR2odSajpoDWQKXC4hyT32olLfefwwf95dS1KCi7y0JNwuITkhNnHbmQ+rXAOIUmoaaA1kipITXWPXQFp6WFmSyfsvLAegojA94rVDJio9OYHXV+Rz+fLJjcBXSqmJ0BrIFKUkuvGOMRdWTWsvm1cW8uW3ruJf3rySGMWOgHs/fGFsL6CUUjYNIFPkSXTT2x86gPQNDNLU5Q1Mf6JzUymlZhNtwpoiT2L4HEhdu7VGeWnQ/FlKKTVbaACZIk+iO+x07jWt1my+pTkaQJRSs48GkCnyJLjDJtFr2qx5IrUGopSajTSATFHyGE1YNa29uGSoe61SSs0mGkCmyJM4Vg2kj6JMD4khBiEqpdTZTr/ZpsiT6MYbZiqTmrYebb5SSs1aGkCmyJMQfiBhTVuvJtCVUrOWBpApCteENeg31LX1DVsCVymlZhMNIFMUbhxIY6cXn99oE5ZSatbSADJFKYluegcGMcYM2x7owqtNWEqpWUoDyBQl24tKjUykn7YHEc7XGohSapaKSwARka+LyB4R2SUij4vIPHu7iMgPRaTS3r8+6DU3i8hR+3FzPModirMqoXdEM1ZNmxVANAeilJqt4lUD+Y4xZq0xZh3wF+DL9vZrgAr7cSvwUwARyQVuBy4ENgK3i0jOtJc6hKFVCYcn0mvbeslOTSQtRqsPKqVUvMUlgBhjOoJ+TAOcBMK1wK+MZRuQLSIlwNXAE8aYFmNMK/AEsGVaCx1GuFUJa1p7NYGulJrV4nZ7LCLfAG4C2oEr7M2lwKmgw07b28JtD3XeW7FqLyxYsCC6hQ7BacIa2ROrpq2X8ry0mF9fKaXiJWY1EBF5UkT2hXhcC2CM+ZIxpgy4D/hEtK5rjLnTGLPBGLOhoCD2K/MFmrCCaiDGGK2BKKVmvZjVQIwxmyM89D7gUawcRw1QFrRvvr2tBtg0YvuzUy5kFAzVQIYCSEevj+7+QeZrF16l1CwWr15YFUE/Xgscsp8/DNxk98a6CGg3xtQBjwFXiUiOnTy/yt4Wd0NJ9KEmrNM6jbtSag6IVw7kWyKyHPADJ4GP2tsfBd4EVAI9wAcBjDEtIvJ1YId93NeMMS3TW+TQkkMk0Z2FpLQLr1JqNotLADHGvDPMdgPcFmbf3cDdsSzXZIRqwnLGgOgodKXUbKYj0acoJWl0AKlt68WT6CIvLSlexVJKqZjTADJFngSnF9ZQDqSmrZd52SmISLyKpZRSMacBZIpCNWGd6fBSmJEcryIppdS00AAyRaEGEjZ395OfrgFEKTW7aQCZIrdLSHTLsLmwmru8GkCUUrOeBpAo8CQMrUrY7/PT0efTBLpSatbTAJSiJk4AAAqrSURBVBIFyYnuQBNWS3c/AHlaA1FKzXIaQKLAk+jCa9dAmrq8AORqDUQpNctpAIkCT6I7kANptmsg+ekaQJRSs5sGkCjwJLoCTVjNdg1Em7CUUrOdBpAoSEl009tv10C6nByI1kCUUrObBpAoCG7Caur2kuR2kaFL2SqlZjkNIFGQnBDUC6urn7z0JJ3GRCk162kAiYLgXljN3f3afKWUmhM0gESBJ3FoIGFzl5fcNE2gK6VmPw0gUeBJdAVWJGzq6idfx4AopeYADSBR4ExlYoyhudurTVhKqTlBA0gUOE1Y3f2D9A34dQyIUmpOiGsAEZFPi4gRkXz7ZxGRH4pIpYjsEZH1QcfeLCJH7cfN8Sv1aJ5EF34D9e19ADqRolJqTojbYAURKQOuAqqDNl8DVNiPC4GfAheKSC5wO7ABMMArIvKwMaZ1eksdmrMmiLMWuk7lrpSaC+JZA/k+8FmsgOC4FviVsWwDskWkBLgaeMIY02IHjSeALdNe4jCSnQDSagUQzYEopeaCuAQQEbkWqDHG7B6xqxQ4FfTzaXtbuO0zQkqgBtID6Ey8Sqm5IWZNWCLyJFAcYteXgC9iNV/F4rq3ArcCLFiwIBaXGMWTaMXhQA1Ex4EopeaAmAUQY8zmUNtFZA2wCNhtT/cxH3hVRDYCNUBZ0OHz7W01wKYR258Nc907gTsBNmzYYEIdE22ehKEcSFqSm5Qk93RcViml4mram7CMMXuNMYXGmIXGmIVYzVHrjTH1wMPATXZvrIuAdmNMHfAYcJWI5IhIDlbt5bHpLns4nqAciHbhVUrNFTNtythHgTcBlUAP8EEAY0yLiHwd2GEf9zVjTEt8ijia04RV39HHuWXZcS6NUkpNj7gHELsW4jw3wG1hjrsbuHuaijUhTg3EbzT/oZSaO3QkehQ4NRDQQYRKqblDA0gUJCcMJc11DIhSaq7QABIFThMW6FroSqm5QwNIFAQ3YeVrDUQpNUdoAImCYTUQTaIrpeYIDSBRkOh2keCy1kDXHIhSaq7QABIlTi1EA4hSaq7QABIlTh4kJ1UDiFJqbtAAEiXJCW6yUxNJdOtHqpSaG+I+En228CS68CRq7UMpNXdoAIkST6KbtCT9OJVSc4d+40XJxzctJSVJm6+UUnOHBpAoefPakngXQSmlppXeMiullJoUDSBKKaUmRQOIUkqpSdEAopRSalI0gCillJoUDSBKKaUmRQOIUkqpSdEAopRSalLEGBPvMsSMiDQCJ6dwinygKUrFOVvMxfcMc/N9z8X3DHPzfU/0PZcbYwrGO2hWB5CpEpGdxpgN8S7HdJqL7xnm5vuei+8Z5ub7jtV71iYspZRSk6IBRCml1KRoABnbnfEuQBzMxfcMc/N9z8X3DHPzfcfkPWsORCml1KRoDUQppdSkaABRSik1KRpAQhCRLSJyWEQqReTz8S5PrIhImYg8IyIHRGS/iPyjvT1XRJ4QkaP2vznxLmu0iYhbRF4Tkb/YPy8SkZft3/mDIjLrFrgXkWwR+b2IHBKRgyJy8Wz/XYvIp+y/7X0i8hsR8czG37WI3C0iZ0RkX9C2kL9bsfzQfv97RGT9ZK+rAWQEEXEDPwGuAVYBN4rIqviWKmZ8wKeNMauAi4Db7Pf6eeApY0wF8JT98/9v785DrSjDOI5/f3WzUksrSHIJC28ZFVqYSFaI7QvtmmEUlRgR7RHWHy1QYCRWFEj7glHEVUqijUqypE0ti7S9KEWzMNtTy19/vO+t6Symh3s6Nff5wMUzc94z847PvfOc952Z9y2bi4GlheWbgFtsDwG+Bc5tSa2a6zbgGdtDgWGk4y9trCUNAC4CRtjeB9gSmEA5Y/0AcFTFunqxPRpozz+TgRmN7jQSSLWRwMe2P7W9DngUOKHFdWoK2ytsL8qvfyCdUAaQjvfBXOxB4MTW1LA5JA0EjgXuycsCxgIduUgZj7kPcAhwL4DtdbbXUPJYk6bt3lZSG9ATWEEJY217HrC6YnW92J4APOTkNaCvpIbm5I4EUm0A8GVheVleV2qSBgP7Aa8D/WyvyG+tBPq1qFrNcitwJbAhL+8ErLH9W14uY8x3A74G7s9dd/dI6kWJY217OTAN+IKUOL4DFlL+WHeqF9suO8dFAglI6g3MAi6x/X3xPaf7vEtzr7ek44BVthe2ui7/sjZgf2CG7f2An6joriphrHcgfdveDegP9KK6m6dbaFZsI4FUWw4MKiwPzOtKSdJWpOTxsO3ZefVXnU3a/O+qVtWvCUYDx0v6nNQ9OZZ0baBv7uaAcsZ8GbDM9ut5uYOUUMoc68OAz2x/bXs9MJsU/7LHulO92HbZOS4SSLU3gfZ8p0YP0kW3OS2uU1Pkvv97gaW2pxfemgOclV+fBTzxb9etWWxfZXug7cGk2L5oeyIwFzg1FyvVMQPYXgl8KWnPvOpQYAkljjWp62qUpJ75d73zmEsd64J6sZ0DnJnvxhoFfFfo6tos8SR6DZKOIfWTbwncZ/vGFlepKSQdBLwMvMtf1wOuJl0HeQzYlTQc/njblRfo/vckjQGusH2cpN1JLZIdgbeAM2yvbWX9upqk4aQbB3oAnwJnk75EljbWkq4HTiPdcfgWMInU31+qWEt6BBhDGrb9K+Ba4HFqxDYn0ztI3Xk/A2fbXtDQfiOBhBBCaER0YYUQQmhIJJAQQggNiQQSQgihIZFAQgghNCQSSAghhIZEAgndkqSdJL2df1ZKWl5Y3qTRWSXdX3iuol6ZCyRN7Jpa19z+yZKGNmv7IWxM3MYbuj1J1wE/2p5WsV6kv5ENNT/4HyBpJtBh+/FW1yV0P9ECCaFA0pA8P8rDwHvALpLukrQgzytxTaHsK5KGS2qTtEbSVEmLJb0qaedc5gZJlxTKT5X0htJ8Mwfm9b0kzcr77cj7Gl6jbjfnMu9IuknSwcAxwC255TRYUrukZyUtlDRP0h75szMlzcjrP5R0dPP/N0PZtf1zkRC6naHAmZ1P50qakp/gbQPmSuqwvaTiM32Al2xPkTQdOAeYWmPbsj1S0vHANaSngS8EVto+RdIwYFHVh6R+pGSxt21L6mt7jaSnKLRAJM0FJtn+RNJo0hPHR+TNDAIOIM0D8bykIf/3J7BDa0ULJIRqn1QM7XC6pEWkE/tepInGKv1i++n8eiEwuM62Z9cocxBpaA1sLya1fCqtJg03c7ekk0ij6f6NpL6kicFmSXqbNDFa/0KRx2xvsP0BaTjv9jp1DGGTRAskhGp/npwltZNmLxyZv/HPBLap8Zl1hde/U/9va+0mlKlie72kEcDhwDjgfP5qWfxZXeAb21XdX52b+YflEDZLtEBC2LjtgR+A7/OQ2Ec2YR/zgfEAkvalRgtH0nbA9rafBC4lTf5Frtt2ALa/BVbkFgqStshdYp3G5RFY9yB1Z33UhGMJ3Ui0QELYuEWkIcDfJ41oOr8J+7gdeEjSkryvJaTZ84r6ALMlbU364ndZXv8IcKeky0lTlk4AZuQ7y3oAM4HFuexyYAHQG5icp2wOoWFxG28ILZYvzrfZ/jV3mT0HtBemXe2KfcTtvqHLRQskhNbrDbyQE4mA87oyeYTQLNECCSGE0JC4iB5CCKEhkUBCCCE0JBJICCGEhkQCCSGE0JBIICGEEBryB9tJmt+O8crsAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "eval_score = [l['avg_return/eval'] for l in all_logs]\n",
    "plt.plot(eval_score)\n",
    "plt.title(f'Learning curve for {env_name}')\n",
    "plt.xlabel('Training step')\n",
    "plt.ylabel('Avg. return')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Visualize trained actor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Simulation done, total reward: 118.843115872755\n",
      "Loading video...\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<video alt=\"test\" autoplay \n",
       "                    loop controls style=\"height: 400px;\">\n",
       "                    <source src=\"data:video/mp4;base64,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\" type=\"video/mp4\" />\n",
       "                 </video>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_actor(env=eval_env, actor=actor, nsteps=300, train_sampling=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Save Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved checkpoints for VPG...\n",
      "models_state_dict global_step n_updates optimizer scheduler\n",
      "Saved checkpoints for BootstrappedContinuous...\n",
      "models_state_dict global_step n_updates optimizer scheduler\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nlyubaykin/anaconda3/envs/relax/lib/python3.6/site-packages/torch/optim/lr_scheduler.py:200: UserWarning: Please also save or load the state of the optimzer when saving or loading the scheduler.\n",
      "  warnings.warn(SAVE_STATE_WARNING, UserWarning)\n"
     ]
    }
   ],
   "source": [
    "actor.save_checkpoint('trained_models', 'vpg_actor')\n",
    "critic.save_checkpoint('trained_models', 'bootstrappedcontinuous_critic')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "actor.load_checkpoint('trained_models', 'vpg_actor')\n",
    "critic.load_checkpoint('trained_models', 'bootstrappedcontinuous_critic')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python Relax",
   "language": "python",
   "name": "relax"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}