notebooks/setfit-optimum-intel.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "76571396-8f54-40ed-9e81-6c7531e6eaee",
   "metadata": {
    "id": "76571396-8f54-40ed-9e81-6c7531e6eaee"
   },
   "source": [
    "# Accelerate SetFit Models with 🤗 Optimum Intel"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4733f283-4736-4d9d-854e-5582280b074d",
   "metadata": {},
   "source": [
    "In this notebook, we'll show how you can use 🤗 [Optimum Intel](https://github.com/huggingface/optimum-intel) to speed up your SetFit models on Intel CPUs.\\\n",
    "[Optimum Intel](https://github.com/huggingface/optimum-intel) is an open-source library that accelerates end-to-end pipelines built with Hugging Face libraries on Intel Hardware. Optimum Intel includes several techniques to accelerate models such as low-bit quantization, model weight pruning, distillation, and an accelerated runtime.\n",
    "\n",
    "Optimizing pre-trained models can be done easily with Optimum Intel; many simple examples can be found [here](https://huggingface.co/docs/optimum/main/en/intel/optimization_inc).\n",
    "\n",
    "In our experiments, we will be using a fine-tuned SetFit model based on [`bge-small-en-v1.5`](https://huggingface.co/BAAI/bge-small-en-v1.5) [Sentence Transformer](https://www.sbert.net/) model and observe our performance increase. \\\n",
    "We'll compare the performance of the Optimum Intel optimization with two other commonly used methods:\n",
    "\n",
    " - Using PyTorch and Huggingface's Transformers library with fp32.\n",
    " - Using [Intel Extension for PyTorch](https://github.com/intel/intel-extension-for-pytorch) (IPEX) runtime with bf16 and tracing the model using torchscript.\n",
    "\n",
    "For further details, please refer to our [blog]()."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d874cea-d8d1-4fdf-b374-7772811fe7b7",
   "metadata": {},
   "source": [
    "### Prerequisites"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4bbde933-a2f1-4159-9b93-cfcf73f28c79",
   "metadata": {},
   "source": [
    "* Linux OS\n",
    "* Python 3.9+\n",
    "* An Intel Xeon Processor which supports [bfloat16](https://www.intel.com/content/www/us/en/developer/articles/technical/pytorch-on-xeon-processors-with-bfloat16.html)\n",
    "\n",
    "Since Google Colab doesn't meet the hardware requirement, it's incompatible with this notebook and will not see speedups from our optimizations.\n",
    "\n",
    "Please run the following to ensure you meet the requirements."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d7d9247a-b053-46af-89f6-e57067ecb8cd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Notebook requirements  met.\n"
     ]
    }
   ],
   "source": [
    "import sys\n",
    "import os\n",
    "from time import sleep\n",
    "\n",
    "os_info = !uname -o\n",
    "os_type = os_info[0]\n",
    "is_os_ok = \"linux\" in os_type.lower()\n",
    "is_python_ver_ok = sys.version_info.major == 3 and sys.version_info.minor >=9\n",
    "cpu_info = !lscpu\n",
    "is_cpu_ok = \"intel\" in cpu_info.grep(\"vendor\")[0].lower() and \"bf16\" in cpu_info.grep(\"flags\")[0]\n",
    "if not is_os_ok:\n",
    "    print(f\"This notebook requires a Linux type operating system, but you have {os_type}\")\n",
    "if not is_python_ver_ok:\n",
    "    print(f\"This notebook requires Python 3.9+ but your version is {sys.version.split()[0]}.\")\n",
    "if not is_cpu_ok:\n",
    "    print(f\"Your CPU does not support bfloat16.\")\n",
    "print(f\"Notebook requirements {'' if is_os_ok and is_python_ver_ok and is_cpu_ok else 'not'} met.\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8855b561-fea5-4137-ab78-e3c32a200748",
   "metadata": {},
   "source": [
    "### Before Running this Notebook\n",
    "1. Please create a `conda` environment and install `jupyterlab`:\n",
    "```bash\n",
    "conda create -n setfit python=3.10 -y && conda activate setfit\n",
    "python -m pip install -U jupyterlab\n",
    "```\n",
    "\n",
    "2. Install **TCMalloc**: Google's fast, multi-threaded `malloc()` implementation:\n",
    "```bash\n",
    "conda install gperftools -c conda-forge -y\n",
    "echo export LD_PRELOAD=${CONDA_PREFIX}/lib/libtcmalloc.so:$LD_PRELOAD >> ~/.bashrc\n",
    "# **** Restart the shell or `source ~/.bashrc` ****\n",
    "```\n",
    "\n",
    "3. To reproduce the maximum speeds reported in this notebook, please launch it using `numactl`.\\\n",
    "`numactl` enables gaining better control on which resources you're running. \\\n",
    "It's best to run on a single socket (`-m 0`):\n",
    "```bash\n",
    "OMP_NUM_THREADS=<NUM_THREADS> numactl -C <CORES_RANGE> -m 0 jupyter lab\n",
    "```\n",
    "\n",
    "**Note**: For even increased throughput, the data can be split into 2 processes, each running seperately on a different socket."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67c08bea-1c62-450b-b532-068bbd6bc92d",
   "metadata": {},
   "source": [
    "#### Experimental Setup\n",
    "- Hardware (CPU): 5th gen Intel Xeon 8580 with 2 sockets, 60 cores per socket.\n",
    "- This notebook was tested on `Ubuntu 22.04.3 LTS` and `Python 3.10.13`.\n",
    "- All runs were evaluated with 60 cores on 1 CPU socket, and cores ranging from 0-59.\n",
    "- TCMalloc was installed and defined as an environment variable in all runs. (see above)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79bc2a85",
   "metadata": {},
   "source": [
    "## Setup\n",
    "Install the required packages to evaluate and visualize results for SetFit models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "839bd550",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python -m pip install -U setfit evaluate matplotlib -qqq"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1f72d255-ab98-4cc3-878b-6a60e08ccf6d",
   "metadata": {},
   "source": [
    "### Install Optimum Intel"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1cefb9e-67b8-40d2-9448-5cb54527555f",
   "metadata": {},
   "source": [
    "Install Optimum Intel along with INC and IPEX backends which we'll use in our optimization later on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "e9466ee2-5791-45b4-b69d-33bb5a82f048",
   "metadata": {},
   "outputs": [],
   "source": [
    "!python -m pip install --upgrade-strategy eager optimum[neural-compressor] -qqq\n",
    "!python -m pip install intel-extension-for-pytorch -qqq"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2493cd04-812f-461b-bb26-9225d76c2774",
   "metadata": {},
   "source": [
    "### Set Up Benchmark"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6a090c0-a065-4b3f-b803-5e518c73299f",
   "metadata": {},
   "source": [
    "First, define the infrastructure for conducting latency, throughput and accuracy benchmarks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "55756fec-fc22-4590-84d7-2f3df37b9256",
   "metadata": {
    "id": "55756fec-fc22-4590-84d7-2f3df37b9256"
   },
   "outputs": [],
   "source": [
    "import sys\n",
    "import os\n",
    "from time import sleep\n",
    "from pathlib import Path\n",
    "from time import perf_counter\n",
    "import warnings\n",
    "import subprocess\n",
    "import sys\n",
    "from collections import defaultdict\n",
    "\n",
    "import evaluate\n",
    "import numpy as np\n",
    "import torch\n",
    "from tqdm.auto import trange\n",
    "from sentence_transformers import SentenceTransformer\n",
    "from setfit.exporters.utils import mean_pooling\n",
    "import logging\n",
    "\n",
    "logging.getLogger(\"sentence_transformers\").setLevel(logging.ERROR)\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "os.environ[\"TOKENIZERS_PARALLELISM\"] = \"false\"\n",
    "\n",
    "def get_dir_size(path):\n",
    "    ps = subprocess.Popen(('du','-hk', path), stdout=subprocess.PIPE)\n",
    "    output = subprocess.check_output((\"awk\", \"{print $1/1024,$2}\"), stdin=ps.stdout)\n",
    "    ps.wait()\n",
    "    return float(output.split()[0].decode('utf-8'))\n",
    "\n",
    "def generate_random_sequences(batch_size, length, vocab_size):\n",
    "    input_ids = torch.randint(0, vocab_size - 1, (batch_size, length))\n",
    "    token_type_ids = torch.zeros((batch_size, length), dtype=torch.int64)\n",
    "    attention_mask = torch.ones((batch_size, length), dtype=torch.int64)\n",
    "    return {\n",
    "        \"input_ids\": input_ids,\n",
    "        \"token_type_ids\": token_type_ids,\n",
    "        \"attention_mask\": attention_mask,\n",
    "    }\n",
    "    \n",
    "class PerformanceBenchmark:\n",
    "    def __init__(self, model, dataset, optim_type, max_sequence_length=512, metric=\"accuracy\",\n",
    "                 model_path=None, autocast_dtype=None, batch_sizes=None):\n",
    "        self.model = model\n",
    "        self.dataset = dataset\n",
    "        self.optim_type = optim_type\n",
    "        self.metric = evaluate.load(metric)\n",
    "        self.model_path = model_path\n",
    "        self.tokenizer = self.model.model_body.tokenizer\n",
    "        self.vocab_size = self.tokenizer.vocab_size\n",
    "        self.max_length = max_sequence_length\n",
    "        self.device = self.model.model_body.device\n",
    "        self.batch_sizes = [2 ** i for i in range(10)] if batch_sizes is None else batch_sizes\n",
    "        self.autocast_dtype = autocast_dtype\n",
    "        self.model.model_body.eval()\n",
    "            \n",
    "    def to_device(self, batch):\n",
    "        for key in batch:\n",
    "            batch[key] = batch[key].to(self.device)\n",
    "        return batch\n",
    "\n",
    "    @torch.inference_mode()\n",
    "    def timed_predict(self, inputs=None, encoded_inputs=None):\n",
    "        if encoded_inputs is None:\n",
    "            encoded_inputs = self.tokenizer(\n",
    "            inputs, padding=\"max_length\", max_length=self.max_length, truncation=True, return_tensors=\"pt\"\n",
    "        )\n",
    "        start = perf_counter()\n",
    "        encoded_inputs = self.to_device(encoded_inputs)\n",
    "        \n",
    "        with torch.no_grad(), torch.cpu.amp.autocast(dtype=self.autocast_dtype, enabled=self.autocast_dtype is not None):                \n",
    "            if isinstance(self.model.model_body, SentenceTransformer) or \"ipex\" in self.optim_type:\n",
    "                embeddings = self.model.model_body(encoded_inputs)[\"sentence_embedding\"]\n",
    "            else:\n",
    "                # Quantized SetFit model bodies don't include mean pooling\n",
    "                outputs = self.model.model_body(**encoded_inputs)\n",
    "                embeddings = mean_pooling(outputs[\"last_hidden_state\"], encoded_inputs[\"attention_mask\"])\n",
    "                \n",
    "            preds = self.model.model_head.predict(embeddings)\n",
    "            latency = perf_counter() - start\n",
    "        return preds, latency\n",
    "\n",
    "    def compute_accuracy(self):\n",
    "        preds, _ = self.timed_predict(self.dataset[\"text\"])\n",
    "        accuracy = self.metric.compute(predictions=preds, references=self.dataset[\"label\"])\n",
    "        print(f\"Accuracy on test set - {accuracy['accuracy']:.3f}\")\n",
    "        return accuracy\n",
    "        \n",
    "    def compute_size(self):\n",
    "        if self.model_path is None:\n",
    "            tmp_path = Path(\"model.pt\")\n",
    "            try:\n",
    "                with open(tmp_path, \"wb\") as f:\n",
    "                    torch.jit.save(self.model.model_body, f)\n",
    "            except:\n",
    "                state_dict = self.model.model_body.state_dict()\n",
    "                torch.save(state_dict, tmp_path)\n",
    "            \n",
    "            size_mb = get_dir_size(tmp_path)\n",
    "            # Delete temporary file\n",
    "            tmp_path.unlink()\n",
    "        else:\n",
    "            size_mb = get_dir_size(self.model_path)\n",
    "        print(f\"Model size (MB) - {size_mb:.2f}\")\n",
    "        return {\"size_mb\": size_mb}\n",
    "\n",
    "    def compute_latency_and_throughput(self, num_iter=100, num_warmup_iter=40):\n",
    "        res = defaultdict(list)\n",
    "        for batch_size in self.batch_sizes:\n",
    "            try:\n",
    "                for _ in trange(num_warmup_iter, desc=f\"Warmup ({batch_size=})\", file=sys.stdout):\n",
    "                    encoded_inputs = generate_random_sequences(batch_size=batch_size, length=self.max_length, vocab_size=self.vocab_size)\n",
    "                    _ = self.timed_predict(encoded_inputs=encoded_inputs)\n",
    "    \n",
    "                batch_latencies = []\n",
    "                for _ in trange(num_iter, desc=f\"Timed Run ({batch_size=})\", file=sys.stdout):\n",
    "                    encoded_inputs = generate_random_sequences(batch_size=batch_size, length=self.max_length, vocab_size=self.vocab_size)\n",
    "                    _, latency = self.timed_predict(encoded_inputs=encoded_inputs)\n",
    "                    batch_latencies.append(latency)\n",
    "        \n",
    "                # Compute run statistics\n",
    "                batch_latencies = np.array(batch_latencies)\n",
    "                if batch_size == 1:\n",
    "                    latency = 1000 * (batch_latencies / batch_size)\n",
    "                    res.update({\"time_avg_ms\": np.mean(latency), \"time_std_ms\": np.std(latency)})\n",
    "                    print(f\"Avg. latency (ms), {batch_size=} - {res['time_avg_ms']:.2f} +/- {res['time_std_ms']:.2f}\")\n",
    "    \n",
    "                throughputs = batch_size / batch_latencies\n",
    "                res[\"throughputs_avg\"].append(np.mean(throughputs))\n",
    "                res[\"throughputs_std\"].append(np.std(throughputs))\n",
    "                print(f\"Avg. throughput (samples/sec), {batch_size=} - {res['throughputs_avg'][-1]:.1f} +/- {res['throughputs_std'][-1]:.1f}\")\n",
    "                res[\"batch_sizes\"].append(batch_size)\n",
    "            except Exception as e:\n",
    "                print(f\"Skipping throughput computation at {batch_size=} since it failed with the following error: \\n\\n{e}\\n\")\n",
    "        return res\n",
    "        \n",
    "    def run_benchmark(self):\n",
    "        print(f\"\\n\\n{'*' * 150}\\nOptimization type: {self.optim_type}\\n{'*' * 150}\")\n",
    "        all_metrics = {}\n",
    "        for run_metric in self.compute_accuracy, self.compute_latency_and_throughput, self.compute_size:\n",
    "            all_metrics |= run_metric()\n",
    "            print(\"=\" * 100)\n",
    "        return {self.optim_type: all_metrics}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2878768f-9ec4-4675-ab20-0f8bca7bbe37",
   "metadata": {},
   "source": [
    "Let's add a function to plot our results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "99110d37-1aa1-4373-8517-f1f8acbbc9a4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "def plot_metrics(perf_metrics):\n",
    "    df = pd.DataFrame.from_dict(perf_metrics, orient=\"index\")\n",
    "    plt.figure(figsize=(6, 4), dpi=120)\n",
    "    for idx in df.index:\n",
    "        df_opt = df.loc[idx]\n",
    "        x, y = df_opt[\"time_avg_ms\"], df_opt[\"accuracy\"] * 100\n",
    "        plt.errorbar(\n",
    "            x,\n",
    "            y,\n",
    "            xerr=df_opt[\"time_std_ms\"],\n",
    "            fmt=\"o\",\n",
    "            alpha=0.5,\n",
    "            ms=df_opt[\"size_mb\"] / 15,\n",
    "            label=idx,\n",
    "            capsize=5,\n",
    "            capthick=1,\n",
    "        )\n",
    "        plt.text(x, y-3, f'{y:.1f}\\n{x:.1f}ms', va='bottom', fontsize=6, color=\"black\")\n",
    "        \n",
    "    plt.suptitle(\"Accuracy vs Latency (batch_size=1)\")\n",
    "    plt.title('\\nBAAI/bge-small-en-v1.5 SetFit model trained on SetFit/sst2', fontsize=8)\n",
    "    legend = plt.legend(loc=\"lower right\")\n",
    "    plt.ylim(63, 95)\n",
    "    # Use the slowest model to define the x-axis range\n",
    "    xlim = max([metrics[\"time_avg_ms\"] for metrics in perf_metrics.values()]) * 1.3\n",
    "    plt.xlim(0, xlim)\n",
    "    plt.grid(True, linestyle='--', alpha=0.7)\n",
    "    plt.ylabel(\"Accuracy (%)\")\n",
    "    plt.xlabel(\"Latency (ms)\")\n",
    "    plt.savefig(\"accuracy_vs_latency.png\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b842dea2-76cc-408b-9431-0b282b9abaed",
   "metadata": {},
   "source": [
    "## 1. Benchmark SetFit using 🤗 Transformers (Float32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d159ea96-f604-478c-b6a2-bf4cc5c57a5a",
   "metadata": {},
   "source": [
    "Load the dataset for running evaluations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "867a9450-d8f0-4697-b3b4-a5ddf6b04906",
   "metadata": {},
   "outputs": [],
   "source": [
    "from datasets import load_dataset\n",
    "from setfit import SetFitModel\n",
    "\n",
    "test_dataset = load_dataset(\"SetFit/sst2\")[\"validation\"]\n",
    "\n",
    "# Prepare variables for benchmarks\n",
    "model_path = \"dkorat/bge-small-en-v1.5_setfit-sst2-english\"\n",
    "perf_metrics = {}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b24b47c9-7076-4510-8c1e-5c17445bae34",
   "metadata": {},
   "source": [
    "We'll now run the benchmark with the standard 🤗 Transformers backend:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "u-w99Y2qW4lU",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "u-w99Y2qW4lU",
    "outputId": "57f0b8f7-6dad-4e90-c779-658a7de6e960"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "******************************************************************************************************************************************************\n",
      "Optimization type: bge-small (transformers)\n",
      "******************************************************************************************************************************************************\n",
      "Accuracy on test set - 0.884\n",
      "====================================================================================================\n",
      "Warmup (batch_size=1):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=1):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. latency (ms), batch_size=1 - 15.69 +/- 0.57\n",
      "Avg. throughput (samples/sec), batch_size=1 - 63.8 +/- 2.4\n",
      "Warmup (batch_size=2):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=2):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=2 - 84.9 +/- 1.0\n",
      "Warmup (batch_size=4):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=4):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=4 - 110.6 +/- 1.3\n",
      "Warmup (batch_size=8):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=8):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=8 - 119.6 +/- 7.4\n",
      "Warmup (batch_size=16):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=16):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=16 - 115.3 +/- 8.2\n",
      "Warmup (batch_size=32):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=32):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=32 - 83.8 +/- 3.6\n",
      "Warmup (batch_size=64):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=64):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=64 - 81.5 +/- 1.2\n",
      "Warmup (batch_size=128):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=128):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=128 - 61.7 +/- 11.2\n",
      "Warmup (batch_size=256):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=256):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=256 - 74.8 +/- 0.5\n",
      "Warmup (batch_size=512):   0%|          | 0/40 [00:00<?, ?it/s]\n",
      "Timed Run (batch_size=512):   0%|          | 0/100 [00:00<?, ?it/s]\n",
      "Avg. throughput (samples/sec), batch_size=512 - 66.1 +/- 3.7\n",
      "====================================================================================================\n",
      "Model size (MB) - 127.33\n",
      "====================================================================================================\n"
     ]
    }
   ],
   "source": [
    "setfit_model = SetFitModel.from_pretrained(model_path)\n",
    "pb = PerformanceBenchmark(\n",
    "    model=setfit_model, \n",
    "    dataset=test_dataset, \n",
    "    optim_type=\"bge-small (transformers)\",\n",
    ")\n",
    "perf_metrics.update(pb.run_benchmark())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f71722e2-7c40-43a3-ab85-de4398d81df7",
   "metadata": {},
   "source": [
    "## 2. Optimize with [Intel Extension for PyTorch](https://intel.github.io/intel-extension-for-pytorch/#introduction) + [TorchScript](https://pytorch.org/docs/stable/jit.html) (BFloat16)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd52ad3a-12ab-4d25-8fc7-bf4703fb6aae",
   "metadata": {},
   "source": [
    "Let's now optimize by using Intel extension for PyTorch (IPEX) runtime with bf16 and tracing the model using torchscript."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "016316c3-30a0-4dd1-9975-1bfb95613122",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "******************************************************************************************************************************************************\n",
      "Optimization type: bge-small (ipex-bfloat16)\n",
      "******************************************************************************************************************************************************\n",
      "Accuracy on test set - 0.884\n",
      "====================================================================================================\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1bc4520e65b74c5b832a2cbcfccf619a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=1):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bcc53fd2605f4bbb8e445d166bbbb4ae",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=1):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. latency (ms), batch_size=1 - 5.67 +/- 0.66\n",
      "Avg. throughput (samples/sec), batch_size=1 - 178.3 +/- 16.5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f9858c80b32b458b8acd1c342bd5648b",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=2):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "731f06d764814b2e9f74cecc52c0349a",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=2):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=2 - 247.4 +/- 7.7\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a386496f1fc14262922fea20ed6494c0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=4):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d6015d0980a849fe85e1a0db3fc3cc76",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=4):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=4 - 338.7 +/- 6.6\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "32c45cd56e7044749aa4a535b57c5ea4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=8):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9dca0ee5539045e88ea7a5ad1fb04efe",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=8):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=8 - 424.2 +/- 6.2\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "67495d6f9c974e3b8b42c75b63e64a76",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=16):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d277d6471762463fb32334a39427922d",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=16):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=16 - 481.5 +/- 6.0\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1ddf01489cff4b61bcd7729533168ae7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=32):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d61e02f9cd7d4721a38c9c637b40d0bc",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=32):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=32 - 537.4 +/- 2.6\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a65148d1257248d5ab06927e197dd95c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=64):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "268ac87eaa824cf3879e60c35da70b58",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=64):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=64 - 573.7 +/- 3.9\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1e18bc4a85084e9484249df0c7a3bc24",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=128):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f3fcba39b7654733944273f634763d8c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=128):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=128 - 582.7 +/- 8.9\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2200520c6f4a401785c8b3b0b957b403",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=256):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "096726c7d7e84b8698493413e4f2087c",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=256):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=256 - 556.3 +/- 1.1\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b2e9a07b250f46e69025177c5117ba39",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=512):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "dee09df5a5b54f978338d9d7b934cf47",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=512):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=512 - 536.9 +/- 1.0\n",
      "====================================================================================================\n",
      "Model size (MB) - 63.74\n",
      "====================================================================================================\n"
     ]
    }
   ],
   "source": [
    "import intel_extension_for_pytorch as ipex\n",
    "\n",
    "setfit_model = SetFitModel.from_pretrained(model_path)\n",
    "model = setfit_model.model_body\n",
    "model.eval()\n",
    "dtype = torch.bfloat16\n",
    "optimized_model = ipex.optimize(model, dtype=dtype)\n",
    "d = generate_random_sequences(batch_size=1, length=model.tokenizer.model_max_length, vocab_size=model.tokenizer.vocab_size)\n",
    "\n",
    "optimized_model = torch.jit.trace(optimized_model, (d,), check_trace=False, strict=False)\n",
    "setfit_model.model_body = torch.jit.freeze(optimized_model)\n",
    "setfit_model.model_body.tokenizer = model.tokenizer\n",
    "setfit_model.model_body.device = model.device\n",
    "\n",
    "pb = PerformanceBenchmark(\n",
    "    model=setfit_model,\n",
    "    dataset=test_dataset,\n",
    "    optim_type=f\"bge-small (ipex-{str(dtype).split('.')[1]})\",\n",
    "    autocast_dtype=dtype,\n",
    ")\n",
    "perf_metrics.update(pb.run_benchmark())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "AiPUhOCNWRny",
   "metadata": {
    "id": "AiPUhOCNWRny"
   },
   "source": [
    "## 3. Optimize with 🤗 Optimum Intel"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0e60d191-db3b-49a3-a281-34ae9063cb25",
   "metadata": {},
   "source": [
    "In order to optimize our SetFit model, we will apply quantization to the model body, using Intel [Neural Compressor](https://www.intel.com/content/www/us/en/developer/tools/oneapi/neural-compressor.html) (INC), part of Optimum Intel.\n",
    "\n",
    "**Quantization** is a very popular deep learning model optimization technique for improving inference speeds. It minimizes the number of bits required to represent the weights and/or activations in a neural network. This is done by converting a set of real-valued numbers into their lower-bit data representations, such as INT8. Moreover, quantization can enable faster computations in lower precision.\n",
    "\n",
    "Specifically, we'll apply post-training static quantization (PTQ). PTQ can reduce the memory footprint and latency for inference, while still preserving the accuracy of the model, with only a small unlabeled calibration set and without any training."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e0557f9-8193-4059-8056-bfc50c00deef",
   "metadata": {},
   "source": [
    "#### Prepare a Calibration Dataset\n",
    "\n",
    "The calibration dataset should be able to represent the data distribution of unseen data. In general, preparing 100 samples is enough for calibration. We'll use the `rotten_tomatoes` dataset in our case, since it’s composed of movie reviews, like our target dataset, sst2. First, we’ll load 100 random samples from this dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "14b2ae65-fc24-4ce2-9125-2100e78ed18b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['. . . plays like somebody spliced random moments of a chris rock routine into what is otherwise a cliche-riddled but self-serious spy thriller .',\n",
       " \"michael moore has perfected the art of highly entertaining , self-aggrandizing , politically motivated documentary-making , and he's got as potent a topic as ever here .\"]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "calibration_set = load_dataset(\"rotten_tomatoes\")[\"train\"].shuffle(seed=42).select(range(100))  \n",
    "calibration_set[\"text\"][:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f471a9fb-52f0-4b79-9126-73d9de2b280f",
   "metadata": {},
   "source": [
    "To prepare the dataset for quantization, we'll need to tokenize each example. We won’t need the “text” and “label” columns, so let’s remove them."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "06480d69-f706-4eb6-90dd-e9bb0001625b",
   "metadata": {},
   "outputs": [],
   "source": [
    "def tokenize(examples):\n",
    "    return tokenizer(examples[\"text\"], padding=\"max_length\", max_length=tokenizer.model_max_length, truncation=True)\n",
    "\n",
    "setfit_model = SetFitModel.from_pretrained(model_path)\n",
    "tokenizer = setfit_model.model_body.tokenizer\n",
    "calibration_set = calibration_set.map(tokenize, num_proc=10).remove_columns([\"text\", \"label\"])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b90c6ba-8f6e-4899-b404-ecebc0f7a78a",
   "metadata": {},
   "source": [
    "#### Run Quantization\n",
    "Now we define the desired quantization process - in our case - Static Post Training Quantization, and use `optimum.intel` to run the quantization on our calibration dataset:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d38c5008-77a0-4985-bb99-7108d57af5d2",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "ONNX export is no supported for model with quantized embeddings\n",
      "2024-03-28 04:07:29 [INFO] Start auto tuning.\n",
      "2024-03-28 04:07:29 [INFO] Execute the tuning process due to detect the evaluation function.\n",
      "2024-03-28 04:07:29 [INFO] Adaptor has 5 recipes.\n",
      "2024-03-28 04:07:29 [INFO] 0 recipes specified by user.\n",
      "2024-03-28 04:07:29 [INFO] 3 recipes require future tuning.\n",
      "2024-03-28 04:07:29 [WARNING] Fail to remove /home/sdp/dkorat/setfit/notebooks/nc_workspace/2024-03-28_04-07-26/ipex_config_tmp.json.\n",
      "2024-03-28 04:07:29 [INFO] *** Initialize auto tuning\n",
      "2024-03-28 04:07:29 [INFO] {\n",
      "2024-03-28 04:07:29 [INFO]     'PostTrainingQuantConfig': {\n",
      "2024-03-28 04:07:29 [INFO]         'AccuracyCriterion': {\n",
      "2024-03-28 04:07:29 [INFO]             'criterion': 'relative',\n",
      "2024-03-28 04:07:29 [INFO]             'higher_is_better': True,\n",
      "2024-03-28 04:07:29 [INFO]             'tolerable_loss': 0.01,\n",
      "2024-03-28 04:07:29 [INFO]             'absolute': None,\n",
      "2024-03-28 04:07:29 [INFO]             'keys': <bound method AccuracyCriterion.keys of <neural_compressor.config.AccuracyCriterion object at 0x7f2ef9a26080>>,\n",
      "2024-03-28 04:07:29 [INFO]             'relative': 0.01\n",
      "2024-03-28 04:07:29 [INFO]         },\n",
      "2024-03-28 04:07:29 [INFO]         'approach': 'post_training_static_quant',\n",
      "2024-03-28 04:07:29 [INFO]         'backend': 'ipex',\n",
      "2024-03-28 04:07:29 [INFO]         'calibration_sampling_size': [\n",
      "2024-03-28 04:07:29 [INFO]             100\n",
      "2024-03-28 04:07:29 [INFO]         ],\n",
      "2024-03-28 04:07:29 [INFO]         'device': 'cpu',\n",
      "2024-03-28 04:07:29 [INFO]         'diagnosis': False,\n",
      "2024-03-28 04:07:29 [INFO]         'domain': 'nlp',\n",
      "2024-03-28 04:07:29 [INFO]         'example_inputs': 'Not printed here due to large size tensors...',\n",
      "2024-03-28 04:07:29 [INFO]         'excluded_precisions': [\n",
      "2024-03-28 04:07:29 [INFO]         ],\n",
      "2024-03-28 04:07:29 [INFO]         'framework': 'pytorch_ipex',\n",
      "2024-03-28 04:07:29 [INFO]         'inputs': [\n",
      "2024-03-28 04:07:29 [INFO]         ],\n",
      "2024-03-28 04:07:29 [INFO]         'model_name': '',\n",
      "2024-03-28 04:07:29 [INFO]         'ni_workload_name': 'quantization',\n",
      "2024-03-28 04:07:29 [INFO]         'op_name_dict': None,\n",
      "2024-03-28 04:07:29 [INFO]         'op_type_dict': None,\n",
      "2024-03-28 04:07:29 [INFO]         'outputs': [\n",
      "2024-03-28 04:07:29 [INFO]         ],\n",
      "2024-03-28 04:07:29 [INFO]         'quant_format': 'default',\n",
      "2024-03-28 04:07:29 [INFO]         'quant_level': 'auto',\n",
      "2024-03-28 04:07:29 [INFO]         'recipes': {\n",
      "2024-03-28 04:07:29 [INFO]             'smooth_quant': False,\n",
      "2024-03-28 04:07:29 [INFO]             'smooth_quant_args': {\n",
      "2024-03-28 04:07:29 [INFO]             },\n",
      "2024-03-28 04:07:29 [INFO]             'layer_wise_quant': False,\n",
      "2024-03-28 04:07:29 [INFO]             'layer_wise_quant_args': {\n",
      "2024-03-28 04:07:29 [INFO]             },\n",
      "2024-03-28 04:07:29 [INFO]             'fast_bias_correction': False,\n",
      "2024-03-28 04:07:29 [INFO]             'weight_correction': False,\n",
      "2024-03-28 04:07:29 [INFO]             'gemm_to_matmul': True,\n",
      "2024-03-28 04:07:29 [INFO]             'graph_optimization_level': None,\n",
      "2024-03-28 04:07:29 [INFO]             'first_conv_or_matmul_quantization': True,\n",
      "2024-03-28 04:07:29 [INFO]             'last_conv_or_matmul_quantization': True,\n",
      "2024-03-28 04:07:29 [INFO]             'pre_post_process_quantization': True,\n",
      "2024-03-28 04:07:29 [INFO]             'add_qdq_pair_to_weight': False,\n",
      "2024-03-28 04:07:29 [INFO]             'optypes_to_exclude_output_quant': [\n",
      "2024-03-28 04:07:29 [INFO]             ],\n",
      "2024-03-28 04:07:29 [INFO]             'dedicated_qdq_pair': False,\n",
      "2024-03-28 04:07:29 [INFO]             'rtn_args': {\n",
      "2024-03-28 04:07:29 [INFO]             },\n",
      "2024-03-28 04:07:29 [INFO]             'awq_args': {\n",
      "2024-03-28 04:07:29 [INFO]             },\n",
      "2024-03-28 04:07:29 [INFO]             'gptq_args': {\n",
      "2024-03-28 04:07:29 [INFO]             },\n",
      "2024-03-28 04:07:29 [INFO]             'teq_args': {\n",
      "2024-03-28 04:07:29 [INFO]             }\n",
      "2024-03-28 04:07:29 [INFO]         },\n",
      "2024-03-28 04:07:29 [INFO]         'reduce_range': None,\n",
      "2024-03-28 04:07:29 [INFO]         'TuningCriterion': {\n",
      "2024-03-28 04:07:29 [INFO]             'max_trials': 100,\n",
      "2024-03-28 04:07:29 [INFO]             'objective': [\n",
      "2024-03-28 04:07:29 [INFO]                 'performance'\n",
      "2024-03-28 04:07:29 [INFO]             ],\n",
      "2024-03-28 04:07:29 [INFO]             'strategy': 'basic',\n",
      "2024-03-28 04:07:29 [INFO]             'strategy_kwargs': None,\n",
      "2024-03-28 04:07:29 [INFO]             'timeout': 0\n",
      "2024-03-28 04:07:29 [INFO]         },\n",
      "2024-03-28 04:07:29 [INFO]         'use_bf16': True\n",
      "2024-03-28 04:07:29 [INFO]     }\n",
      "2024-03-28 04:07:29 [INFO] }\n",
      "2024-03-28 04:07:29 [WARNING] [Strategy] Please install `mpi4py` correctly if using distributed tuning; otherwise, ignore this warning.\n",
      "2024-03-28 04:07:29 [INFO]  Found 12 blocks\n",
      "2024-03-28 04:07:29 [INFO] Attention Blocks: 12\n",
      "2024-03-28 04:07:29 [INFO] FFN Blocks: 12\n",
      "2024-03-28 04:07:29 [INFO] Attention Blocks : \n",
      "2024-03-28 04:07:29 [INFO] [['encoder.layer.0.attention.self.query', 'encoder.layer.0.attention.self.key', 'encoder.layer.0.attention.self.value', 'encoder.layer.0.attention.output.dense'], ['encoder.layer.1.attention.self.query', 'encoder.layer.1.attention.self.key', 'encoder.layer.1.attention.self.value', 'encoder.layer.1.attention.output.dense'], ['encoder.layer.2.attention.self.query', 'encoder.layer.2.attention.self.key', 'encoder.layer.2.attention.self.value', 'encoder.layer.2.attention.output.dense'], ['encoder.layer.3.attention.self.query', 'encoder.layer.3.attention.self.key', 'encoder.layer.3.attention.self.value', 'encoder.layer.3.attention.output.dense'], ['encoder.layer.4.attention.self.query', 'encoder.layer.4.attention.self.key', 'encoder.layer.4.attention.self.value', 'encoder.layer.4.attention.output.dense'], ['encoder.layer.5.attention.self.query', 'encoder.layer.5.attention.self.key', 'encoder.layer.5.attention.self.value', 'encoder.layer.5.attention.output.dense'], ['encoder.layer.6.attention.self.query', 'encoder.layer.6.attention.self.key', 'encoder.layer.6.attention.self.value', 'encoder.layer.6.attention.output.dense'], ['encoder.layer.7.attention.self.query', 'encoder.layer.7.attention.self.key', 'encoder.layer.7.attention.self.value', 'encoder.layer.7.attention.output.dense'], ['encoder.layer.8.attention.self.query', 'encoder.layer.8.attention.self.key', 'encoder.layer.8.attention.self.value', 'encoder.layer.8.attention.output.dense'], ['encoder.layer.9.attention.self.query', 'encoder.layer.9.attention.self.key', 'encoder.layer.9.attention.self.value', 'encoder.layer.9.attention.output.dense'], ['encoder.layer.10.attention.self.query', 'encoder.layer.10.attention.self.key', 'encoder.layer.10.attention.self.value', 'encoder.layer.10.attention.output.dense'], ['encoder.layer.11.attention.self.query', 'encoder.layer.11.attention.self.key', 'encoder.layer.11.attention.self.value', 'encoder.layer.11.attention.output.dense']]\n",
      "2024-03-28 04:07:29 [INFO] FFN Blocks : \n",
      "2024-03-28 04:07:29 [INFO] [['encoder.layer.0.intermediate.dense', 'encoder.layer.0.output.dense'], ['encoder.layer.1.intermediate.dense', 'encoder.layer.1.output.dense'], ['encoder.layer.2.intermediate.dense', 'encoder.layer.2.output.dense'], ['encoder.layer.3.intermediate.dense', 'encoder.layer.3.output.dense'], ['encoder.layer.4.intermediate.dense', 'encoder.layer.4.output.dense'], ['encoder.layer.5.intermediate.dense', 'encoder.layer.5.output.dense'], ['encoder.layer.6.intermediate.dense', 'encoder.layer.6.output.dense'], ['encoder.layer.7.intermediate.dense', 'encoder.layer.7.output.dense'], ['encoder.layer.8.intermediate.dense', 'encoder.layer.8.output.dense'], ['encoder.layer.9.intermediate.dense', 'encoder.layer.9.output.dense'], ['encoder.layer.10.intermediate.dense', 'encoder.layer.10.output.dense'], ['encoder.layer.11.intermediate.dense', 'encoder.layer.11.output.dense']]\n",
      "2024-03-28 04:07:29 [INFO] Pass query framework capability elapsed time: 406.4 ms\n",
      "2024-03-28 04:07:29 [INFO] Get FP32 model baseline.\n",
      "2024-03-28 04:07:29 [INFO] Save tuning history to /home/sdp/dkorat/setfit/notebooks/nc_workspace/2024-03-28_04-07-26/./history.snapshot.\n",
      "2024-03-28 04:07:29 [INFO] FP32 baseline is: [Accuracy: 1.0000, Duration (seconds): 0.0000]\n",
      "2024-03-28 04:07:29 [INFO] Quantize the model with default config.\n",
      "2024-03-28 04:07:38 [INFO] |******Mixed Precision Statistics******|\n",
      "2024-03-28 04:07:38 [INFO] +---------------+-----------+----------+\n",
      "2024-03-28 04:07:38 [INFO] |    Op Type    |   Total   |   INT8   |\n",
      "2024-03-28 04:07:38 [INFO] +---------------+-----------+----------+\n",
      "2024-03-28 04:07:38 [INFO] |     matmul    |     24    |    24    |\n",
      "2024-03-28 04:07:38 [INFO] |     Linear    |     25    |    25    |\n",
      "2024-03-28 04:07:38 [INFO] +---------------+-----------+----------+\n",
      "2024-03-28 04:07:38 [INFO] Pass quantize model elapsed time: 8978.15 ms\n",
      "2024-03-28 04:07:38 [INFO] Tune 1 result is: [Accuracy (int8|fp32): 1.0000|1.0000, Duration (seconds) (int8|fp32): 0.0000|0.0000], Best tune result is: [Accuracy: 1.0000, Duration (seconds): 0.0000]\n",
      "2024-03-28 04:07:38 [INFO] |**********************Tune Result Statistics**********************|\n",
      "2024-03-28 04:07:38 [INFO] +--------------------+----------+---------------+------------------+\n",
      "2024-03-28 04:07:38 [INFO] |     Info Type      | Baseline | Tune 1 result | Best tune result |\n",
      "2024-03-28 04:07:38 [INFO] +--------------------+----------+---------------+------------------+\n",
      "2024-03-28 04:07:38 [INFO] |      Accuracy      | 1.0000   |    1.0000     |     1.0000       |\n",
      "2024-03-28 04:07:38 [INFO] | Duration (seconds) | 0.0000   |    0.0000     |     0.0000       |\n",
      "2024-03-28 04:07:38 [INFO] +--------------------+----------+---------------+------------------+\n",
      "2024-03-28 04:07:38 [INFO] [Strategy] Found a model that meets the accuracy requirements.\n",
      "2024-03-28 04:07:38 [INFO] Save tuning history to /home/sdp/dkorat/setfit/notebooks/nc_workspace/2024-03-28_04-07-26/./history.snapshot.\n",
      "2024-03-28 04:07:38 [INFO] [Strategy] Found the model meets accuracy requirements, ending the tuning process.\n",
      "2024-03-28 04:07:38 [INFO] Specified timeout or max trials is reached! Found a quantized model which meet accuracy goal. Exit.\n",
      "2024-03-28 04:07:38 [INFO] Save deploy yaml to /home/sdp/dkorat/setfit/notebooks/nc_workspace/2024-03-28_04-07-26/deploy.yaml\n",
      "Model weights saved to /tmp/bge-small-en-v1.5_setfit-sst2-english_opt/pytorch_model.bin\n",
      "Configuration saved in /tmp/bge-small-en-v1.5_setfit-sst2-english_opt/inc_config.json\n"
     ]
    }
   ],
   "source": [
    "from optimum.intel import INCQuantizer\n",
    "from neural_compressor.config import PostTrainingQuantConfig\n",
    "\n",
    "setfit_body = setfit_model.model_body[0].auto_model\n",
    "quantizer = INCQuantizer.from_pretrained(setfit_body)\n",
    "optimum_model_path = \"/tmp/bge-small-en-v1.5_setfit-sst2-english_opt\"\n",
    "quantization_config = PostTrainingQuantConfig(approach=\"static\", backend=\"ipex\", domain=\"nlp\")\n",
    "quantizer.quantize(\n",
    "    quantization_config=quantization_config,\n",
    "    calibration_dataset=calibration_set,\n",
    "    save_directory=optimum_model_path,\n",
    "    batch_size=1,\n",
    ")\n",
    "tokenizer.save_pretrained(optimum_model_path);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f1e8d9eb-354c-4d0f-8b68-d1cb3417dd29",
   "metadata": {},
   "source": [
    "#### Benchmark Optimum Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "424628ae-091b-47bd-952f-96e1d6e7a1b2",
   "metadata": {},
   "source": [
    "This next class defines a wrapper around our SetFit model which plugs in our quantized model at inference (instead of the original model body)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "cf891fe1-03d4-44ce-a77e-9e86a9399f75",
   "metadata": {},
   "outputs": [],
   "source": [
    "class OptimumSetFitModel:\n",
    "    def __init__(self, setfit_model, model_body):\n",
    "        model_body.tokenizer = setfit_model.model_body.tokenizer\n",
    "        self.model_body = model_body\n",
    "        self.model_head = setfit_model.model_head"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72e7215f",
   "metadata": {},
   "source": [
    "Time to run the performance benchmark on our optimized SetFit model!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4941e895-72e2-43e6-a3f4-9af39a5f0315",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Passing the argument `library_name` to `get_supported_tasks_for_model_type` is required, but got library_name=None. Defaulting to `transformers`. An error will be raised in a future version of Optimum if `library_name` is not provided.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "\n",
      "******************************************************************************************************************************************************\n",
      "Optimization type: bge-small (optimum-int8)\n",
      "******************************************************************************************************************************************************\n",
      "Accuracy on test set - 0.881\n",
      "====================================================================================================\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "0202ef5c0a024ef7825a988e4517d9d1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=1):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1c512578457f47199620a48edd5f0bc3",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=1):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. latency (ms), batch_size=1 - 4.55 +/- 0.25\n",
      "Avg. throughput (samples/sec), batch_size=1 - 220.2 +/- 9.4\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "5178bbc2b023465e9ac8ecc1dc1cc186",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=2):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "14e60d4b43ef4e2f999fff535c08ec79",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=2):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=2 - 254.3 +/- 9.4\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "01ebec0131664f27a79fc79a1bb6dec5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=4):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "2306a27ceb484a6eb357a1e03205fdec",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=4):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=4 - 384.8 +/- 9.0\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "346210aa843f4eb795759a9dc555f731",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=8):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "bf9a5625454f46669f7628c026380f49",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=8):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=8 - 536.3 +/- 9.5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1fc1686607bb4ce8a3d6a475905aac62",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=16):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4c26da7184d84fbdacfb1d1c6a7f0016",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=16):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=16 - 646.3 +/- 13.8\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "841cb888cf2d44d8942686c0e6733af4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=32):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d5c26ebfd5ad4a5082832775c2320846",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=32):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=32 - 790.2 +/- 7.9\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "a29d0415cfa64e0281f15ce50f3f9b2e",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=64):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b07cb21638d44bc7820c9ac4b0ff55ca",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=64):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=64 - 896.2 +/- 9.1\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c412bb3ec3904c57bdd50ba2fc1c7407",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=128):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d89726879ff24880a0d1235bb16e9b68",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=128):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=128 - 935.2 +/- 3.5\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6ffce123197b4dcc93224e10329825f9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=256):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3467fa5ce9314ccd97fc9f09fe22c519",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=256):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=256 - 923.4 +/- 10.8\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "13dac04952ac4e089309711fbfd85bf9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Warmup (batch_size=512):   0%|          | 0/40 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f836a2679b9044799b36427686e440e4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Timed Run (batch_size=512):   0%|          | 0/100 [00:00<?, ?it/s]"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Avg. throughput (samples/sec), batch_size=512 - 857.1 +/- 2.2\n",
      "====================================================================================================\n",
      "Model size (MB) - 44.65\n",
      "====================================================================================================\n"
     ]
    }
   ],
   "source": [
    "from optimum.intel import IPEXModel\n",
    "\n",
    "optimum_model = IPEXModel.from_pretrained(optimum_model_path)\n",
    "optimum_setfit_model = OptimumSetFitModel(setfit_model, model_body=optimum_model)\n",
    "\n",
    "pb = PerformanceBenchmark(\n",
    "    model=optimum_setfit_model,\n",
    "    dataset=test_dataset,\n",
    "    optim_type=f\"bge-small (optimum-int8)\",\n",
    "    model_path=optimum_model_path,\n",
    "    autocast_dtype=torch.bfloat16,\n",
    ")\n",
    "perf_metrics.update(pb.run_benchmark())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da6b2998-6605-4e42-9fb5-247c5eee750f",
   "metadata": {},
   "source": [
    "#### Plot Latencies and Compute Latency Speedup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "O8jpZ3gdWRn9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "O8jpZ3gdWRn9",
    "outputId": "8d31c81a-67e4-4074-cf35-9f56d6dcdd20"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Latency speedup for 'bge-small (optimum-int8) vs bge-small (transformers)': 3.4x\n"
     ]
    }
   ],
   "source": [
    "plot_metrics(perf_metrics)\n",
    "speedup = perf_metrics['bge-small (transformers)']['time_avg_ms'] / perf_metrics['bge-small (optimum-int8)']['time_avg_ms']\n",
    "print(f\"Latency speedup for 'bge-small (optimum-int8) vs bge-small (transformers)': {speedup:.1f}x\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48568c7b",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true
   },
   "source": [
    "#### After optimizing, inference with batch size 1 is 3.4x faster than before, with minimal drop in accuracy!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abd6872f-1765-48a5-beba-a989555b5461",
   "metadata": {},
   "source": [
    "#### Throughput Speedup"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e043497-89c0-44a7-b08f-1f8012cd9fba",
   "metadata": {},
   "source": [
    "Let's look now at the throughput, which is the number of samples the model can predict per second.\\\n",
    "We'll plot this value for our optimized and baseline models, as a function of the batch size."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "c701770a-d643-4f32-bd56-2574ba7dc691",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA/0AAALACAYAAADBrBKnAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/H5lhTAAAACXBIWXMAABJ0AAASdAHeZh94AAEAAElEQVR4nOzdd1hT59sH8G8GIyQkQNhDQFHcC3cduMW9tY6K1r1qW1ut/qxobV11dVi1ratW66qjdW9ttbVqsYoLFRRBARmBMELG8/7BmyMxQRHQA+H+9OIqPDnjOXeOgfs8S8AYYyCEEEIIIYQQQojVEfJdAUIIIYQQQgghhLwelPQTQgghhBBCCCFWipJ+QgghhBBCCCHESlHSTwghhBBCCCGEWClK+gkhhBBCCCGEECtFST8hhBBCCCGEEGKlKOknhBBCCCGEEEKsFCX9hBBCCCGEEEKIlaKknxBCCCGEEEIIsVKU9BNCCCGEEEIIIVaKkn5CCCGEEEIIIcRKUdJPCCHkjQsNDYVAIOC7Gq/dxo0bIRAIsHHjRr6rQt6g06dPQyAQICIigpfz031HCCGkIEr6CSGEFJtAIHilL0pCyobY2FgIBAKEh4fzXZWXMtb1+S+JRIKqVati0qRJiIuLK7XzlMWY6PV6fP/992jTpg1cXFxgY2MDd3d31K1bF6NHj8b+/fv5riIhhJAyTMx3BQghhJRfc+fONStbuXIlVCoV3nvvPTg5OZm8Vr9+/TdTMWJ1FAoFpk2bxv2ckpKCM2fOYPXq1di1axeuXLkCHx8f/ir4muj1enTv3h2HDx+Gk5MTunXrBl9fX+Tl5SEqKgpbt27FrVu30LNnT26fPn36oFmzZvDy8uKx5oQQQsoKSvoJIYQUm6Xuyxs3boRKpcK0adMQEBDwxutErJOTk5PF+61nz5747bff8P333/PWnf512rZtGw4fPox69erhzJkzUCgUJq9nZ2fj77//NilTKBRm2xFCCKm4qHs/IYQQ3uh0OnzxxReoWrUq7Ozs4OfnhxkzZiAvL89sW4FAgNDQUDx58gSjR4+Gj48PRCKRyZCBHTt2oHXr1lAoFJBIJKhTpw4WLlwIjUZT6PEsCQ8Ph0AgQGxsrEk5YwyrVq1CzZo1YW9vDx8fH0yePBkqlQoBAQEvfMhx6tQphIaGwtHREXK5HN26dcPNmzcLPff9+/exfPlyVK9eHfb29vD19cX777+PjIyMEl1LREQEAgMDAQCbNm0q8vCL3NxcODk5wd3dHTqdzuI2EyZMgEAgwO+//86VnTt3Dj169ICvry/s7Ozg6emJZs2aYd68eYWe61V07NgRAJCcnGxSnpCQgPnz5+Ott96Cp6cnbG1t4e3tjSFDhuDGjRsm275KTI4ePYoePXrA3d2du2d79eqF48ePW6xfZGQkunXrBicnJzg4OKBNmzY4f/58ka/PuG14eLjFRN7BwQFt27Y1KbM0pt94HxT2Zene3bZtG9q2bQsnJyfY29ujRo0aWLBggcV/T4QQQsouauknhBDCmyFDhuDcuXMICwuDXC7HwYMHsWTJEiQlJWHDhg1m26empqJZs2aQyWTo27cvhEIhPDw8AACzZs3CwoUL4erqiiFDhkAmk+HQoUOYNWsWjhw5gqNHj8LW1rZE9Z00aRK+++47eHt7Y+zYsbC1tcX+/ftx8eJFaLVa2NjYWNzv999/x759+xAWFobx48fjxo0bOHjwIP755x/cuHEDrq6uZvu8//77OHv2LAYOHIhevXrhyJEjWLlyJc6dO4c//vgD9vb2xbqG0NBQpKenY9WqVahXrx569+7Nvfai4Rf29vYYNGgQ1q1bh0OHDqFHjx4mr2s0Gmzfvh0eHh7o0qULAODw4cPo1q0b5HI5evbsCR8fH6SmpuLmzZtYvXq1xeEhr8qYbDdq1Mik/OzZs1i0aBHatm2Lfv36QSaTITo6Grt27cL+/fvx559/ol69eq8Uk7lz52L+/PmQyWTo3bs3/Pz8kJCQgPPnz2PLli3o0KGDSR0uXbqEJUuWoHnz5hg9ejQePnyI3bt3o3379oiMjERwcPBLr0+pVAIA7ty5U5zwcHr37m0xsb927Rp+/fVXODg4mJSPGjUKGzZsgK+vL/r16wcnJyf89ddfmDNnDk6cOIFjx45BLKY/IwkhpFxghBBCSCny9/dnAFhMTEyh27Rp04YBYA0bNmQpKSlcuVqtZlWqVGFCoZA9fvzYZB8ADAAbPnw402q1Jq+dP3+eAWB+fn4m+2m1Wta9e3cGgH3++edmx2vTpo3F+o0YMcLsGs6ePcsAsGrVqrG0tDSuXKPRsFatWjEAzN/f3+Q4GzZsYACYSCRix48fN3lt5syZDABbvHixxXMrlUoWGxvLlev1eta3b18GgM2fP79E1xITE8MAsBEjRljcpzDGOPfr18/stR07djAA7IMPPuDKjPWNjIw02z45OblI5zTWVaFQsLlz53JfU6dOZfXr12cikYiFh4eb3ROJiYksIyPD7HiRkZFMKpWyLl26WDxPYTE5cuQIA8ACAwPZo0ePzF6Pi4vjvj916hR3v27YsMFkuzVr1jAAbMKECUW6/itXrjAbGxsmEAjYsGHD2O7du03uC0uM993z57ZUZx8fH2Zvb88uXLhgtn+fPn1Ydna2yT5z585lANjKlSuLVH9CCCH8o+79hBBCeLN48WK4uLhwP0ulUgwdOhQGgwGXLl0y297W1hZffvmlWQvj+vXrAQD/+9//4OnpyZWLxWIsW7YMQqEQP/zwQ4nqumnTJgDA7NmzTSYotLW1xcKFC1+47+DBg9G+fXuTsrFjxwIALl68aHGf9957D/7+/tzPQqEQS5cuhVAo5K73TWvevDmqVauG3377DampqSavGeMzYsQIs/0kEolZmaXeDS+iUqkwb9487uurr75CZGQkGjdujEGDBpndE+7u7nB0dDQ7Tr169dCuXTucOnUKWq22yOf/+uuvAQDLli2zOGGgr6+vWdlbb71lthrAqFGjIBaLC33fn9egQQNs2bIFHh4e2LJlC/r164eAgAAolUr06dMHv/32W5GvoaDMzEx069YNCQkJ+Omnn9CsWTPutVWrVkEsFmP9+vVm792cOXOgVCrx888/F+u8hBBC3jxK+gkhhPDm+S7ZAODn5wcASEtLM3stICAA7u7uZuVXrlwBALRr187stWrVqsHX1xcxMTFQqVTFruu///4LAGjZsqXZa82aNXthV+dXvU4AaNOmjVlZ5cqV4efnh9jYWKSnpxel2qVuxIgRyMvLwy+//MKVJSYm4siRI2jQoAHq1q3LlQ8dOhQA0LRpU4wfPx7bt2/Ho0ePinVef39/MMa4r/T0dJw6dQo5OTno2rUr1q1bZ7bPgQMH0KNHD3h5ecHGxoYbv/7bb79Bo9Hg6dOnRT7/X3/9BYFAwA1dKApL77uNjQ08PDwKfd8tGThwIB4+fIgjR45gzpw56N69OwwGA/bu3YuePXtixIgRYIwV+Xh6vR4DBw7Ef//9h8WLF6N///7ca9nZ2bh69SqcnZ2xcuVKREREmHx99tlnsLOzszgfBSGEkLKJBmMRQgjhzfNL+gHgkme9Xm/2WsFW/IKMyXxhS5R5eXnh4cOHSE9PL/as5sZzGOcQKEgkEnFjry151ess7DxAfgwePHgAlUpl8biv2zvvvIM5c+Zg06ZNmDhxIgDg559/hk6nM2vl79u3L37//XcsW7YM69evx9q1awEAISEhWLhwITcJX3EoFAqEhoZi165dqFq1KmbMmIHhw4dzLdOrVq3CtGnT4OzsjI4dO6JSpUpwcHCAQCDA3r17cfXq1VeakC49PR3Ozs4Wey0UprD3RywWF/q+F8bGxgadOnVCp06dAOTfN7t378aoUaOwefNm9OnTx2QugheZNGkSDh8+jHHjxuGjjz4yeS0tLQ2MMSQnJ5faZIuEEEL4RS39hBBCyg2BQGCx3JjIP3nyxOLrjx8/NtnOeKzCZqG31Ioul8sB5LdqP0+v1yMlJaXwiheDpfMAz66xJNdSEr6+vmjXrh0uXryIW7duAcjv2m9jY4MhQ4aYbd+tWzecPHkSaWlpOHHiBN5//31ERUWhe/fuZrPoF0dQUBBcXFyQnp7OTXan0+kQEREBT09PREVFYfv27Vi6dCnmzZuHiIiIQh+ovIiTkxPS0tKQk5NT4jqXBpFIhIEDB+L9998HAJw8ebJI+y1ZsgRr165FWFgYvv32W7PXjfdVgwYNTHpWWPoihBBSPlDSTwghpNxr0KABAOD06dNmr929exePHj1CYGCgScurs7Mz4uLizLbX6/WIjIws9Bx//PGH2Wt//fVXoUl3cZ05c8as7P79+4iLi0NAQECJrkUkEnGvF4dxnPqmTZsQGRmJ//77D2FhYXBzcyt0H6lUinbt2mH58uWYNWsW8vLycOjQoWKdvyCdTofMzEwAgMFgAAA8ffoU6enpaNGihVnvD7VazQ0HKehlMWnWrBkYYzh8+HCJ61yajPMWFCUJ37VrF2bOnIl69eph+/bt3DUXJJPJUKtWLURFRZnN20AIIaR8oqSfEEJIuTdq1CgAwIIFC0zWa9fr9Zg+fToMBgPeffddk32aNGmChw8f4ujRoyblCxYswIMHD8zO8c477wAAPv/8c5O5AfLy8jBr1qxSuxajVatWmdTDYDDgo48+gsFgwMiRI022fdVrcXZ2hkAgwMOHD4tVt759+0Iul2PLli3cWvDPT1gH5C+bZ+lhiLEXw/PLxBXHN998A61WC6VSidq1awPIn8TPwcEBly9fhlqt5rbVarV47733LI7lf1lMpkyZAgD48MMPER8fb/a6pbLSsG3bNhw7dox7oFHQkydP8P333wMAWrdu/cLjXLhwAcOHD4e3tzcOHDhgcZJDow8++AB5eXkYNWqUxZ4iaWlpFh+cEEIIKZtoTD8hhJByr0WLFvj444+xZMkS1K5dG/3794dUKsWhQ4dw/fp1tGzZ0mzs8vTp03HkyBH06tULgwYNgouLC86fP4+YmBiEhoaa9Rpo06YNxo4di3Xr1qFWrVro168fbGxs8Ntvv0GhUMDb2xtCYek9S3/rrbdQv359DBo0CAqFAkeOHMHVq1cREhKCjz/+uETXIpPJ0LRpU5w7dw5Dhw5FtWrVIBKJ0LNnT5OJ+AojkUgwYMAA/Pjjj1i9ejWUSiW6detmtt3UqVMRHx+Pt956CwEBAbC1tcXly5dx8uRJ+Pv7Y/DgwUWOR3p6OiIiIrifMzIycOXKFZw5cwZCoRCrV6+GjY0NgPyVDqZOnYpFixahTp066NWrF/Ly8nDq1Cmkpqaibdu2OHXq1CvFpFOnTvjf//6HBQsWoEaNGujduzf8/PyQmJiIP/74A82aNeMegJSmv//+G6tWrYKnpydatmyJwMBAAEBMTAwOHDiAnJwc9OrVy2QyPkveffdd5ObmomnTptyDgoKcnJwwbdo0APkP0S5fvozVq1ejSpUq6Ny5MypVqoTU1FTExMTg7NmzGDlyJNasWVPq10sIIeQ14GGZQEIIIVbM39/fbF3457Vp04YV9iuosDXG8YK16I22bdvG3nrrLSaTyZidnR2rWbMmW7BgAcvJybG4/b59+1hISAizs7NjLi4ubNCgQSw2Ntbi2vaMMabX69ny5ctZcHAws7W1ZV5eXmzixIksPT2dyWQyVq9evSJdy4uuyXjue/fusS+//JIFBwczOzs75u3tzd577z2mUqlK5Vqio6NZ9+7dmYuLCxMIBEVa172gc+fOcWvRT5482eI227dvZ4MHD2ZBQUFMKpUyR0dHVqtWLTZr1iyWlJRUpPPExMRw5yn4ZWNjw3x9fdngwYPZxYsXzfbTarVs2bJlrEaNGsze3p55eHiwYcOGlTgmBw4cYJ07d2bOzs7M1taW+fr6st69e7MTJ05w25w6dYoBYHPnzrV4Tf7+/szf379I1//w4UP2zTffsN69e7Nq1aoxR0dHZmNjwzw9PVlYWBj76aefmF6vN9nH0n1n/HdZ2Jel+vz222+sW7duzM3NjdnY2DAPDw/WuHFjNnv2bHbz5s0i1Z8QQgj/BIzRTCyEEEJISURHR6NatWoYPHgwtm3bVqJjhYeHY9OmTYiJiUFAQEDpVJAQQgghFRaN6SeEEEKK6MmTJ2Zjq7Ozs7lu0X369OGhVoQQQgghhaMx/YQQQkgRrVy5Etu2bUNoaCi8vLzw5MkTnDhxAo8ePUJYWBgGDBjAdxUJIYQQQkxQ0k8IIYQUUceOHXH16lUcPXoUqampEIvFqFatGqZOnYpp06ZBIBDwXUVCCCGEEBM0pp8QQgghhBBCCLFSNKafEEIIIYQQQgixUpT0E0IIIYQQQgghVoqSfkIIIYQQQgghxEpR0k8IIYQQQgghhFgpSvoJIYQQQgghhBArRUk/IYQQQgghhBBipSjpJ4QQQgghhBBCrBQl/YQQQgghhBBCiJWipJ8QQgghhBBCCLFSlPQTQgghhBBCCCFWipJ+QgghhBBCCCHESlHSTwghhBBCCCGEWClK+gkhhBBCCCGEECtFST8hhBBCCCGEEGKlKOknhBBCCCGEEEKsFCX9hBBCCCGEEEKIlaKknxBCCCGEEEIIsVKU9BNCCCGEEEIIIVaKkn5CCCGEEEIIIcRKUdJPCCGEEEIIIYRYKUr6CSGEEEIIIYQQK0VJPyGEEEIIIYQQYqUo6SeEEEIIIYQQQqwUJf2EEEIIIYQQQoiVoqSfEEIIIYQQQgixUpT0E0IIIYQQQgghVoqSfkIIIYQQQgghxEpR0k8IIYQQQgghhFgpSvoJIYQQQgghhBArRUk/IYQQQgghhBBipSjpJ4QQQgghhBBCrBQl/YQQQgghhBBCiJWipJ8QQiqw8PBwDBs2jO9qvHYbN26Er68v93NFue6yolatWti0aRPf1bDohx9+QEBAQJG3f/5eep0ePnwImUyG+/fvv9bz+Pr6YuPGja/1HHx7U7EkhJCyiJJ+Qgh5g0JDQ2FrawuZTAaZTAYfHx9MnjwZOTk5ZtueOnUKAoEAgwcPfuEx27RpA4FAgJs3b5q9JhAIcPz4cZOy6OhoODk5IS8vr2QXQ0rd6NGjUatWLYjF4iI9lIiNjYVAIIBUKuXuKZlMBpVKVeg+T58+xdixY+Hn5weZTAYPDw+0b98e165dK3I9AwIC8MMPP5iURUREQCQSmdRDJpMhKysLUVFRGDFihEmd7969W+TzlSel+VCgUqVKUKvVqFy5cqkcryy5evUqunfvDnd3dzg6OsLf3x99+/Yt8udSYffR85+xMpkMrVu3Novli96nzMxMSKVSPHnypNjXFxoaiv/9738mZQcPHkSHDh3g5uYGJycnNGnSBL/99luxz0EIIUVFST8hhLxhH3/8MdRqNdRqNf7880+cOHEC8+fPN9tu9erVUCqV+PXXX5GYmGjxWFFRUTh79iyUSiW+++67Ip1/9+7d6NGjB2xtbUt0HaT01a1bF8uXL0fPnj1fab+rV69y95RarYZCoSh02+HDh+PJkyf4+++/oVarcfPmTUyYMAFisbik1Ufz5s1N6qFWqyGVSkt8XGuk1Wr5rgJvsrKy0KFDBzRq1Aj3799HRkYG/vjjD3Ts2BGMsRIfv+BnrFqtxtmzZ19p/wMHDqBhw4bw9PQscV0KSktLw4QJE3Dnzh2kpKTgww8/xIABA3Dp0qVSPQ8hhDyPkn5CCOFRQEAAunTpgv/++8+k/PHjx9i7dy++/vpruLi44Pvvv7e4/+rVq1GrVi3MnTsXmzdvRlZW1kvPuXv3bvTt25f7OS8vD6NHj4aTkxMqVaqEJUuWmGx/8OBB1K5dGzKZDO3atcPcuXNNukPr9XosW7YMNWrUgEKhQEhICE6cOPHCOuzYsQO1atWCXC6Hq6srOnTowL0WGhqKKVOmYNCgQZDL5fD19cUvv/yCa9euoXnz5nB0dESTJk1w+/Ztbp+dO3ciJCQEzs7OcHV1Rc+ePRETE/PSWLzIrVu30L17d3h4eMDHxwcTJ040iW9AQADmz5+Prl27wtHREVWqVMGePXsKPd6QIUPw7rvvmpRduXIFtra23EOdqVOnonPnzpDL5SWq+4v88ccfCA8Ph7e3NwDAxcUF/fv3R40aNbhtXnTtYWFhePjwISZPngyZTIZatWq99JwFewYYt69Xrx5kMhnGjx9vcZ+IiAi0bNkSn376Kby8vCCXy/Hxxx8jLS0NgwYNgkKhQEBAAPbt22ey38aNG1G7dm3I5XLUrl3bbFjBkSNHUKdOHe5+jouLM3m9OPez0blz5zB+/HgkJCRwrcw///wzgPxeNytWrECLFi0glUqxe/duXL9+He3bt4ebmxsUCgWaNm2KkydPcsd7vjXb2Dq9du1aBAQEQKFQYMCAAcjIyOD2SU9Px4QJE+Dv7w+lUomuXbuadGlXq9V49913oVQq4ePjg1WrVr30uuLj4zFw4EB4eHjAw8MDgwYNQkJCAvd6eHg4Bg8ejMmTJ0OpVMLDwwNz5swp9Hi3bt3C06dPMXXqVMhkMggEAvj5+WHChAmws7Pjtjt48CCaNm0KZ2dnVK1aFV999RX3WlHvI0uxfNH7BJh+Rp46dQqNGjWCQqGAUqnEW2+9hbS0NACFf46NHz8e586dw5IlS7jjA8DQoUPRr18/ODs7QyQSYdCgQahevTrOnTv30veAEEJKhBFCCHlj2rRpw2bPns39fPfuXRYcHMwWLVpksl1ERARzd3dneXl5bMaMGczPz4/pdDqTbTIzM5mjoyNbuXIlS01NZfb29mzt2rUm2wBgx44d435+8OABk0qlLDs7mzHG2IgRI5hYLGZr1qxheXl57MKFC8zZ2Zlt2bKFq5+NjQ1bv34902q17MKFC8zNzY35+/tzx5w7dy6rV68eu3XrFtPr9ezXX39lDg4O7O7duxZjkJWVxWxsbNiJEycYY4zl5ORw3xtjpFAo2JkzZ5her2crV65kDg4OrGvXriwmJoZpNBrWt29f1qlTJ26fQ4cOscjISKbT6VhycjLr3r07a9asGff6hg0bmI+PD/fziBEj2NChQy3WjzHGkpOTmaurK1u+fDnLzc1lycnJrH379mz06NHcNv7+/szPz49dvnyZ6fV6tmzZMubo6MhUKpXFY544cYLJZDKWmZnJlY0fP5716dPHbNuX1c8oJiaGAWDe3t5MqVSy5s2bs19//fWF+3Tv3p1VrVqVffvtt+yff/5heXl5xbr277//3mS/uXPnsrfeesviOQtub6xzdHT0C+s5d+5cJhaL2cqVK1leXh67ePEiE4vFLCQkhJ09e5aLubOzM8vKymKMMbZr1y7m6OjIjh8/znQ6HTt27BiTSqVsz549jDHG7t+/z2xtbdnatWtZXl4e+/PPP5mrq+sr3c/P30vPK+x1ACw4OJhFRUUxg8HAsrOz2bVr19jRo0dZdnY2y83NZXPnzmVyuZwlJiZajNWGDRuYSCRi06ZNY9nZ2SwhIYEFBQWxTz/9lDHGmMFgYKGhoWzIkCEsJSWF5ebmso8//pjVqFGDe5/HjBnDGjRowOLi4pharWbh4eFMJBKxDRs2WLwenU7H6tevzwYPHszS09NZWloaGzBgAAsJCeE+k0aMGMFsbW3Ztm3bmE6nY+fPn2disZidPHnS4jEzMzOZl5cX69ChA/v555/ZrVu3zLY5efIkUygU7Pjx40yv17Nr164xX19f7rOpsPvo+c9YI0uxtPQ+ZWdnM5lMxmJjYxljjHl7e7P169czg8HANBoNO3/+PFOr1UX6HLNUj4IePHjA7O3t2alTp164HSGElBQl/YQQ8ga1adOG2dnZMYVCwaRSKQPAWrVqxTIyMrhtdDod8/HxYR9//DFjLD/xFggEbO/evSbH+u6775i9vT1LTU1ljDE2fPhwVr9+fZNtnk/6V6xYwfr378/9PGLECNawYUOTfT7++GPWrl07xhhjn332GWvcuLHJ6x9++KFJkiSXy9nhw4dNtunQoQP77LPPLMYgKyuLOTg4sG+++YYlJydbjNGoUaO4n9PT0xkAtnXrVq5s165dzMnJyeLxGWPsypUrDAAX11dN+pctW2by0IAxxv744w9ma2vLJTr+/v5s3rx53OtqtZoBYH/99ZfFYxoMBlalShUu+c3KymIKhYIdOHDAbNuiJv2ZmZns/PnzTKPRsOzsbLZ582Zma2tr8ZgF67lkyRLWqlUr5uDgwKRSKRs5ciRLS0t7pWu3lPSLRCKmUCi4r7Fjx5pt/ypJf+XKlU3K6tevzx2TMcaePn3KALDIyEjGGGOdOnVi06ZNM9ln6tSprHPnzowxxhYsWGB2v3/wwQevdD+XJOlfs2ZNofsZKRQKtn//fsaY5UTVzs7O5EHN9OnTWZcuXRhjjF2+fJnZ2NiYPFjS6XTM3t6enTt3jun1emZnZ8cdn7H8f18CgaDQpP/8+fNMIBBwnzOM5cddIBCwCxcuMMby79e2bdua7NeoUSOzh5kFPXz4kL3//vusbt26TCwWMw8PD/bFF18wg8HAGGOsR48ebObMmSb7LFiwgLVv395ibIwKfsYavzZv3lzkpH/Pnj2sUaNG3M8BAQFs9uzZ7NGjRybbFeVz7EVJf3p6OmvYsGGR/p0TQkhJUfd+Qgh5w6ZPn4709HSo1WokJSXBw8MDnTt35l7ft28fEhISMHr0aABAlSpV0LZtW6xevdrkON999x3XVRQAxowZg8jISFy4cKHQcz/ftR8AAgMDzX42dnmOj4+Hv7+/yesFu/YnJiYiIyMDAwYMgJOTE/d1/vx5xMfHA4DJhFpffPEFHBwccPjwYRw/fhzBwcGoU6eOWRdjLy8v7nvjmPDnyzIzM7mfz5w5g/bt23PdwNu0aQMASEpKKjQWRl988YVJHYH8yQ4vX75sck1du3aFQCAwmdzL2EW+YD2N9Xr+ugUCAUaNGoUff/wRQP6QBEdHR3Tp0uWldSyMTCZD8+bNYWtrC4lEguHDh+Ptt9/Gli1bCt1HKpXio48+wtmzZ5GRkYH9+/fj+PHjmDZt2itduyXNmjVDeno697V27dpiXxtg+p4b627p3jDGPC4uDlWqVDHZJygoCA8fPgQAPHr0yOL9blSU+7kknj/3w4cPMXjwYFSqVAlyuRxOTk7IyMh44X3r6uoKGxsb7ueC/xaio6Oh0+ng6+vL1V2pVALIj01ycjI0Go1JPRQKBVxcXAo9X1xcHFxcXLjPGQBQKpVwdnbm4gqY/lt4vl6W+Pn5Yfny5bh69SrS09OxcOFCfPrpp9xwjOjoaKxatcrkfVi0aBEeP35c6DGNjJ+xxq/hw4e/dB+j5z8j9+/fj/v37yMkJARBQUGYO3cudDpdkT7HCpOcnIx27dohODjY6ldNIISUDSWftYcQQkixubm5YcSIEejRowdSUlKgVCq55L5Vq1bcdmq1GtnZ2bh79y6CgoLwxx9/4L///kNMTIzZZFOrV69G8+bNzc6VmJiIy5cvo3v37iblsbGxZj8bZ7X28fHB5cuXTV5/8OAB972TkxPs7e3x+++/o3Xr1havUa1Wm5W1atUKrVq1AmMMZ86cQZcuXVCzZk107NjR4jFeJC8vD927d8enn36KvXv3wtHREf/++y8aNmxYpEnBZs2ahVmzZpmUeXp6omXLlibjq1+VpesODw/H3LlzERUVhR9++AEjR46EUFi6z9+FQmGRJ0MTiURo164dBgwYgGPHjgEo2rWXpM6lfb0F+fn54d69eyZl9+7dQ6VKlQDkL033zz//mLxe8P4vyv38Mi+6vudfGzNmDBQKBf755x94eHiAMQZnZ+diT2bn6ekJW1tbJCcnmzwYMDIYDLCzs0NsbCxq164NAFCpVNwYdUv8/PyQlpaGtLQ0LvFPTU1FWloaF9eSkkqlGDlyJL755htcuXIF4eHh8PT0xNtvv41PP/3U4j4lvY8s7a/VavH777+bzEdQp04dbN26FQAQGRmJzp07w9fXF2PGjHnh51hh9YuLi0PHjh3RunVrrFmz5rX+eyCEECP6pCGEEB6lpaXhp59+gp+fH5RKJe7cuYOTJ09i48aNiIyM5L7u3LkDX19fbob+1atXo06dOrhz547JditWrMDOnTvx9OlTs3Pt2bMH7du3h6Ojo0n51atX8cMPP0Cn0+HixYv4/vvvMXLkSADA4MGDERkZic2bN3Ovb968mdvXzs4O48ePx8cff4ybN2+CMYacnBycPXsWd+7csXjNT548wc6dO5Geng6BQAAnJycIBIJizx6fl5eHnJwcODs7w9HREQkJCWZLZb2qkSNH4t9//8Xq1auRnZ0Nxhji4uKwd+/eEh3X29sbYWFhmDFjBs6fP49Ro0aZvJ6Xl4fc3Fzo9XoYDAbk5uZCo9EUerxz587h5s2b0Ov1yMvLw7Zt27B161a8/fbbhe7z/vvv4++//+au6/Lly/j111+5JLco1+7p6WkykeKrcHNzg1AoLPb+LzJ69GisX78ep0+fhl6vx8mTJ/Hjjz9i7NixAIC3334b165d4+73v/76q8T38/M8PT3x9OlTpKSkvHRblUoFmUwGZ2dnZGVl4ZNPPrH4sKioWrZsidq1a2PChAlcb4G0tDTs3r0b2dnZEAqFGDZsGCIiIhAfH4+srCx8+OGHEAgEhR6zSZMmqF27NiZPnoyMjAyoVCpMmjQJ9evXR+PGjYtVz1u3bmHBggW4ffs2dDodtFotduzYgevXr3P34XvvvYevv/4aJ06cgE6ng06nw/Xr17mZ+Et6H1l6n06cOAFfX19Uq1YNQP6/xw0bNiA5ORlAfq8IkUgEsVj80s8xT09Ps3vm9u3beOutt9C1a1esW7eOEn5CyBtDnzaEEPKGFZzRuVq1asjOzsahQ4cA5HfZr1q1KoYNGwZPT0/uy9vbGx988AE2btyIuLg47N69GzNmzDDZxtPTE+PHj4ezszPXhbwgS137AaBPnz7466+/4Orqin79+mH69OncGvFBQUH49ddfsWjRIjg5OWHmzJkYNWoU7O3tuf2//PJLvP3221yX6ICAACxcuLDQJckYY1izZg0qV64MmUyG/v374/PPP0fbtm2LFU+ZTIYffvgBCxYsgEwmQ1hYGAYMGFCsYxlVqlQJFy5cwLFjx1ClShU4OTmhc+fOr7SWfWFGjx6NAwcOoH379iZDJQCgU6dOkEgk2LJlC7Zt2waJRILg4GDu9S+++MJktvzbt2+je/fukMvl8PT0xKpVq/DTTz+9dMm/0aNHw8vLCwqFAoMHD8bgwYPx5ZdfFvnaP/30U+zbtw9OTk6oW7fuK12/RCLBF198wa0YMXHixFfa/0UGDBiAZcuWYeLEiXBycsKUKVOwatUq7r6vXLky9uzZg5UrV8LJyQmzZs3ChAkTTI7xqvfz89q1a4devXqhWrVqcHJy4lqJLfnqq69w9epVODs7o2bNmvDx8Sl07fiiEIlEOHbsGBwcHNC0aVM4OjqiXr162LNnD5fYr1ixAnXq1EGdOnVQrVo11KlT54VL04lEIvz+++/QaDQICgpC1apVodPpsH//fohEomLV09HREbdu3UJYWBicnJzg7u6OxYsX49tvv0X//v0BAL1798ZPP/2ETz/9FO7u7nB3d8fo0aO5B5olvY8svU+WPiN37dqFWrVqQSqVok2bNggPD8eIESNe+jn24Ycf4vbt23B2doaTkxMAYNGiRYiLi8O6detMhv+8bOUBQggpKQErbh8yQggh5UZqaip8fX3x6NGjF47fLYpp06bh5s2bOHLkSCnVjhBC+KXX6+Hl5YXjx4+/8oMsQggp66ilnxBCKoCUlBQsX768WAn/77//jqdPn0Kv1+P48eNYv349hg4d+hpqSQgh/EhNTcUHH3xACT8hxCpRSz8hhJAXmj17NtatW4fs7Gx4e3tj3LhxLx0HTAghhBBCygZK+gkhhBBCCCGEECtF3fsJIYQQQgghhBArRUk/IYQQQgghhBBipSjpJ4QQQgghhBBCrBQl/YQQQgghhBBCiJUS812B8iw9PR1nzpyBn58f7Ozs+K4OIYQQQgghhBArp9FoEBcXhzZt2sDJyeml21PSXwJnzpxB7969+a4GIYQQQgghhJAKZu/evejVq9dLt6OkvwT8/PwA5Ac7KCiI59oUTqvVQqVSQaFQwMbGhu/qVBgUd35Q3PlBcecHxZ0fFHd+UNz5QXHnB8WdH+Ul7nfv3kXv3r25fPRlKOkvAWOX/qCgINSqVYvn2hROq9UiJSUFSqWyTN+81obizg+KOz8o7vyguPOD4s4Pijs/KO78oLjzo7zFvahDzGkiP0IIIYQQQgghxEpR0l8BiEQiODk5QSQS8V2VCoXizg+KOz8o7vyguPOD4s4Pijs/KO78oLjzw1rjTt37KwChUAh7e3u+q1HhUNz5QXHnB8WdHxR3flDc+UFx5wfFnR8Ud35Ya9yppb8C0Ov1yMjIgF6v57sqFQrFnR8Ud35Q3PlBcecHxZ0fFHd+UNz5QXHnh7XGnZL+CsBgMCA7OxsGg4HvqlQoFHd+UNz5QXHnB8WdHxR3flDc+UFx5wfFnR/WGnfq3v+aMcaQmZmJjIwMaLVaMMbeeB0MBgPy8vKgVqshFNJznjeF4s6P8h53gUAAGxsbyOVyODo6QiAQ8F0lQgghhBBSjlHS/xrpdDrEx8cjOzsbACAWiyEUCt/4H/ECgQC2traUPLxhFHd+lOe4M8ag1+uRm5uLzMxMODg4wMfHB2IxfVQTQgghhJDiob8kX6O0tDRkZ2dDoVDA3d2dtz/cDQYD9Ho9RCJRuWz5LK8o7vywhrjrdDokJSVBpVIhLS0Nbm5ufFeJEEIIIYSUU+XzL+JyQq1WQyQSwcvLi9eWOoFAAJFIVC5bPsszijs/rCHuYrEYXl5eEIlEUKvVfFenSEQiERQKhdUtcVPWUdz5QXHnB8WdHxR3flDc+WGtcaeW/teIMQaxWMx78iEQCHivQ0VEceeHtcTd+PCCj3lAikMoFEIikfBdjQqH4s4Pijs/KO78oLjzg+LOD2uNO7X0VwDGccLlJXmwFhR3flhT3MvTwwu9Xo/MzEyrW+KmrKO484Pizg+KOz8o7vyguPPDWuNOSX8FwBiDwWCwiiSoPKG484Pizg+DwYCsrCyrW+KmrKO484Pizg+KOz8o7vyguPPDWuNOST8hhBBCCCGEEGKlKOknxRYREQGBQICnT5/yXZUyKTY2Fra2tti4cSNXZoxZUXXt2hVjxox5DbV7Mw4fPoz69evD3t4eAoEA6enpfFfpjWrWrBk+/vhjvqtBCCGknIuKikLr1q0hl8tRs2ZNnD59GgDw119/oW7dunBycoK7uztGjBhhMgFsQEAAHBwcIJPJIJPJMH78eIvH12g0GDVqFCpVqgS5XI5mzZrhwoULb+LSCCFvACX9hJRRf/75J44ePYoZM2ZwZTdu3EBERARiY2P5q1gRpaSkYODAgZBIJPj222/x008/QSqV8l2tN2rGjBn49ttv8eTJE76rQgghpJzSarXo1asX+vfvj7S0NHz11Vfo378/UlJSEBQUhEOHDiE9PR2xsbEwGAyYN2+eyf5Hjx6FWq2GWq3GmjVrLJ5Dp9MhICAAf/zxB9LT0zFt2jT06NGj3KwgQwh5MUr6KwCBQAChUFiuJgWzBsZ4FzfuS5cuRfv27REUFMSV3bhxA/PmzSsXSf8///yDzMxMfPbZZ3j33XcxbNgw2NjYvPbzlqX7vVevXpDL5Vi9ejXfVXntRCIRHB0drW6Jm7KO4s4Pijs/Kmrcb9++jbS0NEydOhUikQgdOnRAgwYNsGfPHri6usLHxwdA/pw2QqEQ9+7de+VzSKVSfPrpp6hUqRKEQiEGDx4MW1tb3L59u8LGnW8Ud35Ya9wp6a8ArGHd8vKoJEl/UlISDhw4gIEDBxb7/Iwx5OTkFHv/kkpKSgIAODk5ldoxs7KyXrrN677fi1IHI6FQiP79+2Pz5s1WP7GgUCiEVCqFUEi/Vt4kijs/KO78qMhxf/53CGMMUVFRAICHDx/CyckJMpkMu3fvxuTJk0227d+/Pzw8PNCnTx88ePCgSOeLjo5GamoqgoKCKnTc+URx54e1xt26roZY9LqXMHv69CkGDhwIuVwOpVKJ9957D7m5uSbb5OTkYOrUqXB1dYWjoyN69uyJ+Ph4CAQCREREmGwbHx+PUaNGwcPDA3Z2dqhVqxbWr19fpLpotVrMmzcPVatWhb29PZRKJVq2bIljx45x24SHh0Mmk+Hhw4fo3r07ZDIZfHx88O233wIArl27hnbt2kEqlcLf3x9bt241OUdqaiqmT5+OOnXqQCaTQS6XIywsDFevXjXZzhjv4sT9wIED0Ol06NChA1e2ceNGDBgwAADQtm1bbj1647i+gIAAdO/eHUeOHEGjRo0gkUiwdu1aAMCGDRvQrl07uLu7w87ODjVr1sR3331ndl7jMf744w80adIE9vb2qFy5MjZv3vxKcQ4NDcWIESMAAI0bN4ZAIEB4eDi3/86dOxESEgKJRAJXV1cMGzYM8fHxJucwvk/37t1D165d4ejoiKFDhwLIT+wnT56MnTt3ombNmpBIJGjevDmuXbsGxhi+++47BAUFwd7eHqGhoRZ7Rvz999/o0qULFAoFHBwc0KZNG/z5558m2xjnYLhx4waGDBkCZ2dntGzZEgDw5MkTjBw5Er6+vrCzs4OXlxd69epldq6OHTviwYMHiIyMNKuDNbHW2W7LOoo7Pyju/KiocQ8ODoaTkxOWL18OrVaLw4cP48yZM9xD6EqVKiE9PR2JiYn45JNPUKlSJW7frVu3IjY2FtHR0ahUqRJ69er10vjl5ORg2LBh+OSTT6BQKCps3PlGceeHtcZdzHcFyOtnXMLMmCSWtoEDByIgIAALFy7EX3/9ha+++gppaWkmiWJ4eDh27NiB4cOHo1mzZjhz5gy6detmdqzExEQ0a9aMS+rc3Nxw6NAhvPvuu8jIyMC0adNeWJeIiAgsXLgQo0ePRpMmTZCRkYFLly7hypUr6NixI7edXq9HWFgYWrdujSVLluDnn3/G5MmTIZVKMXv2bAwdOhR9+/bFmjVr8M4776B58+YIDAwEANy/fx979+7FgAEDEBgYiMTERKxduxZt2rTBjRs34O3tDaBkSf/58+ehVCrh7+/PlbVu3RpTp07FV199hVmzZqFGjRoAwP0fyO8C+Pbbb2PcuHEYM2YMgoODAQDfffcdatWqhZ49e0IsFuO3337DxIkTYTAYMGnSJJNz3717F/3798e7776LESNGYP369QgPD0dISAhq1apVpDjPnj0bwcHBWLduHebPn4/AwEBUqVIFQP7Di5EjR6Jx48ZYuHAhEhMTsWrVKvz555/4999/TXoG6HQ6dO7cGS1btsSXX34JBwcH7rVz585h//79XP0XLlyI7t2746OPPsLq1asxYcIEpKenY8mSJRg1ahROnjzJ7Xvy5EmEhYUhJCQEc+fOhVAo5B6MnDt3Dk2aNDGJyYABA1C1alV88cUX3PvZr18/REVFYcqUKQgICEBSUhKOHTuGhw8fIiAggNs3JCQEQP4cDQ0aNHiFu6B8Ma5ra2tra3VPx8syijs/KO78qKhxt7Gxwd69ezFlyhR8/vnnaNSoEQYNGgRfX1+T7dzd3dGlSxcMGTIEFy9eBAC0aNECAGBvb4/ly5dDoVDg3r17qFq1qsVzabVaDBgwAEFBQfj0008BVNy4843izg9rjTsl/Tz5/ux93H/6ZiZHYQxgzACBQAhLOX9lVxnGtK5c7OMHBgZi3759AIBJkyZxY5inT5+OunXr4sqVK9ixYwemTZuGFStWAAAmTpyIkSNHmrWOz549G3q9HteuXYNSqQQAjB8/Hm+//TYiIiIwbtw4SCSSQuty4MABdO3aFevWrXthnXNzc7mn2AAwZMgQeHt7Y9SoUdi2bRsGDRoEIL+Vtnr16ti0aRPXI6FOnTq4c+eOyQfB8OHDUb16dfz444+YM2fOK0TPslu3bpkkjgBQuXJltGrVCl999RU6duyI0NBQs/3u3r2Lw4cPo3PnziblZ86cMYnb5MmT0aVLFyxfvtws6b99+zbOnj2LVq1aAch/qOPn54cNGzbgyy+/BPDyOHfs2BHx8fFYt24dwsLC0KhRIwD5f0zMmDEDtWvXxtmzZ2Fvbw8AaNmyJbp3744VK1aYTECk0WgwYMAALFy40Owct2/fNomTs7Mzxo0bh88//xxRUVFwcnKCUCiEXq/HwoULERsbi4CAADDGMH78eLRt2xaHDh3iHoSNGzcOtWrVwv/+9z8cPXrU5Fz16tUz6fGRnp6O8+fPY+nSpZg+fTpXbryfCvLx8YGtrS1u3LhhMVaEEELIy9StWxdnzpzhfm7RogWGDRtmtp1Op8Pdu3ctHsPY+FNYY4TBYMDw4cMhEAiwadMmGhZKiBWxnscX5cz9p2pcj894M18JKkQlZOJ6gsri6yV9+PB80jhlyhQAwMGDBwHkL9sG5Cf6lrYzYoxh9+7d6NGjBxhjePr0KffVuXNnqFQqXLly5YV1cXJyQlRUFKKjo19a79GjR5vsFxwcDKlUajKO3til7v79+1yZnZ0dl/Dr9XqkpKRAJpMhODj4pfUrqpSUFDg7O7/yfoGBgWYJPwCThF+lUuHp06do06YN7t+/D5VKZbJtzZo1uYQfANzc3BAcHGwSg1eJc0GXLl1CUlISJk6cyCX8ANCtWzdUr14dBw4cMNtnwoQJFo/Vvn17kwcjTZs2BQD07dsXjo6OZuXG+kdGRiI6OhpDhgxBSkoKd49lZWWhffv2OHv2rFmXrueXOJJIJLC1tcXp06eRlpb20ut2dnampS0JIYQU23///Yfc3FxkZ2dj6dKlMBgM6NKlC37//Xfcvn0bjDEkJCRg9uzZaNeuHYD8sf4XLlyAVqtFVlYWPvroI1SqVInrefe8cePG4fHjx9i5cyfEYmoXBApfKvHAgQNo0aIFFAoFvL298cEHH0Cr1XL7devWDe7u7lAoFGjatGmRlj+8cOEChEIhFixY8Louh1Rg9C+aJ5VdZW/sXEVp6S+J57uIValSBUKhkBvb/ODBAwiFQq57vFHBWekBIDk5Genp6Vi3bl2hLcjGyeGeXwJNoVBAIpFg/vz56NWrF6pVq4batWujS5cuGD58OOrWrWuyvb29Pdzc3MyO4evra/ZkW6FQmCR2BoMBq1atwurVqxETEwO9Xs+9ZuydUBqKMyzg+Rgb/fnnn5g7dy4uXLiA7Oxsk9dUKhUUCgX3c8GxgEbOzs4mMShqnJ9nnEDIOOygoOrVq+OPP/4wKROLxWbdFwurp/Ea/Pz8LJYb6298UGGcc8ASlUpl8tDl+bja2dlh8eLF+PDDD+Hh4YFmzZqhe/fueOedd+Dp6Wl2PMYYtZgQQggptg0bNmDDhg0wGAzo2LEj9u7dCyD/76GpU6ciMTERCoUCYWFhWLJkCQAgMzMTY8eORUxMDDf3zf79+7lZyb/44gucO3cOhw4dwoMHD/DDDz/A3t4erq6u3HkPHTqEZs2avfHrLQuMSyVOnToVp06dwqlTp9C/f3/cvn0bGRkZiIiIQKtWraBWq9G3b18sXboUs2bNAgAsWbIEwcHBEIvF2L9/P3r16oXExMRC/xYwGAx4//330bhx4zd5iaQCoaSfJyXpTv+qjGP639QyZsU9h7F1ddiwYYUmZMak0svLy6R8w4YNCA8PR+vWrXHv3j3s27cPR48exQ8//IAVK1ZgzZo1Ji37hS3DUVh5wQT8iy++wJw5czBq1Ch89tlncHFxgVAoxLRp00xaiEsye79SqSxSC/LzLA19uHfvHtq3b4/q1atj+fLl8PPzg62tLQ4ePIgVK1aYtWoXJQZFjXNJFexV8bwXvYeW7nVj/Y3Xu3TpUtSvX9/iMWQy0wdhluJqXMN47969OHLkCObMmYOFCxfi5MmTZmP309PTTf6IskbWOtttWUdx5wfFnR8VOe4rVqzghkgWNHr06EJ/79aqVQvXrl0r9JjGBBUA/P39zRobdAYdcnW5SMtLg0akQaomFSKdCALkDxMQQAChIP+9EAqEJuUCgeBZ2XPbl5eH4AWXSgRgslRiwZhLJBIMHz4cv/32G1dmnAOJMQaRSITk5GRkZGSYNLIUtG7dOjRt2tSk92VFvt/5ZK1xp6S/AjAuYfa6REdHm7SE3r17FwaDget67e/vD4PBgJiYGJNeAc+POXNzc4OjoyP0er3JrPWWFJyNH3j24QoALi4uGDlyJEaOHAm1Wo3WrVsjIiKi1JLRXbt2oW3btvjxxx9Nyp9P7EqS9FevXh27d+82Ky/OsX777TdoNBrs37/fpHX81KlTr3ysgooTZ+PEhLdv3+a6Hxrdvn3bZOLC4nrZ/W7s1iiXy196n71MlSpV8OGHH+LDDz9EdHQ06tevj2XLlmHLli3cNvHx8cjLyzOZcNEaGde1JW8WxZ0fFHd+UNxfH71Bj1x9LnJ0OcjR5UCVq8I91T3EZsTiUeYjGJgBEhsJHMQOkIglcLD5//+LHUy+l4glEAlf/jen8WGAxYcGBR8QFHxoUJQHDXj2UKE0HkS8aKnEgs6ePWvytygAdO/eHceOHUNeXh7Gjx9faMKfkpKClStX4q+//jKZsJrud35Ya9wp6a8AGGNc9+LX8XT122+/RadOnbifv/76awBAWFgYAKBz586YPXs2Vq9ebfKU2ridkUgkQr9+/bB161Zcv34dtWvXNnk9OTmZ65JfWLKWkpJi0sVeJpMhKCgIcXFxJbhCUyKRyOyXwM6dOxEfH28yZKEks/c3b94cP/zwA+7fv4/KlZ/1CpFKpQDyHzC8Sn2fr4dKpcKGDRteuV5GxY1zo0aN4O7ujjVr1mDUqFGws7MDkN998ObNm9xMwSXxstUqQkJCUKVKFXz55ZcYMmSIWat+wfusMNnZ2RAKhSbzElSpUgWOjo7QaDQm216+fBnAsxmUrZXBYIBGo3lh7wxS+iju/KC484PiXjqeT/BzdblIzU1FbEYsHmQ8QKwqFgnqBDAU+DuGARAUrfHBVmRr9nDA+L2lhwbGhwliIb9pScEHDQAgdBdCKpdi1uezMGLcCFw4cwFnzpyBu6874jLiuL8zDu07hGPHj+HEhRNIyk7iHihs2rkJWq0Wxw4eQ3Z2NlQaFXd844MLgUCAmbNmYvLUyZDKpTAwAwzMwP3tTvf7m2etnzOU9FcAjDHo9XqIRKLXkvTHxMSgZ8+e6NKlCy5cuIAtW7ZgyJAhqFevHoD8JKtfv35YuXIlUlJSuCX77ty5A8D0F8iiRYtw6tQpNG3aFGPGjEHNmjWRmpqKK1eu4Pjx40hNTX1hXWrWrInQ0FCEhITAxcUFly5dwq5duzB58uRSu97u3btj/vz5GDlyJFq0aIFr167h559/NknOgZIl/d26dYNYLMbx48cxduxYrrx+/foQiURYvHgxVCoV7Ozs0K5dO7i7uxd6rE6dOsHW1hY9evTAuHHjoFar8f3338Pd3R2PHz9+5boBxY+zjY0NFi9ejJEjR6JNmzZ4++23uSX7AgIC8P777xerPs970f0uFArxww8/ICwsDLVq1cLIkSPh4+OD+Ph4nDp1CnK53KSLniV37txB+/btMXDgQNSsWRNisRh79uxBYmIiBg8ebLLtsWPHUKlSJaterg/Ij7lKpYJSqbSqX5JlHcWdHxR3flDcX13BBD9Xl/9/jV6DlNwUxKr+P8nPiEVKTkqhx3CTuEFsEEMr1CJbl40cXc4Lz5mnz0OePg/pmvRXqquN0Mas1wD3c4GHB1Ib6bMHCWIH2IhsXuk8hWFg0LNn8zRBBKzavAoLZy7E6mWrUat+LXTu3Rkuni7I0GYAAC6eu4gZ02Zg9bbVEMgFSM5JNjtuw44N0ad1H3gGeyKouul8Vjf/u4nzf53HlM+m4E7aHWTkZeBpzlPcSL0BpmNANuDv5Q9He+treS6rrPVzhpJ+UmLbt2/Hp59+ipkzZ0IsFmPy5MlYunSpyTabN2+Gp6cntm3bhj179qBDhw7Yvn07goODTVpLPTw8cPHiRcyfPx+//vorVq9eDaVSiVq1amHx4sUvrcvUqVOxf/9+HD16FBqNBv7+/liwYAE++uijUrveWbNmISsrC1u3bsX27dvRsGFDHDhwADNnziy1c3h4eKBr167YsWOHSdLv6emJNWvWYOHChXj33Xeh1+tx6tSpFyb9wcHB2LVrF/73v/9h+vTp8PT0xIQJE+Dm5oZRo0YVq34liXN4eDgcHBywaNEizJgxA1KpFH369MHixYvh5ORUrPq8qtDQUFy4cAGfffYZvvnmG6jVanh6eqJp06YYN27cS/f38/PD22+/jRMnTuCnn36CWCxG9erVsWPHDvTr14/bzmAwYPfu3Xj33XfLzRhGQggh5ZOBGbjE3pjkawwaGJgBCeqE/JZ8VX6Sn6XNsngMgUAAX5kv/OX+CFAEoJpTNShtlchMz4RELoFBaIBGr4Faq+bOk63LRo42h3sgkK3N5r43/mz8v7H3gCVagxYqjQoqjarQbSwRCUUmDwle1Jug4JAEG6HNS383B9cKxsbfNnI/Dw0bih4DegAA/rv8Hz4c/SGW/bgMtRvULuQI+XRaHeJi4syS/kvnLyHmXgza1ckf8qjOUEMkEiEuNg7zV8xHjjYHsapYOGocoZQo4WhLyT8pHgErTjMkAZC/jEft2rVx/fp1s3E8wLMlwp5vAX7TDAYD1/JZlp5YRUZGokGDBtiyZQuGDh3Kd3VKXUnjfu7cOYSGhuLWrVtmKySQwpWl+33v3r0YMmQI7t27Zzb5ZFGUlc+QotBqtdywDxub0ml1IS9HcecHxZ0fFPdnCib4ufpc5GhzoDHkDy/T6DWIy4jjuus/zHwIrV5r8Ti2IltUcqyEAEUAAuT5X872zpDZyCC1kcJGZGMx7gZmgM6gg9aghVavzf///38Zy01azZHf81Gj15g8JDB+zz08eO6hgfGBgYEZLFW/REQCkcnDAEtzFSTfS0bloMqwE9nh8JbDOHfoHLYd2Ybom9EY3Xc05q2Yh7ZhbU2O++jBI9y7fQ/NWjeDQCDA9o3b8fUXX+PgxYNw9TCd1DcnOwdZmc8ewCyctRC+/r549713IZPKkKPKgUQhgVCc//eMndAOSokSTnZO1JjwmpSXz5mX5aHPo5Z+8kbk5OSYzYC+cuVKCIVCtG7dmqdalW2tWrVCp06dsGTJEnz//fd8V4cUw+LFizF58uRiJfyEEEII8CzBL9hNX6PXcK3m6jx1oePxnyezleUn9/+f5HtLvSGzzU/wjd3mi0IoEMJWZAtbkS1QSF6kN+gtPgwo+JDgRS3/Rowx5BnyzHoTGB8UFPz++Z4Fzz94MKkf00Odp4Y6T13oNtfXX8fDEw8BBrjVc0OdyXXwvz//h/+++Q9pqWn4YMwH+P8pABBYPxCTv50MdaIa6xevR/yYeIhEIlQOrozlPy2Hi7sLAGDdinW48tcVrNm+BhIHCSQOz2JuL7GHg9QBcoUcBp35gw6NQYOErAQkZSfBxd4FzvbOvM+FQMoHaukvgfLS0v+ml+yzZN68ebh8+TLatm0LsViMQ4cO4dChQxg7dizWrl3LS51et7IQ94rImuJeVj5DikKv1yMrKwtSqfS1rhZCTFHc+UFx50dFiDtjjGu9LzgGv+Ckeim5KVyC/7Lx+K4SVwQoAuAv90egPBDO9s5wEDtAaiOFzFYGiVjCTSpXmNcZ98IeBhh/1jFdsY/NGIPWoDV7SGAcdpClzcr/3sLDBK3Bcs+IkhBAAHuxvcWVD8x6GYgdYC+yh4zJYOdQ+IRyQgjhZOcEpUSZ/wCGlFh5+Zyhln5i5nUv2VcULVq0wLFjx/DZZ59BrVajUqVKiIiIwOzZs3mt1+tUFuJeEVHc+SESiSCXy/muRoVDcecHxZ0f1hZ3xphJ672xNb9g67eBGfBY/RixGbFca35hLdMCgQA+Mh8EyPOTfH+5P2S2MtgIbbju+lIb6Su3DL/OuNsIbWAjtIEElnsYMMaePRQo+DCgQFlhDwYEAgHXG0FhZ3m5vMLoDDqLwwxe1NMgR5eDPH1eocdkYNx2ReVo64gQjxA08mwEF3sXs9cNMCBVk4pUTSrkNnIoJUo42Di80rUSU9b2OWNESX8F8LqX7CuKjh07omPHjrycmy9lIe4VEcWdHwaDAXl5ebC1teV9LoWKhOLOD4o7P8pz3I0JvjG5N7bmP9+9PU+fh7jMOG5m/YeZDwtNJAuOx/eX+8PP0Q+2IluIBCJIxVJIbfOTfDuRXYnqzmfcBQIBbEQ2L5yh38AMFh8GFOw9YMCrzQkgFooht5NDbvdqyZ/OoEOuLtd8foKCvQksTHyo0WvMjsXAkKHJwKm4UzgddxrVnKuhsWdjVHepDpHQvHEjQ5uBDG0GJGIJXCWukNtaX+L6JpTnz5kXoaS/AnjdS/YRyyju/KC480Ov1yM9Pd3qlrgp6yju/KC486O8xN04YR23TJ4+BxqdxmLiqc5Tc8vmPch4gHh1fKFL/cpsZVw3fX+FP7ykXty67xKxhGvNl4glpfr7r6zHXSgQwk5k98KHG8/PL2Bp8sGizC/wMmKhGDJbGWS2slfaT2/Qm66GoMtBanYqLsVfwmPNY0AA3Em7gztpd7jW/8aejeFs72x2rBxdDuIy42ArtOUm/XvZEA7yTFm/34uLkn5CCCGEEEKKwZjgG7vmZ+uyC03wGWNIzU01mXTvac7TQo/tKnHlls4LkAfAxd6FS+bthHbcBHwOYgeLLb/kGZFQBJFQBHvYF7rN8w8DTHoNlHB+gaLU7/mHBQa5AfUk9ZAuTsel5EuITIqERq9BZl4mTsedNmn9r6GsYZbY5xny8DjrMZKzk+Fs7wwXexea9K8Co3eeEEIIIYSQItDoNcjR5piMxS+s67iBGfA46zEeqJ615GfmZVrc1jge31/uz43JL5gAigXi/Mn3CiylR0qXcX6BwhgnBnzR5IMvWi2guLykXuil6IWwwDD8l/wfLj65iEeZjwAUrfVfx3RIzklGSk4KFHYKKCXKEg/5IOUPJf2EEEIIIaRIIiMjMWnSJFy7dg2urq6YNWsWRo8ejbS0NEyePBlHjhyBra0tZs6cialTp3L7ZWZm4v3338evv/4Kg8GAnj17YvPmzWbHT0pKwtSpU3HmzBnk5OSgUaNG+Prrr1GjRo03eZkAwLXgG7tdvyjBB/LH4z/KfJQ/6Z4q9oXj8W1ENvB39Oda8o3j8Y2EEMLBxuGVl9Ijr0/BiQELU3B+gcImH3zV+QWMbEW2aOTZCI08GyFBnYB/nvyDf5P+RZ4+z6z1v4lXE1R3qW7S+m+AAWmaNKRp0uBo4wilRAmpjbRYdSHlDyX9FYBxQjMa3/xmUdz5QXHnh1AohL29vVWNfysPKO78qMhxHz58OAYMGIBz584hMjISbdq0wVtvvYVFixbBYDAgLi4Ojx8/RocOHVCzZk106NABADBq1ChUqlQJsbGxkEgkuH79usXjq9VqNGvWDKtWrYKrqyuWL1+OXr164c6dO6817nn6PJPkPlef+9JW2yxtlsnSeS8ajy+1kSJAHsB11feSeZl1x5aIJK+0lN6bUpHv91dVlPkFzIYNPDf5IDe/gBAQ24kBC2H3lnmjV9DLW/8beTZCY4/GcLJ3Mtk/U5uJTG0mJCIJlBIl5LZy+rvp/1nr/S5ghX06kZd62fqI5WmNbUJI2UOfIYSQssbR0RFXrlxB1apVAQBNmjTBJ598gjFjxuDkyZOoW7cuAODzzz/HzZs3sWXLFty4cQPdu3dHdHT0Ky+pqtFoIJFIkJycDKVSWSrXUJwEnzGGtNy0Z+PxM2KRnJ1c6PZKiZJL8v3l/lDaK82SqpIupUesU8FlCtVaNVQaFfIMhS8FaPR8639B1ZyroalXUwS7BFt8mGQjtIGLvQtc7F3KzMMm8mIvy0OfR58uFQAtYcYPijs/KO78YIxBq9XCxsaG4v4GUdz5UZHjPmXKFGzZsgVz5szBlStX8PDhQzRr1gwATFq5GWOIiooCAPzzzz+oWrUqRowYgUOHDqFq1apYvnw5WrRo8dLznTt3Dh4eHlAqlcWKu1avRbYum0vuc3W5RZqQzcAMeJL1JD/J//8x+YWOx4cA3jJvLsEPkAdYnL29tJfSe1Mq8v3OB+MyhWKhGGImhpvEDTm6HGTkZUClURV6/xpb/7sEdslv/X98EfHqeADPWv/ltnJuiICTnRO3r9agRWJ2Ip7mPIWznTNcJC4vnN/Amlnr/U6PcioA4xJmpd2pIyIiAgKBAE+fFj7zbEUWExMDkUiEDRs2cGXGmBVV165dMWbMGO7n06dPQyAQ4PTp06VZ1VL1KveFTqfDxx9/DD8/PwiFQvTu3RtA/i+8iIiIYp3/dd3vJTFz5kw0bdqU72q8VjqdDqmpqdDpXt/sxsQcxZ0fFTnuYWFh2Lx5M+zt7dGiRQssXrwYXl5e6NKlCz7//HOo1WpER0fjxx9/RFZWFgAgPj4eR48eRdu2bfHkyRNMmzYNPXv2RHp6+gvP9fTpU4wbNw6LFi0C8PK4a/VaZORlIDErEQ8yHuB26m3cSb+DR+pHeJr7FGqtutCESavXIkYVg1MPT2HD9Q2Yf2E+vvn3G/x+73dce3rNJOG3EdmgilMVtKvUDqNqj8KnLT7FpAaT0K1yN9R2rc0l/AII4CB2gLvEHYHyQAQ7B8NP7gcXe5dyk/ADFft+51PBuDvYOMBT6olgl2AEyAPgbOcMkcByrxk7kR0aezbGpAaTMLnBZDTxasLNQ5CRl4GTD09i6cWl2BS1CTdTbsLAns0xoGd6PM19iui0aMSr45Gry30j11qWWOv9Ti39hJRRf/75J44ePYpbt27xXZXXZv369Vi6dCmmTZuGhg0bolKlSm/kvAkJCVi3bh169+6N+vXrm7x2+/ZtrFmzBn///TeuXLkCjUaDmJgYBAQEWDxWZmYmPvvsM+zcuRMJCQlwdXVF8+bNsXnzZjg4OAAApk2bhpUrV2L//v3o2bPna746Qgh5PVJTU9GtWzesX78effr0QVRUFLp06YI6dergq6++wpQpUxAUFASlUomhQ4fi/PnzAACJRIKAgAC8++67AIDBgwfj888/x99//43OnTtbPFdmZibCwsIwaNAgjBgxwux1rUH7rHv+/0+29ypLqmVrs7lu+g8yHuBR5iOT5Kcg43h8f0V+K76X1KvQJfJoKT3yuhmHg3gxL2RqM6HSqKDOU1ucINBb5o3eQb2fjf3//9Z/BobbqbdxO/W2xdZ/BoZ0TTrSNemQ2cigtFda7L1Cyg9K+gkpo5YuXYr27dsjKCiIK2vdujVycnJga1v4zLHlycmTJ+Hj44MVK1a80fMmJCRg3rx5CAgIMEv6L1y4gK+++go1a9ZEjRo1EBkZWehxVCoV2rRpg0ePHmHs2LEICgpCcnIyzp07B41GwyX9np6e6NWrF7788ktK+gkh5da9e/cglUrRv39/AEDdunXRokULnDlzBg0bNsTPP//MbTtr1iw0adIEAFC7dm2zXm4v6vWWk5OD7t27IyQkBF988QVXnpGXgeTsZKQKU8FERe/NxRhDmibNZOm8pOykQrc3jsc3zqxvaTy+ES2lR/giEAggt5VDbiuHgRmQmff/DwC06vyJAAswtv439myMeHU8Lj6+iKvJV5Gnz+Na/089PIVqLtXQxLOJydh/tVYNtVYNe5E9lPZKKOwUVtXtvaKgpJ+QMigpKQkHDhzAmjVrTMqNM4pai6SkJDg5OfFdDRPGLqeOjo748ssvX5j0f/LJJ3jw4AGuXLmCwMBArnzGjBlm2w4cOBADBgzA/fv3aWI+Qki5VK1aNWRnZ2Pfvn3o2bMnbt68iXPnzmH8+PG4e/cuXFxcIJfLcfToUWzevBmXL18GAISGhoIxhk2bNmHYsGHYs2cPHj9+bHHYk1arRb9+/eDt7Y3Vq1fnlxm0eKJ+gvScdOTociBhEghfMELVOB6fa8lXPUBGXobFbYs6Ht+IltIjZZFQIITCTgGFnQI6g457AJClyzLb1kfmgz5V+6Br5a64mnQV/zz5p9DW/8aejaGwUwAAcvW5iM+KR2J2IpQSZf4QA+rJUm7QmP4K4HUvYfb06VMMHDgQcrkcSqUS7733HnJzTccA5eTkYOrUqXB1dYWjoyN69uyJ+Ph4i2O34+PjMWrUKHh4eMDOzg61atXC+vXri1QXrVaLefPmoWrVqrC3t4dSqUTLli1x7Ngxbpvw8HDIZDI8fPgQ3bt3h0wmg4+PD7799lsAwLVr19CuXTtIpVL4+/tj69atJudITU3F9OnTUadOHchkMsjlcoSFheHq1asm2xnjXZy4HzhwADqdjlvqyMjSmP7Q0FDUrl0bly9fRosWLSCRSBAYGGj2wADInwV57ty5CAoKgp2dHfz8/PDxxx9Do9Fw24wYMQL29va4efOmyb6dO3eGs7MzEhISinQNL7ovYmNjIRAIcOrUKURFRXH354vmKvj3338RFhYGuVwOmUyG9u3b46+//jLZJjU1FR999BEaNGgAuVxu8b05ffo0GjduDAAYOXIkd+6NGzcCAFxcXODo6PjS60tPT8eGDRswduxYBAYGIi8vzySOzzO+l/v27XvpscsjoVAIOzs7q1vipqyjuPOjosZdoVBgx44dmDt3Lvf5+sEHH6BDhw64ePEiatasCYVCgfnz52P37t3w8PAAANjY2GDfvn345ptvoFAosHDhQuzdu5d76Dt+/HiMHz8eAHD+/HkcOnQI+/bt4z7vneROuB1zGxAAIlsR8Nyv1YLj8Tde34jPLnyGb/79Br/d+w3Xkq+ZJPw2wqKNxy9IIpLA1d4V/o7+CHYJhr/cH64S1wqT8FfU+51vxY27WCiGs70zAhQBqOZcDR4OHpCIzO9VO5Edmng1waQGkzCpwSQ09mxsNvZ/ycUl2By12WTsv47pkJidiOj0aDzJegKtXlvyiy1DrPV+p5b+CkAgEEAsfn1v9cCBAxEQEICFCxfir7/+wldffYW0tDRs3ryZ2yY8PBw7duzA8OHD0axZM5w5cwbdunUzO1ZiYiKaNWsGgUCAyZMnw83NDYcOHcK7776LjIwMTJs27YV1iYiIwMKFCzF69Gg0adIEGRkZuHTpEq5cuYKOHTty2+n1eoSFhaF169ZYsmQJfv75Z0yePBlSqRSzZ8/G0KFD0bdvX6xZswbvvPMOmjdvzrXk3r9/H3v37sWAAQMQGBiIxMRErF27Fm3atMGNGzfg7e0NoGRJ//nz56FUKuHv71+k7dPS0tC1a1cMHDgQb7/9Nnbs2IEJEybA1tYWo0aNAgAYDAb07NkTf/zxB8aOHYsaNWrg2rVrWLFiBe7cuYO9e/cCAFatWoWTJ09ixIgRuHDhAkQiEdauXYujR4/ip59+4q7vZV50X7i5ueGnn37iJn1auHAhAKBGjRoWjxUVFYVWrVpBLpfj448/ho2NDdauXYvQ0FCcOXOGay26f/8+9u3b98L3pkaNGpg/fz4+/fRTjB07Fq1atQKAIs0iXdAff/yB3NxcBAUFoX///ti7dy8MBgOaN2+Ob7/91mzYgEKhQJUqVfDnn3/i/ffff6VzlQcikQjOzs58V6PCobjzoyLHvXPnzhbH4Q8ZMgRDhgwpdL+6devin3/+sfhawYfUbdq0AWMMefo8JKgTzFoq7R3t88fjpz/rqh+fGV/oknuvMh7fiJbSM1WR73c+lUbcbYQ2cJW4wlXiijx9HlQaFVQaFTQG00aK51v/Lz65iAR1AhgYbqXewq3UW1DYKRDiEcK1/uuZHim5KUjNTYWjraPVPAiz1vu9Yn+KVRCvewmzwMBArvVy0qRJkMvlWL16NaZPn466deviypUr2LFjB6ZNm8aN3Z44cSJGjhxp1jo+e/Zs6PV6XLt2jVuPd/z48Xj77bcRERGBcePGQSIp/APlwIED6Nq1K9atW/fCOufm5mLYsGH45JNPAOT/seLt7Y1Ro0Zh27ZtGDRoEACgY8eOqF69OjZt2sT1SKhTpw7u3Llj8gRw+PDhqF69On788UfMmTMHwLOli4ozi/ytW7cKnTjOkoSEBCxbtgwffPABAGDcuHFo2rQpPvnkEwwfPhw2NjbYunUrjh8/jjNnzqBly5bcvrVr18b48eNx/vx5tGjRAk5OTvjxxx/RuXNnLFq0CEOGDMH06dPRu3dvDBs2rMh1etl9MWzYMPzwww8QiUQvPe7//vc/aLVa/PHHH1zX+HfeeQfBwcH4+OOPcebMGQD5783t27dNerc8/954eHggLCwMn376KZo3b/5K11RQdHQ0gPwu/lWqVMHmzZuhUqkwb948tGvXDlFRUfDy8jLZp3Llyrhx40axzlfWMcag0+kgFotprN8bRHHnB8X99WGMISU3BcnZySYTk6nz1DgddxrRadFIykmC4Pnm/v+ntFdyCf7LxuMbldel9N4Uut/5UdpxtxXZws3BDW4O+UsAqjQqZORlQGt41lJvbP1v4tUE8ZnxuPjk2dh/lUbFjf0PdglGE68mqOZcDUKBEBl5GcjIy4BULIVSooSj7ct7TJZV1nq/U9LPl/NfA0+j39z5DAagsG4qrlWBFlOKfehJkyaZ/DxlyhSsXr0aBw8eRN26dXH48GEA+Yn+89sZu1QD+f/Idu/ejYEDB4IxZrLkW+fOnfHLL7/gypUreOuttwqti5OTE6KiohAdHY2qVau+sN6jR4822S84OBh3797FwIEDufLg4GA4OTnh/v37XJmd3bM/BvR6PdLT0yGTyRAcHIwrV66YXE/B/7+KlJQU+Pj4FHl7sViMcePGcT/b2tpi3LhxmDBhAi5fvoxmzZph586dqFGjBqpXr24S23bt2gEATp06xbV2d+rUCePGjcP8+fOxa9cu2NvbY+3ata90DS+7L4pKr9fj6NGj6N27t8lYeC8vLwwZMgTff/89MjIyIJfLYWdnB4PBAL0+v8UnIyPD4ntTGtRqNYD8nhwnTpyATJbfJbRBgwZca/+CBQtM9nF2dsa///5bqvUoK3Q6HVJSUqBUKmFjQxNZvSkUd35Q3F+PXF0uErISkKPL4coYY4hMisTv939HtjYbBr0BQpEQEOSPx/eSeXEJvr/cv0jJhgACSMQSrjVfIpZY1R/3pY3ud368zrhLxBJIxBJ4Sj2Rrc3mHgAUXAHDx9EHfRxf3vrfyCN/5n+FnQJZuixkZWbBTmgHpUQJJzuncvdvy1rvd0r6+fI0Gnh89eXblRbGgNf0j+755LpKlSoQCoWIjY0FADx48ABCodBkojMAJrPSA0BycjLS09Oxbt26Qlvqk5LyZ9t98uSJSblCoYBEIsH8+fPRq1cvVKtWDbVr10aXLl0wfPhwsyTT3t4ebm5uZsfw9fU1+3BSKBRIS0vjfjYYDFi1ahVWr16NmJgYLsEEwPVOKA2v8rDA29sbUqnUpKxatWoA8sfPN2vWDNHR0bh586bZdRsZY2v05ZdfYt++fYiMjMTWrVvh7u7OvabX65GcnGyyvYuLi8mqAi+7L4oqOTkZ2dnZCA4ONnutRo0aMBgMiIuLQ61atWAwGLBy5Up89913r/W9AcD1OOnRoweX8ANAs2bNEBgYyC1VVZCxxw0hhJBnGGNIzknG05ynJrOOp+amYm/0XtxNv8uV+Tr4orpbdQQ4B6CSvFKRW+VpKT1CLHOwcYCDjQM8mSfUWjUy8jKQmZfJDZmx1PofmRwJrV4LlUaFEw9P4OTDkwh2CUZTr6ao6lwVGoMGCVkJSMpOgou9C5ztnSv8MBm+UfT54vriVuhS97KW/lJU3KTGYMjvxjds2DCLa/IC4JL357tNb9iwAeHh4WjdujXu3buHffv24ejRo/jhhx+wYsUKrFmzxqRlXySy/Mu+sPKCCfgXX3yBOXPmYNSoUfjss8/g4uICoVCIadOmcddQUkql0uRBQ2kwGAyoU6cOli9fbvF1Pz8/k5///fdf7kHAtWvX8Pbbb3OvxcXFmT3EOXXqFEJDQws9/5tIdo3vTXh4OObNmwdXV9dSf2+MjHMbGCeqKsjd3d3i+5eWlgZXV9dSrQchhJRn2dpsJKgTTMYYG5gB5+PP49iDY1zXY5mtDD0CeiBQFAgHJwcIxS+eZIuW0iPk1QgEAjjaOsLR1pFbAjBDk4FMbSb3MM7Y+h8WGIb/kv/D34//xuOsx2at/409G6ORRyPI7eRIyknC05yncLJzglKi5CYLJG8WJf18KUF3+lfF/r+7s0gkguA1zEQZHR1tkgDevXsXBoOBG5Pu7+8Pg8GAmJgYk9bfu3fvmhzHzc0Njo6O0Ov1ZrPWP6/gbPwAUKtWLe57FxcXjBw5EiNHjoRarUbr1q0RERFhkvSXxK5du9C2bVv8+OOPJuXp6emlltBVr14du3fvLvL2CQkJyMrKMmntv3PnDgBw70OVKlVw9epVtG/f/qUJeFZWFkaOHImaNWuiRYsWWLJkCfr06cPNeu/p6Wn2HtSrV8/k55fdF0Xl5uYGBwcH3L592+y1W7duQSgUcg8sjO/NunXrIBKJuHkXnn9vSuMBREhICID81Sael5CQgOrVq5uVx8TEmMWJEEIqIgMzICk7CSm5KSblCeoE7Ineg3j1s8/Wxp6N0TmgM+wF9shR5Tx/KAC0lB4hpangEoB6gz5/CcA8FbK0WWBgsBfbo4lXEzT2bIx49bOx/8bW/+MPjuPEgxOorqyOJp5NUNW5KlI1qUjVpEJuI4dSooSDjQPfl1mhWNdaBMSi171kn3GpO6Ovv/4aABAWFgYA3Cy/xvV2n9/OSCQSoV+/fti9ezeuX79udp6C3ck7dOhg8mVs+U9JMf3jQSaTISgo6IVLqb0qkUhk1vV+586dZslfSWbvb968OdLS0kzmEngRnU5nMuY+Ly8Pa9euhZubG5ecDhw4EPHx8fj+++/N9s/JyUFW1rMZkmfMmIGHDx9i06ZNWL58OQICAjBixAgujvb29mbvwfMznb7svigqkUiETp06Yd++fSZDAxITE7F161a0bNkScrmc2/b5SSstvTfGhyPp6emvVJeCgoODUa9ePezbt89kjoSjR48iLi7OZLUIAFCpVLh3794rrxJQXggEAtjY2NDwhTeM4s4PinvJqPPUuJt+1yTh1+q1OBJ7BN/++y2X8CslSoyuMxp9qvbJTxAEgNBGyC3ZV9GX0ntT6H7nR1mJu0gogpO9E/zl/qjqXBWeDp7cvzGBQABfR1/0rdoXnzT5BL2CesFT6gkAYGC4mXITm6I24ct/vsTJhyeRoclAhjYDMRkxuK+6D5VGxeelWVRW4l7aqKW/AnjdS/bFxMSgZ8+e6NKlCy5cuIAtW7ZgyJAhXItmSEgI+vXrh5UrVyIlJYVbss/YEl3wH9WiRYtw6tQpNG3aFGPGjEHNmjWRmpqKK1eu4Pjx40hNTX1hXWrWrInQ0FCEhITAxcUFly5dwq5duzB58uRSu97u3btj/vz5GDlyJFq0aIFr167h559/NplkruB1FedDo1u3bhCLxTh+/DjGjh370u29vb2xePFixMbGolq1ati+fTsiIyOxbt06bhKS4cOHY8eOHRg/fjxOnTqFt956C3q9Hrdu3cKOHTtw5MgRNGrUCCdPnsTq1asxd+5cNGzYEED+8InQ0FDMmTMHS5YsKdI1vOy+eBULFizAsWPH0LJlS0ycOBFisRhr166FRqMxqY/xvRkzZswL35sqVarAyckJa9asgaOjI6RSKZo2bYrAwECoVCruAcWff/4JAPjmm2/g5OQEJycnk3tpxYoV6NixI1q2bIlx48ZBpVJh+fLlqFatGiZMmGByzuPHj4Mxhl69er3y9ZcHYrG41OdNIC9HcecHxb149AY9nmQ/Qbom3aT8fvp97Lm7Byk5+Q8BBBCgtV9rtPNrZ9It30ZsA6WbkpbSe8PofudHWYy7jdAGSokSSokyv1U/L38JwFx9LuzF9mjq1RRNPJsgXh2Pvx//jf+S/4PWoEW6Jt1i63+OLgdJ2UncpH9CAf/t0WUx7qWBPi0rgIKt0q/jqdX27dvx6aefYubMmRCLxZg8eTKWLl1qss3mzZvh6emJbdu2Yc+ePejQoQO2b9+O4OBg2Nvbc9t5eHjg4sWLmD9/Pn799VesXr0aSqUStWrVwuLFi19al6lTp2L//v04evQoNBoN/P39sWDBAnz00Ueldr2zZs1CVlYWtm7diu3bt6Nhw4Y4cOAAZs6cabJdSWbv9/DwQNeuXbFjx44iJf3Ozs7YtGkTpkyZgu+//x4eHh745ptvMGbMGG4boVCIvXv3YsWKFdi8eTP27NkDBwcHVK5cGe+99x6qVauGzMxMjBo1Cg0aNMDs2bO5fVu1aoX33nsPy5YtQ9++fdGsWbOX1qko90VR1apVC+fOncMnn3yChQsXwmAwoGnTptiyZQuaNm3KbTdr1iyo1Wps27bthe+NjY0NNm3ahE8++QTjx4+HTqfDhg0bEBgYiLS0NG7ZRaNly5YByB+qUjDpb9u2LQ4fPow5c+Zg1qxZcHBwQO/evbFkyRKTyf2A/B4HLVu2RJUqVYoVg7KOMQaDwQChUGh1T8fLMoo7Pyjur06lUeFJ1hOT2cGztdk4HHsYl55c4sqM64V7y7xN9lfYKuDh4AEhhBT3N4zud36U9bjbiGzgKnGFq8QVubpcZORlQKVRIc+QB19HX/g6+qJb5W6ITIrExScX8STrCdf6fzPlJpzsnNDYszFCPEKQZ8hDcnYynO2d4WzvDBshf3NwlPW4F5eAFScjIQCAqKgo1K5dG9evXzcZU25k7Jr9fCvjm2YoMKZf+BrG9BdXZGQkGjRogC1btmDo0KF8V6fUlTTu586dQ2hoKG7duvXC5QdDQ0Px9OlTi0MiKqKyeL8/efIEgYGB+OWXX16ppb+sfIYUhVartcolbso6ijs/KO5FpzVo8UT9BBnaDJPy60+vY/+9/VDn5S9/aiO0QUf/jmjh08Kktc9WaAsvqRdktjKKO08o7vwor3HP1mZzDwCMD/kYY3ikfoSLjy9yrf9GAghQQ1kDjT0bo6pzVYgEIm7Sv6KuzlGaykvcX5aHPo9a+skbkZOTwy1xZrRy5UoIhUK0bt2ap1qVba1atUKnTp2wZMkSi+PwSfmxcuVK1KlTx2q79hNCiCVpuWlIzE7klv4C8lv899/bj5spN7myIKcg9K7aGy72LlyZAAK42LvA3cG9THT5JYQUjXEJQA8HD2TrsqHSqJCRlwE/Rz/4OfpZbP2/kXIDN1JuwNneGY08GiHEMwRpmjQ42jhCKckf0kNKhpJ+8kYsWbIEly9fRtu2bSEWi3Ho0CEcOnQIY8eONVsqjjxz6NAhvqtASsGiRYv4rgIhhLwxefo8PM56DLVWzZUZmAGXnlzCoZhD0OjzJ4WViCXoXrk76rvXN+lGKxFJ4CXzogn5CCnHBAIBN/+GF/NCpjYTKo0KQgjRzLsZmno1xaPMR7j45Fnrf1puGo49OIbjD4+jpktNNPZqjCCnIEjFUiglSsht5VbV5f5NoqSfvBEtWrTAsWPH8Nlnn0GtVqNSpUqIiIgwGTdOCCGEkPLtac5TJGcnwwADV5acnYw90XsQmxHLldVzq4dulbtBZvts/hMhhHBzcIPSXkl/2BNiRQQCAeS2csht5TAwQ/4SgBoVKskrwU9uofWfMUSlRCEqJYpr/W/k2Qgu9i5wsXeBs50zREIR35dVrlDSXwGUZBb50tKxY0ezJcys3ZuK++nTp1/r8cubsnC/V0TGVUIo7m8WxZ0fFHdzubpcJGQlIEeXw5XpDDqce3QOJx+e5Lr4K+wU6B3UG8EuwSb7O9o4wkvqZTJb//Mo7vyguPPDWuMuFAihsFNAYaeAzqBDRl4GMjQZJq3/+87uw45FO5DxIAO2cltE9ovE8U7HkbI3BRe2XICtnS0E//+fWp3fo+jnn3/GuHHjuPPo9XpoNBokJSXB1dXVrB4RERHYuXMnbt68ifXr1yM8PByA9cadkv4KwLjeJHmzKO78oLjzQywWW/ylSl4vijs/KO7PMMbyW/dzksHwbG7ouIw47I7ejaTsJK6shXcLdAzoaDI5l1gghqfUEwo7xUvPRXHnB8WdHxUh7mKhmGu9Ny4B6CB2wInFJ9C/R3+EDA3B4T8PY/+H++FS3QVJOUnwCfVB2w/borFnYzTyaIRHmY/gKnHF0KFDTSYGX7x4MQ4dOlRoDIOCgrB8+XJ8/vnnpnWy0rhT0l8BvO4l+4hlFHd+UNz5QXHnB8WdHxT3fNnabDzOeoxcfS5XptFrcCz2GM4nnOfK3B3c0a9qP/jJTefwcbZzhoeDR5G76VLc+UFx50dFi3vBJQAfxz3GqGGjoPRTooVPC0StjYJ7ljseCx4DyJ8k9GjsURx7cAw1XWqiiVcT1HOrBzeJGzdk6KeffsL7779f6PmGDRsGAFixYoVJubXGnZL+CoAxxi1hZk03b1lHcecHxZ0fOp2uXCxxY20o7vyo6HE3MAOSspOQkptiUn479Tb23t0LlUYFABAJRGhbqS1a+7aGWPjsT047oR28Zd5wsHF4pfNW9LjzheLOj4oc9ylTpuDg7oOYM2cOzl88j7QnaZjcazI2P92MLQe24MjfR2DnaodqA6qBNWfc2P/Gno3xlvdbyLifgXv37mHAgAGvfG5rjTsl/YQQQgghpEiytFlIUCcgz5DHlanz1DgQcwBXk65yZf5yf/Sp2gfuDu5cmQACuEpc4SZxo4eyhJBChYWF4Z133uG63v/4449oWaMlJMMlmDRhEkQyEfYc3oMFExdA6iaFIkhh0vqfsDUBjds3Rq44F1KDlCb9AyX9hBBCCCHkJfQGPRKzE5GmSePKGGOITIrE7/d/5ybwsxXZIiwwDI09G0MoEHLbOogd4C3zNhnPTwghz0tNTUW3bt2wfv169OnTB1FRUejSpQvq1KmDkIYhAPI/e/zf9seds3dgE2eD5p2b4+Lji0jKToJBZ8B/x/5D/Un1MePsDDTxbIL2/u0RKA984USh1o6SfkIIIYQQUqiMvAw8Vj+Gjum4stTcVOyN3ou76Xe5shrKGuhZpafJpHwigQgeDh5wtnd+o3UmhJRP9+7dg1QqRf/+/QEAdevWRYsWLXDmzBk0bNgQQP5Ye0dbRzjaOULhoMDAagPRuVJnRKVG4ec9PwMMcKvvhtTcVByOPYyjD46ilrIWQv1C0cyrmclSoRWF8OWbkPKOljDjB8WdHxR3fggEAgiFQor7G0Zx50dFibvWoEVcZhziMuO4hN/ADPgj/g+suryKS/hltjIMqTEEw2oMM0n45bZyVHGqUmoJf0WJe1lDcedHRY17tWrVkJ2djX379oExhhs3buDcuXOoU6cO9u/fD5VKBYPBgJMnT+Lnn39Gjx49oLBToJKiEjr5d4LmogaDBg1Cz6o9ueFFBmbAtafX8PW/X2PKySn44doPiEmNQW5uLgwGA7RaLfe9tcZdwApOUUheSVRUFGrXro3r16+jVq1aZq/fv38fAFC5cuU3XTVCiBWgzxBCCF/Sc9ORmJ1o0rqfoE7Anug9iFfHc2WNPBuhS0AXk0n5bIQ28JJ6wdHW8Y3WmRBiHY4cOYIZM2bg3r17cHFxwcSJEzFjxgwMHjwYR44cgU6nQ2BgIGbPno1BgwZx+6nVanh4eOQ/JKhfB6pcFa4+vYpzj85h0/xNYIyh3oR6AIB/v/oXcafiTM576tQphIaGvslLLbaX5aHPo6S/BCp60h8REYF58+YhOTnZKtezLKnY2FgEBgZiw4YNCA8PB/AsZkX9Z9e1a1f4+Pjg+++/f401tcxS/a1BeHg4Tp8+jdjYWF7OP3PmTJw6dQp///33S7e19s8QQkjZo9VrkZCVALVWbVJ2Mu4kzj46y/3+Utor0btqb1RxqmKyv9JeCXcHd5Px/IQQwqc8fR4eZz3GqYen8GfCn0jKTjJ53VXiinZ+7dAlsAuUEiVPtXw1r5r00ydyBWDstmIwGPiuSoVijHdx4/7nn3/i6NGjmDFjRmlWy8zWrVuxcuXK13qON+l13u8HDx5ERESExddyc3OxcOFC1KxZEw4ODvDx8cGAAQMQFRVlst20adNw9epV7N+/v9TrxyetVosnT55Aq9XyXZUKheLOD2uNe0pOCu6m3zVJ+GNUMfjq369wJu4MGGMQQIDWvq0xteFUk4RfIpKgsqIyPKWery3ht9a4l3UUd35Q3EuPrcgW/nJ/hNcOx1dtv8KMxjMQ4hHCzer/NOcpdtzZgTFHx+DzC5/jzO0z0ORpeK516aKJ/Agpo5YuXYr27dsjKCjotZ5n69atuH79OqZNm2ZS7u/vj5ycHKtaoxQAvv/++2I/EDh48CC+/fZbi4n/0KFDsX//fowZMwYNGzZEQkICvv32WzRv3hzXrl2Dv78/AMDT0xO9evXCl19+iZ49e5bkUgghpFRo9BokqBOQrcvmynJ0OTgccxj/PPmHK/OWeaNv1b7wlnlzZUII4ebgBqW90urGwBJCrI/ERoIWPi3QwqcFkrKTcPzBcZyKO4Wk7CTomR6XEi/h8pPLqOJZBb62vnxXt9RQ0k9IGZSUlIQDBw5gzZo1vNVBIBDA3t6et/O/Lq/jIUZ8fDx+/fVXTJ8+HUuXLuXKW7VqhXbt2uHXX3/F+++/z5UPHDgQAwYMwP3796nrPiGEN4wxPM15iuScZDA8G3Z2/el17L+3H+q8/BZ/sVCMTv6d0MKnhUkrvsxGBi+pF2xFtm+87oQQUlLuDu4YUmMIBgcPRmRyJI7EHsGVxCsIkgXBw8GD7+qVKureT0rs6dOnGDhwIORyOZRKJd577z3k5uaabJOTk4OpU6fC1dUVjo6O6NmzJ+Lj4yEQCMxaTePj4zFq1Ch4eHjAzs4OtWrVwvr164tUF61Wi3nz5qFq1aqwt7eHUqlEy5YtcezYMW6b8PBwyGQyPHz4EN27d4dMJoOPjw++/fZbAMC1a9fQrl07SKVS+Pv7Y+vWrSbnSE1NxfTp01GnTh3IZDLI5XKEhYXh6tWrxYieZQcOHIBOp0OHDh3MXrt//z4GDBgAFxcXODg4oFmzZjhw4IDJNqdPn4ZAIMD27dsxa9YseHp6QiqVomfPnoiLezZpSWhoKA4cOIAHDx5AIBBAIBAgICAAQP6YfoFAgI0bN3LblzR2ERERFluCNm7cCIFAYDLOPiAgAN27d8fp06fRqFEjSCQS1KlTB6dPnwYA/Prrr6hTpw7s7e0REhKCf//9t0ixDQ8P566x4HV++eWXWLduHapUqQI7Ozs0btwY//zzj8l+xus0xsp4LZmZmQAADw/TXxBeXl4AAIlEYlJufF/37dtXpDoTQkhpy9Hl4L7qPpJykriEX6VR4acbP2Hrza1cwh/kFIRpDaehpW9LLuEXC8TwlfnCX+5PCT8hpNwTCoVo6NEQnzT9BOs6rMPgwMF8V6nUUUs/KbGBAwciICAACxcuxF9//YWvvvoKaWlp2Lx5M7dNeHg4duzYgeHDh6NZs2Y4c+YMunXrZnasxMRENGvWDAKBAJMnT4abmxsOHTqEd999FxkZGWZd0J8XERGBhQsXYvTo0WjSpAkyMjJw6dIlXLlyBR07duS20+v1CAsLQ+vWrbFkyRL8/PPPmDx5MqRSKWbPno2hQ4eib9++WLNmDd555x00b94cgYGBAPKT7r1792LAgAEIDAxEYmIi1q5dizZt2uDGjRvw9vYurHpFdv78eSiVSq5LeMH4tGjRAtnZ2Zg6dSqUSiU2bdqEnj17YteuXejTp4/J9p9//jkEAgFmzJiBpKQkrFy5Eh06dEBkZCQkEglmz54NlUqFR48eYcWKFQAAmezFa5eWJHav6u7duxgyZAjGjRuHYcOG4csvv0SPHj2wZs0azJo1CxMnTgQALFy4EAMHDsTt27eLdR4gf5hDZmYmxo0bB4FAgCVLlqBv3764f/8+bGxsMG7cOCQkJODYsWP46aefTPatUqUKfH19sWzZMgQHB6NBgwZISEjAxx9/jMDAQAwebPrLQ6FQoEqVKvjzzz9NegAQQsjrZmAGJGUnITU3lUv2DcyAS08u4VDMIWj0+eNYJWIJulXuhgbuDUwe1jrZOcHDwQNiIf0JSQixPjIbGTwk1tXKD1DSz5tNUZsQo4rhuxoAgEBFIEbUGlH8/QMDuRbLSZMmQS6XY/Xq1Zg+fTrq1q2LK1euYMeOHZg2bRqXWE6cOBEjR440ax2fPXs29Ho9rl27BqUyf/bM8ePH4+2330ZERATGjRtn1mpa0IEDB9C1a1esW7fuhXXOzc3FsGHD8MknnwAAhgwZAm9vb4waNQrbtm3jlv/o2LEjqlevjk2bNnE9EurUqYM7d+5AKHzWUWb48OGoXr06fvzxR8yZMwcAuNcLbldUt27dMmmNNlq0aBESExNx7tw5tGzZEgAwZswY1K1bFx988AF69eplcr7U1FTcvHkTjo75yyY1bNgQAwcOxPfff4+pU6eiY8eO8PHxQVpaGoYNG1akupUkdq/q9u3bOH/+PJo3bw4AqFmzJjp37owxY8bg1q1bqFSpEgDA2dkZ48aNw9mzZxEaGlqsmD98+BDR0dFwds5fTzo4OBi9evXCkSNH0L17dzRv3hzVqlXDsWPHzGJlY2OD3bt3Y8iQISbj9ENCQnD+/Hk4OTmZna9y5cq4cePGK9ezrLKxsYGnpyff1ahwKO78KK9xz9JmIUGdgDxDHleWnJ2MPdF7EJsRy5XVc6uHbpW7QWb77CGwrdAW3jJvSG2kb7LKJspr3Ms7ijs/KO78sNa4U9LPkxhVDG6m3uS7GqVi0qRJJj9PmTIFq1evxsGDB1G3bl0cPnwYALhW2YLbFew6zhjD7t27MXDgwPxxhk+fcq917twZv/zyC65cuYK33nqr0Lo4OTkhKioK0dHRqFq16gvrPXr0aJP9goODcffuXQwcOJArDw4OhpOTE7d0GgDY2dlx3+v1eqSnp0MmkyE4OBhXrlx54TmLKiUlBT4+PmblBw8eRJMmTbiEH8hvmR87diw++eQT3LhxA7Vr1+Zee+edd7iEHwD69+8PLy8vHDx4EFOnTi12/Yobu1dVs2ZNLuEHgKZNmwIA2rVrxyX8Bcvv379f7PVVBw0axCX8QP54fOMxi8LZ2Rn169fHgAED0KxZM9y9excLFy7EgAEDcOzYMbP5EZydnYs8JIEQQkpCb9AjMTsRaZo0rkxn0OHco3M4+fAk9EwPAFDYKdA7qDeCXYK57QQQwFXiCleJKy3DRwgh5RQl/TwJVBSvu3NxGQyGQls/S1qX55PrKlWqQCgUcuOzHzx4AKFQaNbF+/lZ6ZOTk5Geno5169YV2lKflJS/ruaTJ09MyhUKBSQSCebPn49evXqhWrVqqF27Nrp06YLhw4ejbt26Jtvb29vDzc3N7Bi+vr5mY84VCgXS0p79oWQwGLBq1SqsXr0aMTEx0Ov13GvG3gkAuLWMjf9/VZb2e/DgAZfgFlSjRg3u9YJJ//PvjUAgQFBQUInWqC9J7F5VwcTeeDwA8PPzs1ielpYGxhgyMzORlZVlUp+XPbV9/lzGBwBFqb9KpUKrVq3w0Ucf4cMPP+TKGzVqhNDQUGzYsAETJkww2YcxZlUzXet0OqSmpsLFxQViMf1qeVMo7vwoT3HPzMtEgjoBOqbjyuIy4rA7erfJWtXNvZujU0An2ImePdiWiCXwlnrDXlw2JnUtT3G3JhR3flDc+WGtcbeeKylnStKd/lUZDAbo9XqIRKJidXt+VcVNZIzLqA0bNgwjRliOjzF5N06QZrRhwwaEh4ejdevWuHfvHvbt24ejR4/ihx9+wIoVK7BmzRqT1mmRSGTx+IWVF0zAv/jiC8yZMwejRo3CZ599BhcXFwiFQkybNs1kKbiSJP1KpbJEyfLrVJLYFXZvFHxwUtJzMcawc+dOk/f7+Xq86jFfZvfu3UhMTDRbgq9NmzaQy+X4888/zZL+tLQ0uLq6vvTY5QVjDAaDodgPuUjxUNz5UR7irjPo8DjrMTLyMrgyjV6DY7HHcD7hPFfm7uCOvlX7opL82YNPkUAEdwd3uNi7vNE6v0x5iLs1orjzg+LOD2uNOyX9pMSio6NNWvHv3r0Lg8HAjUn39/eHwWBATEyMScvz3bt3TY7j5uYGR0dH6PV6i7PWF1RwNn4AqFWrFve9i4sLRo4ciZEjR0KtVqN169aIiIgwSwKLa9euXWjbti1+/PFHk/L09PRSS+KqV6+O3bt3m5X7+/tbnKzu1q1b3OsFRUdHm/zMGMPdu3dNej68ydZmY+t5enq6yTj3Bw8elOp5OnbsiCNHjpT6Q67CYpWYmAjA/OEFYwx6vR46nc5sn5iYGNSrV69U60cIIUD+LPxPsp6YtO7fTr2NvXf3QqVRAchP7NtWaovWvq1NJuWT28jhKfOEjbD0lzclhBDCDxqcRUrMuIyZ0ddffw0ACAsLA5A/Hh8AVq9ebXE7I5FIhH79+mH37t24fv262XmSk5O57zt06GDyZWz5T0lJMdlHJpMhKCgIGo2mOJdmkUgkMnv6t3PnTsTHx5faOZo3b460tDSz8eRdu3bFxYsXceHCBa4sKysL69atQ0BAAGrWrGmy/ebNm7nl5ID8BxaPHz/m3hsAkEqlUKlUpVb3F6lSpQoA4OzZs1xZVlYWNm3aVKrn8fLyMrtHSoNUmj+BVXp6ukl5tWrVAAC//PKLSfn+/fuRlZWFBg0amJSrVCrcu3cPLVq0KJV6EUIIAGj1WjzMeIhH6kdcwq/OU2P77e3YFLWJS/j95f6Y0nAK2lVqxyX8YoEYfo5+8JP7UcJPCCFWhlr6SYnFxMSgZ8+e6NKlCy5cuIAtW7ZgyJAhXCtmSEgI+vXrh5UrVyIlJYVbsu/OnTsATFtPFy1ahFOnTqFp06YYM2YMatasidTUVFy5cgXHjx9HamrqC+tSs2ZNhIaGIiQkBC4uLrh06RJ27dqFyZMnl9r1du/eHfPnz8fIkSPRokULXLt2DT///DMqV65caufo1q0bxGIxjh8/jrFjx3LlM2fOxLZt2xAWFoapU6fCxcUFmzZtQkxMDHbv3m3Wsu3i4oKWLVti5MiRSExMxMqVKxEUFIQxY8Zw24SEhGD79u344IMP0LhxY8hkMvTo0aPUrqWgTp06oVKlSnj33Xfx0UcfQSQSYf369XBzc8PDhw9fyzlLU0hICABg6tSp6Ny5M0QiEQYPHowePXqgVq1amD9/Ph48eMBN5PfNN9/Ay8sL7777rslxjh8/DsYYevXqxcdlEEKsUGpuKpKyk7hJ+RhjiEyKxO/3f0eOLgcAYCuyRVhgGBp7NjaZlE9pr4SbxA0ioeVhToQQQso3SvorAIFA8Fonoti+fTs+/fRTzJw5E2KxGJMnT8bSpUtNttm8eTM8PT2xbds27NmzBx06dMD27dsRHBxsMqu5h4cHLl68iPnz5+PXX3/F6tWroVQqUatWLSxevPildZk6dSr279+Po0ePQqPRwN/fHwsWLMBHH31Uatc7a9YsZGVlYevWrdi+fTsaNmyIAwcOYObMmSbbGR9mFKf7vIeHB7p27YodO3aYJP0eHh44f/48ZsyYga+//hq5ubmoW7cufvvtN3Tr1s1iXf/77z8sXLgQmZmZaN++PVavXg0HBwdum4kTJyIyMhIbNmzAihUr4O/v/9qSfhsbG+zZswcTJ07EnDlz4OnpiWnTpsHZ2RkjR44slXO8zvu9b9++mDJlCn755Rds2bIFjDEMHjwYtra2OHfuHD777DMcOHAA27Ztg6OjI3r37o0vvvjCbNjHzp070bJlS67ngzUQi8Xw8LC+dW3LOoo7P8pS3DV6DRLUCcjWZXNlqbmp2Bu9F3fTnw2jq6GsgZ5VekJhp+DK7EX28JJ6wcHGAeVBWYp7RUJx5wfFnR/WGncBs7ZZCt6gqKgo1K5dG9evXzcZU25k7Jpdmi3A1iQyMhINGjTAli1bMHToUL6rU+acO3cOoaGhuHXr1kuXH3ze6dOn0bZtW+zcuRP9+/d/TTUkxfHkyRMEBgbil19+eWlLP32GEEIKwxhDSm4KkrKTwJD/p5yBGXA+4TyOxR6D1qAFAEhtpOgZ1BO1lbW5h9BCCOHm4AalvdKqVhEhhJCK4mV56PNoTH8FwBiDVqvldRbKnJwcs7KVK1dCKBSidevWPNTo9Stp3Fu1aoVOnTphyZIlpVwz61YW7vcXWblyJerUqWN1Xft1Oh2ePn1qcdJC8vpQ3PnBd9xzdDmIUcUgMTuRS/gT1An4LvI7HLx/kEv4QzxC8H7I+6jjWodL7qViKao4VYGrxLXcJfx8x72iorjzg+LOD2uNO3XvrwAKLh3H1y/4JUuW4PLly2jbti3EYjEOHTqEQ4cOYezYsWZrrluL0oj7oUOHSrNKFUJZuN9fZNGiRXxX4bVgjEGn05XZhy3WiuLOD77ibmAGJGcnIyU3hUv2tXotTsadxNlHZ7n6KO2V6F21N6o4PRtCJBaI4eHgASd7pzda59JE9zs/KO78oLjzw1rjTkk/eSNatGiBY8eO4bPPPoNarUalSpUQERGB2bNn8101QgghpMzL1mYjQZ0AjeHZajQxqhj8Gv0rUnLyV64RQIBWvq3QvlJ72IiezcCvsFXAU+ppsjQfIYSQioM+/ckb0bFjR3Ts2JHvalQYoaGhVveEkhBCKiK9QY+k7CSkap6tXpOjy8HhmMP458k/XJm3zBt9q/aFt8ybK7MV2sJL6gWZreyN1pkQQkjZQkk/IYSUUWV1iAIh5M3IzMvE46zH3Bh9ALj+9Dr239sPdZ4aACAWitHRvyPe8nmLW4ZPAEH+MnwObiZL8xFCCKmYKOl/jQQCAfR6Pd/VeO1L9hHLKO78sKa4M8YgFJaPP9jFYjHc3NzKTX2tBcWdH6877jqDDk+ynkCVp+LKVBoV9t/bj5spN7myIKcg9A7qDReJC1cmEUvgLfWGvdge1obud35Q3PlBceeHtcbdOv4yLqPs7OygUqmQl5cHW1tb3upBLYX8oLjzw1rinpeXB61WCweH8rF+tkAggEgk4rsaFQ7FnR+vM+4qjQpPsp5Ax/JnjjYwAy49uYRDMYeg0eeP55eIJehWuRsauDcwWYbP3cEdSonytdSrLKD7nR8Ud35Q3PlhrXGnpP81ksvlUKlUePz4Mby8vHhL/Blj0Ov1EIlEVpMQlQcUd35YQ9zz8vLw+PFjAPmfI+WBTqeDSqWCQqGwmp4W5QHFnR+vI+5avRaPsx4jU5vJlSVnJ2NP9B7EZsRyZXXd6qJ75e4m4/QdbRzhJfUymbzPGtH9zg+KOz8o7vyw1rhbz5WUQVKpFC4uLkhNTcW9e/dgY2MDgUDwxhMRxhg3Nri8JkHlEcWdH+U57sa6a7X543ddXFwglUp5rlXRGOtNE0i+WRR3fpR23FNzU5GUnQQ9yx8SqDPocO7ROZx8eJIrU9gp0CuoF6q7VOf2EwvE8JR6QmGnKJV6lHV0v/OD4s4Pijs/rDXulPS/RgKBAO7u7pBKpcjIyIBGo+HlBmKMcUMMylsSVJ5R3PlRnuMuEAggFArh4OAAuVwOqVRa7q6BEFJ0efo8JKgTkKXL4sriMuKwO3o3krKTuLLm3s3RKaAT7ER2XJmLnQvcHdwhElpfN1RCCCGli5L+10wgEEAmk0Em42+5HK1Wi5SUFCiVStjYWHfXv7KE4s4PijshpKxjjCElNwXJ2ckwwAAA0Og1OBZ7DOcTznPbuTu4o2/Vvqgkr8SV2Qnt4C3zhoNN+ZjvgxBCCP8o6SeEEEIIeUNydblIUCcgR5/Dld1OvY29d/dCpcmfrV8kEKFtpbZo7dsaYmH+n2oCCOAmcYOrxJV6ABFCCHkllPRXAGKxGEql0qomoygPKO78oLjzg+LOD4o7P4oTd8YYkrKTkJKbAob8oX7qPDUOxBzA1aSr3HaV5JXQt2pfuDu4c2VSsRReMi+T7v0VEd3v/KC484Pizg9rjbt1XQ2xSCAQUDdnHlDc+UFx5wfFnR8Ud368atyztdlIUCdAY8hfco8xhsikSPx+/3fk6PJb/G1FtugS0AVNvJpAKMhfH1okEMHDwQPO9s6lfxHlEN3v/KC484Pizg9rjbuQ7wqQ10+v1yMtLQ16vZ7vqlQoFHd+UNz5QXHnR0WNu3GuHOOXUCjEsmXLAOQn1BEREfDz84OTkxPGjBmDvLw8bt/k5GR069YNUqkUwcHBOHHiRKHn+fDDD1GlShU4Ojqibt26+P333wEUPe4GZsBj9WPEZMRwCX9qbio2XN+AnXd2cgl/dZfqeD/kfTTzbsYl/ApbBao4VaGEv4CKer/zjeLOD4o7P6w17mU66Y+OjsbgwYPh6+sLBwcHVK9eHfPnz0d2drbJdufPn0fLli3h4OAAT09PTJ06FWq12ux4Go0GM2bMgLe3NyQSCZo2bYpjx469qcvhjcFggEajgcFg4LsqFQrFnR8Ud35Q3PlRUeOuVqu5rzt37kAoFKJv374AgI0bN2Lnzp3466+/EBcXhydPnmD+/PncvpMmTYKnpyeSk5OxdOlSDBw4EKmpqRbP4+joiEOHDkGlUmHVqlUYNmwYYmJiihR3dZ4ad9PvIlWTf2wDM+CP+D+w6vIq3E2/CwCQ2kjxdo23MbzmcG7ZPRuhDSo5VoKvoy9shNbX2lQSFfV+5xvFnR8Ud35Ya9zLbPf+uLg4NGnSBAqFApMnT4aLiwsuXLiAuXPn4vLly9i3bx8AIDIyEu3bt0eNGjWwfPlyPHr0CF9++SWio6Nx6NAhk2OGh4dj165dmDZtGqpWrYqNGzeia9euOHXqFFq2bMnHZRJCCCGkBLZu3YrmzZsjMDAQAHDgwAGMGzcOPj4+AICZM2di0KBBWLBgAdRqNfbu3Yv79+/DwcEBPXv2RJ06dbBv3z6MHDnS7NgRERHc923btkXNmv/H3n3HR1Xl/x9/3WmZSe8NEgKhN0FRdl1F7GJFkF1UBJWVta1li1tcd9fV1f3pV9cuVqSoqCDCoojoLooKCtJBpJeEkJBeJpl27++Pm0wyTIJAQu5k5vP0kYfklpmb99xM5nPPuecMZO3atXTv3r3N4/GqXoqdxVS6Kv3LDtYeZMGOBRTWFvqXnZZxGmN6jvGPwq+gkGzXp+Frau0XQgghOkLIFv2zZ8+msrKSL7/8kkGDBgEwbdo0VFVl1qxZVFRUkJSUxJ///GeSkpJYvnw58fHxAOTl5XHLLbfwySefcNFFFwHw7bffMnfuXB5//HF+97vfATB58mQGDx7Mfffdx9dff936gQghhBAiZM2ePZs777wzYJmmaQH/LiwspKqqit27dxMbGxtQtA8ZMoQtW7b86PNUVFSwefNmBg4c2OY2Va4qDtUdwqt5AfD4PPz3wH/5ouAL/zGl2FMY22cs+Yn5/v0cZgdZsVk4LI5j+6GFEEKI4xCyl5Krq6sByMjICFielZWFyWTCZrNRXV3NsmXLmDRpkr/gB72Yj42N5d133/UvmzdvHmazmWnTpvmX2e12pk6dysqVKzlw4MBJ/omEEEII0ZE2btzI9u3bmTBhgn/ZJZdcwvTp09m3bx+VlZU8+uijANTV1VFbWxvweQEgPj6+1VsCW1JVlZtuuonx48czYMCAoPUe1cOB6gMU1Bb4C/49VXt4Zt0zfH7gczRNQ0FhVPdR3HXqXf6C34SJjOgMeib0lIJfCCHESROyLf2jR4/m//2//8fUqVN58MEHSUlJ4euvv+bFF1/krrvuIiYmhq+++gqv18uIESMC9rXZbAwbNox169b5l61bt46+ffsG/bE/44wzAP02gZycnDaPp6SkhMOHDwcs27lTvyfP4/Hg8Xj8y81mMyaTCVVVgwaBMJlMmM3mE16naRper/eY1ymKgsViISkpyX+sR6472n5H7tOkaVRLr9cb0KIC+lQXiqKc8Dqfzxd0H83R1jXlfaLrTtbrBPjPN4/Hc0yv04mua3qOI0Xq6xQfH++fx/pk/D7J6xS8zmQykZycjMlkCvoZjXrfi4TXSVGUgPeZluuMeN/r7Ndp5syZ/kH5mn7+m2++mQMHDnDOOefg9Xq55557WLZsGcnJyRw6dIjq6uqA16KyspLo6Gi9MG/jdbrzzjupqqrirbfe8j9PU+7l9eWU1Jfg8XlAhXpvPUv3LWV18WpQ9G772THZXJ1/NVkxWaCB6lWJj4onOy4bE6agnzHcXqeO/H1KSkrCYrGExPtepLxOFosl6H0Gus7fp678OjXl7vV6u9zfp678OiUkJITU+15r61p7/KMJ2aL/kksu4aGHHuKRRx5h0aJF/uX3338/Dz/8MABFRUWA3vp/pKysLFasWOH/vqioqM3tAA4ePHjU43nhhRd48MEHW11XVVVFWVmZ//u4uDhiYmKor6+npqYmYNuYmBji4uJwuVxUVVUFrIuOjiY+Ph63201lZWXAOrvdTmJiIh6PJ2jAoaioKJKSkvB6vQHHAfrJ1TTX5JEXLSwWC6mpqQBB+5lMJtLT01tdB5CZmQlAeXl50C9tU++MysrKoF+GtLQ0zGYzVVVVQSdrSkoKVquV6upqXC5XwLrk5GRsNhs1NTU0NDQErEtMTMRut1NXVxc0yGNCQgIOhwOn00ldXV3AulB8nVRVDVonr5O8TtB1Xqe6ujp5nTrxdXI6nRH7+6SqKm+99Rb/7//9v4BtMjMzefDBB7njjjtQVZXly5czZMgQKisr6dOnD7W1tXz//fekpaUB+kX/CRMmoKpqq6/TQw89xHfffcf//vc/XC6XvyeiR/VQXl+OEqNgtprxOD1sKtnEksIl1Hr1ngM2q42L8i5iROIIVJdKfVU9JsVEkj2J9Ph0rGYrNTU1Yf06tbau6XWCE/t9UhQlpN73wv11UhSF2trakHnfk9epa/x96uqvk8PhwOfzhezrdGQOP0bRjryUEELmzJnDnDlzGD9+PCkpKXz44YfMmDGDZ555hjvvvJPZs2czefJkvvnmG3+LfZPJkyezaNEi/4ufn59Pv379+OijjwK22717N/n5+fz73//mnnvuafNY2mrpHzt2LOvWrfOPOwChdaVKURQURaG6upro6GhMJlPAulC4UnXkunC5ouh2u6mpqSEuLg6TydSlWlK68uukqio1NTXEx8djs9m6TEsKdO3XCfTu0zExMUHLQ70l5ch10HVeJ7fbTXV1tf99puW6UG9Jae/rtGzZMq6//nr27dvnf91A/yBUXV1NTk4OW7ZsYdKkSTz66KNccsklWCwWfv7znxMXF8dTTz3Ff//7X6ZOncrWrVtJT08Pep0eeeQR5s6dy4oVK0hNTfVnWuGs4EDZASzRFswWM9XuahbtXMTWsq3+48hPyOfqvleT4tA/4KNCYlSif6C+SHmdoON+n1RVxel0Eh8fj6Zphr/vRcrr5PP5qKys9E+P2aSr/H3qqq9T0+eZuLg4zGZzl/v71FVfJ1VVqa2tJTExEZMpuCdWqHyO2LJlC8OHD2fz5s0BdWhbQralf+7cuUybNo3t27f7B9wZN24cqqryhz/8gWuvvRaHQ7//7cirTwANDQ3+9QAOh6PN7ZrWH016err/ys2RrFar/0VsqanQa82JrlMUpdXnOto6j8eDy+UiNja21fVHe0zgqOtaftDqqHVmsxmz2dxp607W69T0RmEymQIyPJHX8MfWgbxOTes8Hg9er9ffvb+jf5/asw7C93XyeDw0NDQQExPT5s/Yme977VkHXed1avoQcOT7zI/tFyp/n9qzbvbs2UycODHo7/fhw4e54oorOHjwINnZ2fzlL3/hiiuu8K9/4YUXmDJlCpmZmXTv3p133nnH38Ly5ptv8sgjj/gH9vv73/+OzWYjLy/Pv///e+b/MeqyUfi8PixYWH1oNUv2LMHlc6EoCg6Lg0t7Xcqp6af634ccFn2gvhhr8EWxcH+doON+n5o+z6iqetTHDNW/T8ezLpReJ1VV/bcptrZNqP99atLVXqemzzMmkykgx67y96mrvk5Nt2039f4y+n2vrXVHe/xW9z2urTvRCy+8wPDhw4Omxbnyyit54403WLdunb9rflM3/5aKiorIzs72f5+VlUVhYWGr2wEB2wohhBAitM2aNavV5QMGDPCPudOatLS0oF5/Ta6//nquv/56//cBrcyaSkFNATWeGlSvSqmrlI+3fMy+mn3+bYakDeGKXlcQa4sF9Pv5Ux2ppDnS/BcAhBBCiM4WsqP3FxcXB3XRgOauEl6vl8GDB2OxWFizZk3ANm63m/Xr1zNs2DD/smHDhrF9+3b/vXhNvvnmG/96IYQQQogjeXwe9lTt0Qt+TWV5wXJe2v4Se6v3ApAQlcDkQZO5tv+1/oI/2hJNfmI+6dHpUvALIYQwVMgW/X379mXdunVs3749YPnbb7+NyWRi6NChJCQkcMEFFzBnzpyAAR9mz55NbW1twBQ+11xzDT6fj5dfftm/zOVyMWPGDEaOHHnUkfuFEEIIEZmcHie7q3bT4GvA7XPz1vdvsWz/MnyaDwWFn2b/lHtOu4f+yf0BMCtmsmKy6JnQkyhzlMFHL4QQQoRw9/7f//73LFmyhLPPPps777yTlJQUFi9ezJIlS/jlL3/p747/z3/+kzPPPJNzzjmHadOmUVBQwBNPPMFFF13EJZdc4n+8kSNHMmHCBP70pz9RUlJC7969mTlzJnv37uW1114z6sfsFGazmcTExDbvnREnh+RuDMndGJK7MST3k6uyoZKDdQfR0Kh11zJr6ywKagpAgfSYdMb3G09eQp5/+3hbPJkxmVhNx3evpTg2cr4bQ3I3huRujHDNPaRH7//222/5+9//zrp16ygrK6Nnz55MmTKF++67L2Aggy+//JI//OEPrF27lri4OH7+85/z6KOPEhcXF/B4DQ0NPPDAA8yZM4eKigqGDh3KQw89xMUXX3xCx7dlyxYGDx58zKMmCiGEEKJrOFR3iLIGfcqlw87DvLHlDSoaKgDoldCL6wdej8OiDyJoNVnJiskizhbX5uMJIYQQHeV469CQLvpDXVcp+n0+n38qrXC7ahXKJHdjSO7GkNyNIbl3PJ/qo7C2kBqPftvgnqo9zNk6h3pvPQDD04dzVf5VaC4Nq8NKWnSafxo+cXLJ+W4Myd0Ykrsxukrux1uHyl+oCNA0r21r82qLk0dyN4bkbgzJ3RiSe8fy+Dzsrd7rL/g3lGzg9U2v+wv+83LP45q+12DBguJW6BHbg8yYTCn4O4mc78aQ3I0huRsjXHMP2Xv6hRBCCCE6i9Pj5EDNAbyaF03T+Lzgcz7Z+wmgz8c8rs84Tss4DYDEqETMqtnfvV8IIYQIZVL0CyGEECKitRywz6f6WLhrIWsO6dMBR5mjuH7A9fRO6g1AuiOdRGui/35/IYQQItRJ0S+EEEKIiNVywL4GbwNvb3ubHRU7AEiISuDGQTeSEZOBgkJWTBZJ9iQ8Ho+RhyyEEEIcFyn6I4DZbCYhISGkB6MIR5K7MSR3Y0juxpDcT9yRA/ZVuaqYuWUmh+oOAZAdm83kQZOJt8VjwkS3uG7E2+IByd0okrsxJHdjSO7GCNfcpeiPACaTCYdD7jvsbJK7MSR3Y0juxpDcT4zb5+ZAzQEafA0AHKw9yKwts6h2VwPQL7kfE/tPJMochVkxkxuXS7Q12r+/5G4Myd0YkrsxJHdjhGvuMtxsBPD5fNTU1ODz+Yw+lIgiuRtDcjeG5G4Myf34OT1O9lTt8Rf82yu28/LGl/0F/8iskdww8AaizFFYFAt58XkBBT9I7kaR3I0huRtDcjdGuOYuRX8EUFWVurq6sJt6ItRJ7saQ3I0huRtDcj8+FQ0V7K3ei1fzArD60Gpmbp6J2+cGYEzPMVyZfyUmxUSUKYpeCb2wW+xBjyO5G0NyN4bkbgzJ3Rjhmrt07xdCCCFE2Gs5YJ+qqXy671OWH1gOgNlk5uf9fs6Q1CEAOCwOcuNysZjkY5IQQoiuT/6aCSGEECJsHTlgn1f1Mm/7PDYe3ghAtDWaGwbeQI/4HgDEWePoHtcdkyKdIYUQQoQHKfqFEEIIEZaOHLDP6XEyZ+sc9lbvBSDFkcKUQVNIdaQCkBiVSHZMNoqiGHXIQgghRIeToj8CmM1m4uLiwm7qiVAnuRtDcjeG5G4Myb1tTo+TAzUH/Pfvl9eX88aWNyitLwUgNz6XyQMn+wfpS7WnkhGTcUyPLbkbQ3I3huRuDMndGOGauxT9EcBkMhETE2P0YUQcyd0YkrsxJHdjSO6tq2iooKiuCA0NgP3V+5m9dTZ1njoAhqQOYUK/Cf579jOjM0lxpBzz40vuxujs3J0eJ26fm0R7Yqc9ZyiS890YkrsxwjV3uWEtAoTrKJShTnI3huRuDMndGJJ7sEN1hzhYd9Bf8G8p3cIrm17xF/yjuo/iF/1/gcVkQUGhe2z34yr4QXI3Smfmrmoqr3z3Cs9/+7x/dodIJee7MSR3Y4Rr7tLSHwGa5pu02WyYTHKdp7NI7saQ3I0huRtDcm925IB9mqbx1cGv+Gj3RwAoKFzV+yrOyDoDABMmcuJyiLXFHv9zSe6G6Mzc522dx9JdS3H73KTHpPPLU38ZsWM9yPluDMndGOGauxT9QgghhOjS3D43+6v341JdgN5K++HuD1l5cCUAVrOV6/pfR7/kfgBYFAu58bk4LA7DjlmEri/3f8k7m9+hvL4cq8nKf7b/h27x3bi0z6VGH5oQQpwQKfqFEEII0WUdOWCf2+dm7ra5bCvfBkCcLY4pg6aQHZsNgM1ko0d8D2xmm2HHLELXjrIdvPzdy+yr2seQ9CGYFTMbSzYyc/1MsmKzGJ413OhDFEKI4xY+fRaEEEIIEVEqGirYW73XX/DXuGt4ZeMr/oI/IyaD24bd5i/47WY7eQl5UvCLVpU6S3nmm2fYengr/VP6YzPbMJvMDEgdwK6KXTy/+nkOVB0w+jCFEOK4SdEfAZpGoQyn+1K6AsndGJK7MSR3Y0Rq7pqmBQ3YV+Is4cX1L1JYWwhA78Te/Gror0iMSgQgxhJDXnweVpO13c8fqbkb7WTm3uBt4NlvnmX1wdXkJeQRY2sevdtusdMnuQ8bijfw9DdPU+2q7vDnD2VyvhtDcjdGuOYeXj+NaFW4zjcZ6iR3Y0juxpDcjRGJuftUHwdqDlDWUOZftqtyF9M3TKfSVQnAaRmnMWXQFOwWOwDxtnh6xPfAbOqYnCIx91BwsnLXNI3X1r7GF/u+INmRTEp08GwO8VHxdIvtxsoDK3lx9Yt4VW+HHkMok/PdGJK7McI1dyn6I4CqqtTX14fd1BOhTnI3huRuDMndGJGWu9vnZk/VHv8I/QDritcxY/MMGrwNAFzQ4wLG9RnnL/CTo5LJicvp0JHXIy33UHGycl/4w0KW7FyCqql0j+ve5nYZsRlEW6P5bM9nvLnxTTRN69DjCFVyvhtDcjdGuOYuRX8E8Pl8VFVV4fP5jD6UiCK5G0NyN4bkboxIyr3OU8eeqj3+Efo1TeO/+//Le9vfQ9VUTIqJCX0ncF7uef4CP92RTlZsVocfSyTlHkpORu7fFHzDm5ve5HDdYfqm9P3Ri0M9E3tS565jwbYFfLr70w47jlAm57sxJHdjhGvuMnq/EEIIIUJaRUMFRXVF/vv3faqPD3Z+wHfF3wH6Pdc3DLyBngk9AVBQyIrJIsmeZNgxi9C3p2IPL655kT0VexiSPgST8uNtYYqi0D+1PxuLNzJj/Qyy4rIYnD64E45WCCFOnLT0CyGEECIktTZgX4O3gTe2vOEv+JPsSdx6yq3+gt+Eie5x3aXgF0dVUV/B0988zeaSzfRN7ntcMzpYTBYGpA5ge9l2nv32WYpqik7ikQohRPtJ0S+EEEKIkONTfeyv2R8wYF9lQyXTN0xnV+UuALrFduO2U24jPTodALNipkd8D+Jt8YYcs+ga3D43z337HN8UfkNuQi5xUXHH/RgOq4P8pHzWFa3jqVVPUeeuOwlHKoQQHUOK/ghgMpmIjo4Ou6knQp3kbgzJ3RiSuzHCNfemAftqPbX+ZYW1hby44UVKnCUADEgZwC1DbyHWFguARbGQF59HtDX6pB9fuOYe6joid03TeGP9Gyzfu5wEWwJp0Wkn/FiJ9kQyYzL5uuBrXlzzIj41vO4BbiLnuzEkd2OEa+5yT38EMJvNxMdLq0dnk9yNIbkbQ3I3RjjmXuepo6CmAK/WPCXatvJtvL3tbTw+DwA/zf4pl/W6zH8PdpQpih7xPbCarZ1yjOGYe1fQEbl/uOND/rP9P7h9bvKT8tt9TFlxWdR56vh096dkxWZx/dDr2/2YoUbOd2NI7sYI19zD6xKGaJWqqjQ0NITd1BOhTnI3huRuDMndGOGWe0VDBfuq9wUU/KsOrmL2ltn+gv+yXpdxRf4V/oLfYXGQl5DXaQU/hF/uXUV7c19btJbZG2ZTXFtMv5R+HTaNY35SPtWuauZ/P5/le5d3yGOGEjnfjSG5GyNcc5eiPwL4fD4qKyvDbuqJUCe5G0NyN4bkboxwyb21AftUTWXJniUs2rUIDQ2LycL1A67nZ91+5t8vzhpHXnweFlPndlwMl9y7mvbkfqDqAM9/+zy7KnYxMHUgZpO5w45LURQGpA7gQNUBXl37KttKt3XYY4cCOd+NIbkbI1xzl6JfCCGEEIZpbcA+j8/D3G1zWVGwAoAYawy3DL2FQamD/NskRiWSE5dzTNOsichW7arm6W+eZmPxRvok9yHKEtXhz2ExWeif2p9tpdt45ptnKKkr6fDnEEKIEyV/KYUQQghhiNYG7Kvz1PHa5tfYXLoZgFRHKredchs5cTn+bVLsKXSL7dZh3bNF+PL4PLyw+gVWHlhJt/huxEedvHt1o63R9EzsyXcHv+PpVU/j9DhP2nMJIcTxkKJfCCGEEJ2uzlPHnqo9uFSXf1lpfSnT109nf/V+APIS8rj1lFtJdiT7t8mIziAzJrPTj1d0PZqmMWfjHD7b/Rkx1hgyYjJO+nMmO5JJi07jy/1f8sp3r6Bq4XVfsBCia5KiPwKYTCbsdnvYTT0R6iR3Y0juxpDcjdFVc29twL591ft4cf2L/m7+p6Sdws2Db/ZPwaeg0D22O6mOVEOOuaWumntXd7y5L9u9jA+2fUC9t568xLyTe3AtZMdlY1JMLN21lPlb53fa854scr4bQ3I3RrjmLlP2RQCz2UxiYqLRhxFxJHdjSO7GkNyN0dVy1zSNYmdxwP37AJsOb+Ld7e/65zkfnTOaC3tc6O++b8JETlwOsbbYTj/m1nS13MPF8eS+sXgjM9bP4GDtQYamD+3UW0EURaF3cm82lWzinS3vkB2Xzc9yf/bjO4YoOd+NIbkbI1xzD69LGKJVmqbhdrvRNM3oQ4kokrsxJHdjSO7G6Eq5tzZgn6ZpfFHwBW9vexuf6kNBYVyfcVyUd5G/SLMoFvIS8kKm4IeulXs4OdbcD9Yc5Llvn2NH2Q4Gpg7s9NkdAEyKiQGpA9hbuZeXvnuJHWU7Ov0YOoqc78aQ3I0RrrlL0R8BvF4v5eXleL3eH99YdBjJ3RiSuzEkd2N0ldxbG7BP1VQW7lrIx3s+BsBmtnHj4BsZkTnCv43VZKVnQk8cFkenH/PRdJXcw82x5F7rruXpVU+z7tA68pPysVvsnXiEgWxmG/1T+rP18Fae+eYZypxlP75TCJLz3RiSuzHCNXcp+oUQQghx0rQ2YJ/L52LWlll8W/QtAPG2eH419Ff0Serj38ZuttMzoSc2s63Tj1l0TV7Vy/Q10/nqwFdkxWaRaE80+pCIscWQl5DH6oOrefbbZ3F5XT++kxBCdDC5p18IIYQQJ0V5QzmH6g6h0dxNstpVzcwtMymqKwIgMyaTKYOmkBCV4N8mxhJDTlwOZpO5049ZdE2apjF381yW7VqG3WwnKzbL6EPyS4lOwelx8vnez0mLTuP202+X6SaFEJ1KWvqFEEII0aE0TeNQ3SGK6ooCCv5DdYd4ccOL/oK/b1JffnXKrwIK/nhbPD3ie0jBL47L8r3Lmf/9fGrcNfRK6mX04QTpHt8dVVNZsnMJC39YaPThCCEijBT9EcBkMhEVFRV2U0+EOsndGJK7MSR3Y4Ri7q0N2Aews2InL214iSpXFQCnZ57O5EGTiTJH+bdJjkomJy4n5FtBQzH3SNBW7lsPb+XVta9SUFXAgNQBIXn+KIpC35S+HK47zJub3uTbwm+NPqRjJue7MSR3Y4Rr7tK9PwKYzWaSkpKMPoyII7kbQ3I3huRujFDL3e1zs796f8D9+wBrDq1hwc4F/tGQL867mFHdRwUUZ+mOdNKi0zr1eE9UqOUeKVrLvbi2mGe/eZYfyn5gQOoAQ0bqP1YmxcTAtIFsKtnEi6tfJD0mnbzEPKMP60fJ+W4Myd0Y4Zp7eF3CEK3SNA2PxxN2U0+EOsndGJK7MSR3Y4RS7q0N2KdpGsv2LeP9He+jaRpmxcwv+v+Cc3LO8Rf8CgrZMdldpuCH0Mo9khyZu9Pj5JlvnuG7ou/oldiLaGu0wUf442xmG32T+7K5ZDNPrXqKivoKow/pR8n5bgzJ3RjhmrsU/RHA6/VSVlYWdlNPhDrJ3RiSuzEkd2OESu7lDeXsq96HV2s+Dq/q5b3t7/G//f8DwGFxMHXIVE5JO8W/jQkT3eO6k2TvWq0qoZJ7pGmZu6qpvLTmJVbsX0F6TDpJjq5zDsVFxZETn8M3hd/w/OrncfvcRh/SUcn5bgzJ3Rjhmnvo9oESQgghREjTNI1iZ3HQ/ftOj5M3v3+TPVV7AEiyJ3HT4JtIdaT6tzErZnLjcrtE66wIPe9teY9Pdn+CRbGQHZtt9OEct7SYNJweJ//b8z8yYjL45am/DMmxCIQQ4UFa+oUQQghx3NoasK+8oZyXNr7kL/hz4nK47ZTbAgp+i2IhLz5PCn5xQlYeWMm7W96lor6C/OT8Llss5ybk4va5+c/2//DRjo+MPhwhRBiTln4hhBBCHJe2BuwrqClg5paZ1HnqABiUMoif9/s5VrPVv02UKYoe8T0ClglxrAqrC5mxeQb7qvYxNH0oJqXrtl8pikK/lH5sLNnIrA2zyI7LZnjWcKMPSwgRhrruO6U4ZoqiYLVau+yV8K5KcjeG5G4Myd0YRuRe56ljd9XuoIL/+7LveWXjK/6C/6xuZ3HtgGsDinuHxUFeQl6XL/jlfDdGeUM5729/n+9Lv6d/Sv8ufx4BmE1mBqQOYFfFLp5b/RwHqg4YfUhB5Hw3huRujHDNXVr6I4DFYiElJcXow4g4krsxJHdjSO7G6OzcyxvKOVR3CI3AUY1XHlzJ4l2L0dBQULg8/3J+mv3TgG1irbHkxOV06ZbZJnK+G+PDnR/yRckXNKgNbCvb1mGPa1bMDMscdlz7bD28FafH2WGFgVf1srVkK+9tfY/f/PQ3HfKYHUXOd2NI7sYI19yl6I8AmqahqiomkynsrlqFMsndGJK7MSR3Y3RW7pqmcajuEOWu8oDlqqayZM8Svir8CgCrycrE/hMZkDIgYLvEqESyY7LD5tyQ890Yo3JHUVVfpV906qDY91Ts4esDXx/3fl7Vy7CsYQxKG9QxB9Logl4XdOjjdQQ5340huRsjXHOXoj8CNE09kZKSgtXa9bvCdRWSuzEkd2NI7sbojNx9qo+C2gJqPbUBy90+N+/+8C5by7YCEGuLZcrAKXSL6xawXYo9hcyYzJNybEaR890YPeJ7cG2vazs09weXP3jC+9rN9pBrlT8Z5Hw3huRujHDNXYp+IYQQQrTK5XNxoPpA0P37te5aZm2dRUFNAQDp0elMGTSFJHvgXOkZ0RkBo/YLIYQQovNJ0S+EEEKIILXuWgpqC/BpvoDlh52HeWPLG1Q0VADQK7EX1w+4HofF4d9GQaFbbDcSohI69ZiFEEIIEUyKfiGEEEIEaGvAvj1Ve5izdQ713noAhqcP5+o+V2MxNX+cMGEiJy6HWFtspx6zEEIIIVonRX8EUBQFi8USVoNRdAWSuzEkd2NI7sbo6NzbGrAPYEPJBuZtn+dv+T8v9zzOzz0/4LktioWcuByirdEdcjyhSs53Y0juxpDcjSG5GyNcc5eiPwJYLBZSU+Weys4muRtDcjeG5G6Mjsy9rQH7NE3j84LP+WTvJ4D+gWhcn3GclnFawHZWk5Ue8T2IMkd1yPGEMjnfjSG5G0NyN4bkboxwzV2K/gigac3dM8PtqlUok9yNIbkbQ3I3Rkfl3taAfT7Vx8JdC1lzaA0AUeYorh9wPb2TegdsZzfbyY3PxWoKn5GOj0bOd2NI7saQ3I0huRsjXHM3GX0A4uTzer0UFxfj9XqNPpSIIrkbQ3I3huRujI7IvdZdy56qPUEFf4O3gVlbZ/kL/oSoBG495daggj/GEkNefF7EFPwg57tRJHdjSO7GkNyNEa65S0u/EEIIEaHaGrCvylXFzC0zOVR3CIDs2GwmD5pMvC0+YLt4azzd47qHVWuIEEIIEW6k6BdCCCEizNEG7DtYe5BZW2ZR7a4GoF9yPyb2nxh0r35yVDJZsVmdcrxCCCGEOHFS9AshhBARxKf6OFBzgDpvXdC67RXbeev7t3D73ACMzBrJFflXYFIC7wZMd6STFp3WKccrhBBCiPaRoj8CKIqCyWSS7pedTHI3huRuDMndGMebu8vnYn/1ftyqO2jd6kOr+WDHB/6u/mN6juGsbmcFPLaCQmZMJsn25I75AbooOd+NIbkbQ3I3huRujHDNXYr+CGCxWEhPTzf6MCKO5G4Myd0Ykrsxjif3WnctBbUF+DRfwHJVU/l036csP7AcALPJzM/7/ZwhqUMCtjNholtct6D7+iORnO/GkNyNIbkbQ3I3RrjmLkW/EEIIEebK6ssodhYHDdjnVb3M2z6PjYc3AhBtjeaGgTfQI75HwHZmxUxuXC7R1uhOO2YhhBBCdAyZsi8CeDweDh06hMfjMfpQIorkbgzJ3RiSuzF+LHdN0yiqLeKQM3iEfqfHyeubXvcX/CmOFG495daggt+iWMiLz5OCvwU5340huRtDcjeG5G6McM1dWvqFEEKIMORVvRTUFLQ6YF95fTlvbHmD0vpSAHLjc5k8cHJQYR9liqJHfA+sZmunHLMQQgghOp4U/UIIIUSYOdqAffur9zN762zqPPrFgCFpQ5jQdwIWU+BHAofFQW5cbtByIYQQQnQt8pdcCCGECCNtDdgHsKV0C3N/mItP1deN6j6Ki/IuCpqSL9YaS05cTtByIYQQQnQ98tdcCCGEaKfHHnuMnJwc4uLiGD58ODU1Nf51Xq+XIUOG0Lt374B9ampq+OUvf0lycjKJiYlMnjy5zce/7LLLSE9PJyEhgZEjR7Jy5cpWtyurL2N/zf6ggl/TNL4s/JI3v38Tn+pDQWFs77Fc0vOSoMI+MSqR3LhcKfiFEEKIMCEt/RHAarWSmZlp9GFEHMndGJK7MSI59+eff56PP/6Yr776ipycHDZt2oTNZvOvf+6550hISKC+vj5gv5tvvpnc3Fz27t2Lw+Fg8+bNbT7HY489Rr9+/bBYLCxatIirrrqK4uJif+5NA/aVu8qD9lU1lQ93f8jKg/qFApvZxrX9r6Vfcr+gbVPsKWTGRObreDwi+Xw3kuRuDMndGJK7McI1dyn6hRBCiBPk8/n45z//yYoVK8jNzQVg6NCh/vXFxcW8/PLLPP7449x9993+5Vu3buW7775j7ty5mM1mAIYPH97m8wwaNAjQW+zNZjOHDx+murqahISEow7Y5/a5mbttLtvKtwEQZ4tjyqApZMdmB22bEZ1BqiP1BFIQQgghRCiTvnsRwOv1UlJSgtfrNfpQIorkbgzJ3RiRmntBQQFOp5N58+aRkZFBv379eOWVV/zr//CHP/DnP/+ZmJiYgP1Wr15Nnz59mDJlCikpKfzkJz/h66+/PupzXX755djtdi6//HJuvfVWEhISqGmoYdPeTdS4aoK2r3HX8MrGV/wFf0ZMBrcNuy2o4FdQ6B7bXQr+4xCp57vRJHdjSO7GkNyNEa65S0t/BNA0DVVV0TTtxzcWHUZyN4bkboxIzb2wsJCqqiq2b9/O3r172bFjB+effz79+/fHYrGwY8cOZsyYweeffx603yeffMKrr77KjBkzmD9/PldeeSU7d+4kMTGx1edavHgxbrebxYsXU1tbS2l9KYeqD+HxerBogX/Oi+uKmbllJpWuSgB6J/bmugHXYbfYA7YzYSInLodYW2yHZRIJIvV8N5rkbgzJ3RiSuzHCNXdp6RdCCCFOkMPhAOCvf/0rDoeDoUOHMnHiRD766CPuuusunnrqKRRFaXW/vLw8pk6ditVqZeLEiWRlZfHNN98c9flsNhuXX3U5j/y/R/hq7VdoBH8o2VW5i+kbpvsL/tMyTmPKoClBBb9FsdAjvocU/EIIIUSYk5Z+IYQQ4gT17dsXm80WUNgrikJlZSVr167liiuuAMDtdlNdXU1mZibbt29n8ODBQRcDWrs4cKTyhnKK64pxe9wc2HOAXr17BaxfV7yO+Tvmo2oqABf2uJDROaODHttqstIjvgdR5qgT+rmFEEII0XVIS78QQghxgmJiYrjmmmv45z//icvl4vvvv+edd97huuuuo7CwkPXr17N+/XpeffVVcnJyWL9+PXFxcYwePRpN05g5cyY+n4958+ZRVFTEyJEjg55jz549LPzPQraXbGdf+T5mvjST4oPFDDl1iH8bTdP47/7/8t7291A1FZNiYkK/CZybe25QwW832+mZ0FMKfiGEECJCSEt/BLBYLGRkZBh9GBFHcjeG5G6MSM79+eefZ+rUqaSmppKamspDDz3E2WefHbBNcnIyZrPZPw2Q1Wpl4cKFTJ06lTvuuIN+/frxwQcf+O/nv/XWWwGYPn061Q3V/O0ff2Pntp2YzWZ6D+jN8289T2pGKpqmEZUYxYJdC1hbvBYAu8XODQNvoGdCz6BjjbZEkxuXi9lkPomJhL9IPt+NJLkbQ3I3huRujHDNXYr+CHAsXUZFx5PcjSG5GyOSc09MTGT+/PlH3Wb06NHs3LkzYNnQoUNZvXp1q9tPnz4dn+qjoKYAS7qFNz9+s9XtXD4Xb37/JrsqdwGQZE9iyqAppEenB20bb42nW1w3TIp08muvSD7fjSS5G0NyN4bkboxwzV3+8kcAr9dLaWlp2E09Eeokd2NI7saQ3DtWnaeOXVW7qHJXtbnNwdqDvLj+RbaXbUfTNLrFduO2U25rteBPjkomJz5HCv4OIue7MSR3Y0juxpDcjRGuuUtLfwTQNA2v1xt2U0+EOsndGJK7MST3jqFqKiXOEsoaytrcxqt6WX5gOf878D9UVQUNBiQPYOKAidjMtqDt0xxprV4IECdOzndjSO7GkNyNIbkbI1xzl6JfCCGECAH13noKawpxqa42tzlYe5B52+dxqO4QoHdDHJU+iov7XozFHPwnPSsmi2R78kk7ZiGEEEKEPin6hRBCCANpmkZpfSmH6w+j0XrLQsvW/abWh4yYDMb1GkeyLzmo276CQve47sTb4k/24QshhBAixEnRL4QQQhjE5XNRWFtIvbe+zW0KawuZv31+QOv+uTnnMjpnNCbVRH1V4L5mxUxOXA4x1piTeuxCCCGE6Bqk6I8AFouFtLQ0TCYZwKkzSe7GkNyNIbkfv7L6MkqcJaiora73ql7+u/+/fH7gc38PgMyYTK7pew3ZsdkAaIqGI8nhH23YoljoEd8Du8XeOT9EhJLz3RiSuzEkd2NI7sYI19yl6I8AiqJgNsuczJ1NcjeG5G4Myf3YeXweDtYdpNZT2+Y2hTWFvLf9PUqcJUBg677F1PynW1EUf8EfZYqiR3wPrGbryf0BhJzvBpHcjSG5G0NyN0a45i5FfwTwer1UVVWRkJCAxSIveWeR3I0huRtDcj82Va4qiuqK8Gm+VtcfS+t+S6pPxVXnIjEhkbyEvIALAuLkkfPdGJK7MSR3Y0juxgjX3I/rJ/niiy9O6ElGjRp1QvuJjqFpGh6PJ+ymngh1krsxJHdjSO5H51W9FNUVUe2ubnObgpoC5m2fF9C6f17OeYzOGY3Z1EargwY2zUZuXK4U/J1IzndjSO7GkNyNIbkbI1xzP65PCKNHj/Z3IwQ9lJbft8Xna71FQwghhAh3te5aCmsL8WreVtd7VS+f7f+MLw584W/dz4rJYnzf8a227jdRUEh1pIKq/1sIIYQQojXHVfT/73//C/je5XJx33334XQ6mTZtGv369QNg27ZtvPLKK8TExPDYY4913NEKIYQQXYSqqRTXFVPuKm9zmyNb902KifNyz+Oc7ue03boPOCwOsmOyMWtmypxlHX7sQgghhAgfx1X0n3POOQHf/+Y3v8Fms7Fq1Srs9uaRgq+44gruuOMOzjnnHD7++GMuvPDCjjlaIYQQogtwepwU1hbiVt2trveqXj7d9ykrClb4W/ezY7MZ32c8WbFZbT6uCRPp0emkOFIA8Hg8HX/wQgghhAgr7ZqL4M033+SGG24IKPibREdHc8MNNzBnzpz2PIXoABaLhZSUlLAajKIrkNyNIbkbQ3LXaZpGcV0xe6v3tlnwH6g+wLPrnuWLAr07v0kxcWGPC7ntlNuOWvDHWePondjbX/CD5G4Uyd0YkrsxJHdjSO7GCNfc2/XT1NXVUVRU1Ob6oqIinE5ne55CdABFUbBaZQqnzia5G0NyN4bkDg3eBg7WHqTeV9/qeq/qZdm+ZXxZ8GVA6/41fa8hMyazzce1KBYyYzJJiEoIWie5G0NyN4bkbgzJ3RiSuzHCNfd2tfRfcMEFPP3007z//vtB6+bPn8/TTz/NBRdc0J6nEB3A5/NRUVEhAyp2MsndGJK7MSI999L6UnZX7W6z4N9fvZ9n1j7j785vVsxclHcRtw+7/agFf2JUIvmJ+a0W/CC5G0VyN4bkbgzJ3RiSuzHCNfd2tfQ///zznHfeeUyYMIGsrCx69+4NwK5duzh48CD5+fk8++yzHXKg4sSpqorL5UJVVczmtgeGEh1LcjeG5G6MSM3d4/NQWFtInbeuzfXL9i3jq8Kv/K373WK7Mb7v+KMW+zaTjayYLGJtsUd9/kjN3WiSuzEkd2NI7saQ3I0Rrrm3q+jv1q0bGzZs4KWXXmLJkiXs27cPgEGDBvH73/+eW265BYfD0SEHKoQQQoSSyoZKDjkP4dNabw3YV72P+dvnU1pfCoBZMXN+j/MZ1X0UJqX1jnYKCsn2ZNKj09vcRgghhBDieLR7hAK73c7dd9/N3Xff3RHHI4QQQoQ0r+qlqLaIak91q+ubWve/LPzSv6x7XHfG9xlPRkxGm4/rMDvIis3CYZGL5UIIIYToOOE1LKEQQghxElW7qymqLcKreVtdv696H/O2z6OsvgwAs8nMhT0u5KxuZ7XZcm/CRFp0Gin2FBRFOWnHLoQQQojI1O6if+nSpbz22mvs3r2biooKNE0LWK8oCrt27Wrv04h2sFgsJCcnh93UE6FOcjeG5G6McM/dp/oodhZT4apodb3b5/bfu9+ke1x3rul7DenR6W0+bqw1lqyYLGxm2wkdV7jnHqokd2NI7saQ3I0huRsjXHNv10/z+OOP88c//pGMjAzOOOMMhgwZ0lHHJTqQoijYbCf2gVKcOMndGJK7McI5d6fHSWFtIW7V3er6vVV7mb9j/nG17lsUCxnRGSTaE9t1bOGceyiT3I0huRtDcjeG5G6McM29XUX/008/zXnnncdHH30UlvMZhgufz0dNTQ1xcXFhNQplqJPcjSG5GyMcc9c0jWJnMWUNZa2ub611Pycuh/F9xx+1dT/BlkBmTCYWU/tbEcIx965AcjeG5G4Myd0YkrsxwjX3dg0NXFFRwTXXXHNSC/61a9dy5ZVXkpycTHR0NIMHD+aZZ54J2Obrr7/mrLPOIjo6mszMTO666y5qa2uDHsvlcvGHP/yB7OxsHA4HI0eOZNmyZSft2EOFqqo0NDSgqqrRhxJRJHdjSO7GCLfcG7wN7K7a3WbBv6dqD8+sfcZf8JtNZsb0HMOvTvlVmwW/1WQlNy6X7nHdO6Tgh/DLvauQ3I0huRtDcjeG5G6McM29XZ86zjjjDH744YeOOpYgn3zyCVdccQXDhw/ngQceIDY2ll27dlFQUODfZv369Zx//vkMGDCAJ598koKCAv7v//6PHTt2sGTJkoDHu/HGG5k3bx733HMPffr04Y033uDSSy/lf//7H2edddZJ+zmEEEJ0HaX1pZQ4S9DQgta5fW6W7l3KyoMr/cty43MZ32c8adFpbT5mij1FpuETQgghhCHaVfS/8MILjBkzhhEjRnDdddd11DEBUF1dzeTJk7nsssuYN28eJlPrH5T+/Oc/k5SUxPLly4mPjwcgLy+PW265hU8++YSLLroIgG+//Za5c+fy+OOP87vf/Q6AyZMnM3jwYO677z6+/vrrDj1+IYQQXYvb56awthCn19nq+j1Ve5i/fT7lDeWA3rp/cd7FnJl9ZpvFvN1sJysmi2hr9Ek7biGEEEKIo2lXk8MvfvELvF4vN9xwAwkJCQwaNIihQ4cGfJ1yyikn9NhvvfUWxcXF/POf/8RkMlFXVxfUzaK6upply5YxadIkf8EPejEfGxvLu+++6182b948zGYz06ZN8y+z2+1MnTqVlStXcuDAgRM6TiGEEF1fRUMFuyp3tVrwu31u/rPrP7yy8RV/wZ8bn8vdp97d5mB9CgrpjnR6JfSSgl8IIYQQhmpXS39ycjIpKSn06dOno47H79NPPyU+Pp7CwkLGjh3L9u3biYmJ4YYbbuDf//43drudTZs24fV6GTFiRMC+NpuNYcOGsW7dOv+ydevW0bdv34CLA6DfogD6bQI5OTkd/nOEArPZTGJiYlgNRtEVSO7GkNyN0VVz96geimqLqPHUtLp+d+Vu5u+YT0WDPlWfxWThoryLjtq6H22JJjs2myhz1Ek77iZdNfeuTnI3huRuDMndGJK7McI193YV/cuXL++gwwi2Y8cOvF4vV111FVOnTuXRRx9l+fLlPPvss1RWVvL2229TVFQEQFZWVtD+WVlZrFixwv99UVFRm9sBHDx48KjHU1JSwuHDhwOW7dy5EwCPx4PH4/EvN5vNmEwmVFXF5/MF7GMymTCbzSe8TtM0vF7vMa9TFAWLxUJUVBRerzfgcZvWHW2/pp/vSE2DN3q9XjQt8L5Xi8WCoignvM7n8wX16jjauqa8T3TdyXqdfD4fZrMZn8+Hz+c7ptfpRNeBvE4t1zXlDJyU3yd5nVpfZ7fbUVU16Gc06n3vx9aVO8sprivGpzU+rwIms17I17vq+WTfJ6w6tMq/b15CHuP7jifZlgw+UGnOQDErWEwWUm2pJEQlgKpfUGiZ98l4nTRNC3ifabnOiPe9SPp9aspdUZSQeN871nVd/XWKiorq0N8nk2ZCQQn6dxMVFU3RgtbZFBuKpn8f7q+TyWTCYrEEvM9A1/r71FV/n5reZ1RVDYn3vWNZFw6vk8Vi8f+NDYX3vdbWtfb4R9MxwwefBLW1tTidTm699Vb/aP3jxo3D7Xbz0ksv8Y9//IP6+npA/wNwJLvd7l8PUF9f3+Z2TeuP5oUXXuDBBx9sdV1VVRVlZc0jPMfFxRETE0N9fT01NYGtRzExMcTFxeFyuaiqqgpYFx0dTXx8PG63m8rKyqDjTExMxOPxUF5eHrAuKiqKpKQkvF5vwHGAfnIlJiZSXV1NfX19wNgIFouF1NRUgKD9TCYT6enpra4DyMzMBKC8vDzolzYjIwOAysrKoF+GtLQ0zGYzVVVVQSdrSkoKVquV6upqXC5XwLrk5GRsNhs1NTU0NDQErEtMTMRut1NXV4fTGdg1NyEhAYfDgdPppK6uLmDdyX6dSktLaWhowG63YzKZfvR1SklJQVXVoHXyOh3f69Q06mpSUhJJSUkd/vskr1Prr1N0dHMX9iP3M+J972ivU1JyEgdrD1J0qChgnWJSiE6KZlflLt7d8i6Vbv14rCYr52Wdx+g+ozEpJpwVTjQ18A9yZkYmWXFZVJVXUVYb+Jwn83Wqrq6moqLC/z4Dxr7vRcrvU9P7jN1uJzU11fD3vSbh/jqpqorD4SA+Pr7Dfp9ivbFEm/T3r+jG/1qqppoGGnDgIIYY//Jcay52mj9DhvPr5PP5KCgowGazBXyO7Cp/n7rq71PL9xmLxWL4+15L4fw6qaqK1+ulW7duQHDeofL36cgcfoyiHXkp4Tj5fD7mzJnDhx9+yL59+wDo0aMHl19+Oddff/0Jd40YPHgwW7Zs4fPPP2fUqFH+5V988QXnnHMOM2fOJDo6mgkTJvDFF19w9tlnB+z/85//nBUrVvh7AwwePJiMjAw+++yzgO22bt3KoEGDmD59Or/61a/aPJ62WvrHjh3LunXrGDRokH95KF2pUhQFTdMoLS0lMTHRf/WpaV0oXKk6cl24XFFsaGigvLyc5ORk/xXDrtKS0pVfJ6/XS3l5OSkpKf6W567QktJ07F31dVJVlYqKCpKSkoIGXg2llpR6bz2HGg7h9rnRfIGZuVQXn+z/hG+KvvHnmRefx7j8caQ4UjBZ9J9L9ak0DexvVsxkxmaS5Egy5HVqaGigrKzM/z7Tcl2ot6R05d+npveZ5ORk7Ha74e97x7OuK79OXq+XyspKUlNT/Z9vWtvveH6f/vXlv1i4fSGnZZ92XC39Ww5vYUjGEJ67/Lmwf508Hg8lJSUB7zMtMw31v09d9fep5fuM1Wo1/H3vWNd19depKff09HQsFovh73ttrduyZQvDhw9n8+bNAXVoW9rV0l9VVcXFF1/M6tWriYuLo1evXgAsW7aM+fPn8+KLL7J06dKg++iPRXZ2Nlu2bPFf9WjSdPWkoqKC/Px8AH9h31JRURHZ2dn+77OysigsLGx1u6bnO5r09HT/cx/JarX6X8SWTCZT0Iff9q5TFKXV5zraOo/H4z8JW1t/tMcEjrqu5Zt/R60zm81tXiw6GetO1utksVgwm81BuZ/Ia/hj60Bep5brWj5OR/8+tWcdhO/r1PTHzWQytfkzdub73pHrVE2lxFlCWUOZf51iaf4Av6tyF+/veN9/777NbOPinhfzk6yfBN2733QLQHJUMunR6ZhNzVkY9Tq19v5uxPteJP0+NeWuKErQuqPtdzzr5HUKXteUd0f9PqmK6p+eU1Xanpf7yHVuzY2m6PtFwuvU1vsMhP7fpyZd8XVqyr1ljuH6OaJJKLxOTccWKu97ra072uO3pl2j999///189913PPvssxw+fJi1a9eydu1aSkpKeO6551izZg3333//CT32aaedBhBUqDfde5+WlsbgwYOxWCysWbMmYBu328369esZNmyYf9mwYcPYvn071dXVAdt+8803/vVCCCHCT723nt2Vu/0Ff0sun4sPdn7Aa5te8xf8eQl53HXqXW0O1hdliqJnfE+yYrMCCn4hhBBCiFDUrqJ/wYIF3H777dx+++0BVxusViu33XYbt912G/Pnzz+hx/75z38OwGuvvRaw/NVXX8VisTB69GgSEhK44IILmDNnTsC9H7Nnz6a2tpYJEyb4l11zzTX4fD5efvll/zKXy8WMGTMYOXJk2I7cL4QQkUrTNA47D7Onag8u1RW0fmfFTp767im+LfoWAKvZyhX5V/DLIb8kxZEStL2CQpojjfzEfJmGTwgRXjQNfF7w1IOrBpzlUHsYqougqhDqK/Xlqu9HH0oIEXra1b2/rKyMfv36tbm+f//+QYMmHKvhw4dz88038/rrr+P1ejnnnHNYvnw57733Hn/605/83fH/+c9/cuaZZ3LOOecwbdo0CgoKeOKJJ7jooou45JJL/I83cuRIJkyYwJ/+9CdKSkro3bs3M2fOZO/evUEXFsKN2WwmISHhhMdXECdGcjeG5G6MUMvd5XNRWFtIvTd4kNYGbwNL9ixh9aHV/mU9E3oyvs94kh3JrT6ew+IgOyYbu8V+0o75RIRa7pFCcjeG5H6cfF5Qm748esGuesHnaVzma1zeuE0bzKpGguLDXOmEagXMUWCLBmuLrza6VYsTJ+e7McI193YV/b1792bRokXcfvvtra5ftGiR/777EzF9+nRyc3OZMWMGCxYsoEePHvz73//mnnvu8W9z6qmn8umnn/KHP/yBe++9l7i4OP8Uf0eaNWsWDzzwALNnz6aiooKhQ4eyePHigIECw5HJZMLhcBh9GBFHcjeG5G6MUMq9rL6MEmdJwJR6TXZU7OD9He9T5dJHvbWarYzJG8MZWWe02pXfrJhJc6S12vIfCkIp90giuRsj4nNvKtr9hbuvRUHfWMT7Whbx7Rqr289kUnDYWpQMPhfUu6C+onGBAhb7ERcCHKAorT6eODYRf74bJFxzb1fRf/vtt3PnnXdy6aWXcs8999C3b18AfvjhB5555hmWLVvGc889d8KPb7Va+dvf/sbf/va3o2531lln8dVXX/3o49ntdh5//HEef/zxEz6mrsjn8+F0OomOjg67q1ahTHI3huRujFDI3aN6OFh7kFpPbdC61lr3eyX0YlzfcSTbW2/dj7PGkRWThdV8fIPldKZQyD0SSe7GCLvcVfVHWuKP+NLaHmjwZPKpGk6Xl+goC2ZTa4W8Bt56/YumsVMUvfgPuBAQWj2lQl3Yne9dRLjm3u6iv6SkhH/9618sXbo0YJ3VauWvf/0rt912W7sOULSfqqrU1dVht9vD6uQNdZK7MSR3Yxide5WriqK6Inxa8P2m2yu2s2DHAn/rvs1sY0zPMZyeeXqrrfsWxUJmTCYJUQkn/bjby+jcI5XkboyQz13Tjl60H9k639lFvNp4z77X1fj/Br1Q9zTo/25a5v+/vlzx1GPCDsndISEb4jIhLksv6NukgadO/2qimJt7AdiiwRoDFttJ/7G7qpA/38NUuOberqIf4O9//zt33nknn376Kfv27QOgR48eXHDBBaSmprb7AIUQQoi2eFUvRXVFVLurg9Y1eBv4cPeHfFf8nX9ZfmI+V/e5us3W/cSoRDKjM2VUfiGETlMbi3Ot+d+aBj43uKqgdEdzQd/KRcf2P3/jcx1DcX7UbbwN+jGeIJtPRan4PnChPaHxAkDjRYC4LP3f9sTWu/ZrPnDX6F9N1wJMluaeAE29AkK4d5UQXVW7i36A1NRUJk6c2BEPJYQQQhyTWncthbWFeLXgAah+KP+BBTsW+C8G2Mw2Lu15Kadnnu6f57slm8lGdmw2MdaYk37cQggDHXnfu8ept7q761ov8Nt8HA943eAOvp3I/zzeIwvwIwrzNgv4+ub/e12d2yNAabw/32LXW+QtdjSTDW9tOWZXWeCFjYYq/evwD4GPYXVAbOPFgPgs/d/xWRCdCkdeUFW94KrWv5qYbfpjWGOaLwTIhVgh2qVdRf+nn37Kf//7Xx555JFW199///2cf/75nHfeee15GiGEEMJP1VSK64opdwXPDlPvreej3R8FtO73TuzN1X2uJsmeFLS9gkKKPYW06LRWu/oLIUKcqgaOQP9jI9YfObhdQ5W+zlsPmoZF07BpPqLU5q+m720tlg2pq6KXbwesfKH1Qt7n7twczFawRIHFEVCwY7Xry5r+b4lqsa6Vbcw2OOK9UPWpVNU0kBJjxdRQBjVFUHOo8avx396G5h089VCxR/9qyWSB2PTmHgFNvQNi0/XjauJz618NVc3Lmo636UKAxSEzBghxHNpV9D/00EPk5ua2ub6wsJCHH35Yin6Dmc1m4uLiwuq+lK5AcjeG5G6Mzsrd6XFSWFuIWw3+QH28rfsOs4Os2Cwclq47Sq+c78aQ3DtZY4u62VVLnK8Oc3kl4Gse3E5Tj+jmfpSW8yO6v48r2czoujIynN8RpfowHeOI906vi2ivCkXr2/ez+YvuI4rzpmVNhXlrBbx/P/tJ7RJvVhTi7Db9fG/qzt+SpkFDpX4BoLoo8GJAyxZ81QvVB/WvlhQFHCnNFwLiW1wQsDX2vmp6XVvOGGB1NF8IaPp3GM0YIO8zxgjX3NtV9G/atIkJEya0uf70009n8eLF7XkK0QFMJhMxMdJltbNJ7saQ3I1xsnPXNI3D9YcprS9FO+JDeb23ng93f8ja4rX+Zb0TezOuzzgS7YnBx4qJtOg0UuwprV4M6ErkfDeG5H4SaZre5d7t1LvON1RDxW4o342pYi8xDVUtur63KPRPUIqrFrOq4jjKPPVHUlGoUxQ8lihI6H5Eq3lbRXpr20QFtaqHIpNJIcZ+lJJBUcCRpH+lDwxc565rcRGg8UJAdRHUl+mvNej/d5bqX8WbA/ePijuiZ0Dj/x1J+nniceKfMUAx6XmHyYwB8j5jjHDNvV1Fv8vlwu1uu/uSy+XC6XS25ylEB1BVlfr6ehwOBybpCtVpJHdjSO7GOJm5u3wuCmsKqffVB63bVr6NBTsWUOOuASDKHMVlvS7jtIzTWi3oY62xZMVkYTOHx4jRcr4bQ3LvQD6vXtx7nHqBWF0IZTuhfDeU74HKff4B6DRA1cCkwAldrmu6V7xF4b23fDubq92kxGXgUsy4TWZcpub/u5QjvjeZ8aGwsWInp6b05/mz/tSRaXQsxQQoja3fjf8PWmZqbh0PWqZvp6oa9Q0NOCxg0nx61/tjHbTQFgMp+fpXSz431BRDbeNFgKYLArXFjbdhNHLV6F+l2wP3t9ibex00jRkQlwkxaa3PGGCLbu4V0EVmDJD3GWOEa+7tKvoHDx7MggUL+M1vfhO0TtM03n//fQYOHNjKnqIz+Xw+ampqsNlsYXXyhjrJ3RiSuzFOVu6l9aUcdh5GJXAgK6fHyYe7P2RdyTr/sr5JfRnbZyyJUYlBj2NRLGREZ7Ta8t+VyfluDMm9HTz1LVrxK/VirkxvyadiT4vu28E0WxwuWzJR0bGYbS1b1I+4N91//3eLbvOWqFYHg1v23Uss2lfNiJS84/ghWhTHJmtQkdxqMX1MBfiRy/iRxz7a43RcLyafx0ONqwxbYgoma+NtBGpj8e/zNH659TEUmv7tcx99AEKzDRJz9K+WVJ/e4t/yFoGm/3taXPj1NkDFXv2rJcV8xLgBLXoINI0bYLI2ThsY03whwNwhY5t3KHmfMUa45t6uM/zXv/41kydPZsKECfz1r39lwIABAGzdupV//OMfrFy5ktdff71DDlQIIUTk8Pg8FNYWUuetC1r3fdn3fLDzg2Nu3U+wJZAZk4nFFHof6oQIa6qqt7q668BVqxdoTa34FXug8kDbLcaKWS8Ik3tBck9I7oUvJpvqWg8pCTGYLebjKKaPLJxpXuZI1i8M2BNo7j/Q1JWgZQs4zesAagohOgUyB3dgYF2IyQymxvvo2+K/MODWe3T4/+1pvEDQyoUBkxliM/SvrFOal2uaPrBfwIWAxn+3HPBP8zWvO1J0cuDUgk0XBKLiGnuBRLe4ECAzBojw0q5PQJMmTWLXrl089NBDvP/++/6rIaqqoigKf/nLX5gyZUqHHKgQQojIUNlQySHnIXxHFANOj5PFuxezvmS9f1nfpL5c3edqEqISgh7HarKSHZNNrC32ZB+yEAL06eXcjUW+sxxKtjYX+OW79W7abXEktSjw8/V/OxKbB2mzxYBPBVMZJKeAtYMGrmu6r94kc8N3uGO5MODzNl8A8Hla7z2gqY0XaBL1r/QBgY/hdjbfJtDydgFnWeBFBWe5/lW8JXB/W2yLQQSzWkw3mK2v898eEB1WAwWKyNLuZo+//e1vTJo0iQULFrB7924A8vPzGTt2LPn5+T+ytxBCCKHzql6Kaouo9lQHrfu+7HsW7FxAbeOc2FHmKC7Pv5xT009ttXU/xZ5CenS6TMMnxMniH3CvTi/my3bpXfXLG7vq1xxsHqjtSGYrJPZoLPJ76fd7x3f/8fuufZ04X73oHGaL/vVjFwaaegmoR/QY8HnApjSfSwH7efQxAo68VaC22D9OBKDfalK2U/8KODbbEbcIZENSnv5lT2icOtAuFwJEl9AhfR3z8/P53e9+1xEPJU6CplEow+m+lK5AcjeG5G6M9uZe467hYO1BvFrgCNpOj5P/7P4PG0o2+JcdrXXfbraTFZNFtDX6hI6jq5Hz3RgRmXvTgHvuOqgr0VtLy3Y1t+R7ggfa9ItJay7KkntBSh+wxwcOsHYMIjL3EGB47k0XBjjK+3rL3gEtewk09SDxeaBp5hdNhbojxw1o/LenxQDkPjdU7te/WlJM+jnd1DMgqWfjed0bYlKbxw5oJ8Nzj1DhmnuHFP2rVq3if//7HyUlJdx+++306dMHp9PJtm3b6Nu3L7Gx0rXSSE3zTYrOJbkbQ3I3xonm7lN9FDuLqXAFD961pXQLH+z8gLrGkZjtFjuX97qc4enDg1r3w2kavuMh57sxIiJ3T71e4DdU6y34h7c1t+LXFre9n8Wut4Q2Ffhp/fSWUv8UatFwgh+mIyL3ENQlcjdb9a8fvTDQ2EsgITd4jAGvG1xVgVML1jZeEGg5wKSm6r8DtcVQtD7wORxJ+sWApB76xYCU3vrvQEzacfcI6BK5h6Fwzb1dRb/b7WbixIksXLgQTdNQFIUrrriCPn36YDKZuOiii7j33nu5//77O+p4xQlQVRWXy0VUVFTYXbUKZZK7MSR3Y5xI7k6Pk8LaQtyqO2j5ol2L2Hh4o39Zv+R+XN37auKj4oMeJ8YSQ3ZsdthMw3c85Hw3Rtjlrvr0At/jhKpCfa50/4B7e/XCqC3x2c0tnal99CLHFtvYgt+xo6KHXe7HYfTo0axatQqLRc/z7LPPZsmSJbhcLu68804++OADNE3jkksu4cUXXyQuLo5Vq1Zx//33s3btWmw2G5deeilPP/008fHB76MrVqxgzJgx/u81TcPpdLJmzRqGDx8eHrn7Lwy0MQe7pgUOMtiy90BDtT59ZNWBwB4CdYcDxw2or9C/SrYGPrYttnG2gh7NvV1SeunjB7SRaSSf70YK19zb9U78wAMPsHjxYl588UXOPfdc+vXr519nt9uZMGECCxculKLfYD6fj6qqKlJSUsLq5A11krsxJHdjHE/umqZR4iyhtKE0aF1rrftX9LqCYenDglrwzYqZjOgMkuxJHfeDdDFyvhujy+feNOBefSWUfA+Hv29sxd+jT5fWFltMi67M+ZA+UO/O3DTqeQd1a25Ll8+9nV599VUmTZoUsOy5555j7dq1bNu2DZvNxvjx43nkkUd49NFHqays5I477uCiiy5C0zRuuukmfve73/Hyyy8HPfbZZ59NbW2t//t33nmHP/7xj5x66ql4vd7IyF1RGseSsNHqhYHsU5ovDDTdPuCua7wFYJ8+G0XlAagubBw3oMXFMnet/rtW8n2L5zPpvzMJOXrvmJTeeg+BxB6QkINPUyIj9xATru8z7Sr63377bW677TamTZtGWVlZ0PoBAwbw3nvvtecphBBChJEGbwOFtYU0+BoCltd56li0axGbDm/yL+uf3J+xfcYSbwtulYq3xZMZk4lVRtwW4ug0rXlE/cp9cGgzlO3Qi/zK/frAaK1RTBDfrbHA7wVpA/XCJKpxNHOrQwYwCwF79+7lkksuISUlBYCxY8eydOlSAC655JKAbadNm8Zvf/vbY3rc2bNnM2nSpIi6XeqYNF0YaBpo0pEECd2hx5nN22iafmGtqqC5t0xV49gAVQX6BQBN03sIeOr122dKt8OOZc1TTposmOOyiLOlYkrtAQlZjVMZpuv/j0lr7LUgxLFpV9FfUlLCkCFD2lxvNptxOp1trhdCCBE5SutLKXGWoBE4ovfm0s0s3LnQ37rvsDi4PP9yhqUFt+5bFAtZsVmtXggQQqC3QjZNmVe8uXnavPLdgfOZHykqXi/uk3tBaj/9PuTo1OapymTOcsPde++93HvvvQwbNownnniCoUOHMmXKFH77299SUlJCVFQU77//PldffXWr+3/xxRcMGjToR5+npKSEpUuX8uSTT3b0jxAZFAWsdkjtrX+1pGl69//KffpAmBV79IsClfv1wQVpvBjgc0HFHmy+XSjFaxovBihA40UBxQSOZIjLaJxiMCPwokBsOtgT5cKc8GtX0Z+Tk8O2bdvaXP/VV1/Ru3fvNtcLIYToWG3d97lw4UL+8Ic/UFRURHR0NNdeey2PP/44ZrP+Qb6mpoZ7772X999/H1VVufLKK5k1a1arz3HZZZexevVqXC4X/fv356mnnmLEiBFtHpPb56awthCnN/AicK27lv/s/k9A6/6AlAFc1fuqVov65Khk0qPTMUvxIYRO05oH3CvbqY+oX/qDXuBXFQTea9ySyQyJuY2D7fWG9EGQlKt30bfFSAtiCHrssccYOHAgZrOZZ599ljFjxrBt2zZ69+5NWloamZmZAFx00UXceuutQfuvXLmSF154gS+//PJHn2vu3Lmcdtpp9O3bt8N/joinKBCdrH9lDw9c53Y2zxZQuQ+tbA/e0r2YPZX6AINojT0EfHoPnZoifWrMpsfF1OLCgKLfOhCboQ+kGZvZ4oJAi4sDrU2NKcJSu4r+6667jieffJLx48f73xiaWmVeeeUV3n33Xf71r3+1/yhFu5hMJqKjo8PqvpSuQHI3huTe+n2fI0aMYMWKFaSlpVFRUcE111zD9OnTueOOOwC4+eabyc3NZe/evTgcDjZv3tzm4z/22GP069cPi8XCokWLuOqqqzh48GCruVc0VHCo7hAqgcXHptJNLNq5KKB1/8r8KxmaNjSodT/KFEV2bHbETMN3POR8N4ZhuTcNuFdborfiF2+B8l36vfju2rb3i05uHk0/fSCkDQBHYmM3fftxHUJbFxZvvfVW5syZ49/O7XbTr18/Nm3SL+q1drHwpz/9aavPceSsT06nk8cff5x77rknYs/3M844w//v++67j9dff51Vq1YxY8YMrFYrVVVVqKrKLbfcwu9///uAVvotW7Zw9dVXM3v2bAYOHPijzzV79mxuuukm//fyPtNJbNGQ3l//AvD58NTVYYuJ0ccPqCvRf/ebZg6oLWkeULC2WO/lo2mABqqqvyeU1egXBJsuBLS8KICpcbaBzMYLAy0vCDReFLAnnvBsG11VuJ7v7Sr677//flatWsWoUaMYMGAAiqJw7733Ul5eTkFBAZdeein33ntvRx2rOEFms7nVkVrFySW5G0Nyb123bt0CvjeZTOzatQuArVu38t133zF37lx/y//w4cODHqNJU/dQTdMwm80cPnyYuro6EhIS/Nt4VA9FtUXUeGoC9q1117Jo1yI2lzZfVBiQMoCxvccSZwucIkdBIdWRSpojTe4rbYOc78botNy9LnDVwOEf9HvxSxunzaspavxw39rB2RqnzOupd9PPGKQPFNbUTb8Dfpdau7A4ffp0pk+f7v9+zJgxjBw50v99axcLi4uLW/3dbjmg3MGDB8nNzWXcuHFyvrdgMpnQNI0NGzbw5JNP+qcYmzx5csAA2rt27eLiiy/mySef5LLLLvvRx922bRsbN27kF7/4hX+Z5G6MgNzNjb1zEnNb31hVoaGyxUWBFhcHaoqgprhx2sHG2wc0Ve8xUFusT0tYROu9BcxW/QJAXFYbvQXS9fE9wki4nu/tKvptNhsff/wxb775JvPmzcPn8+FyuRg6dCgPP/wwN9xwg3xQCwGqquJ2u7HZbGF31SqUSe7GkNxbv+8T4Msvv+Syyy6jurqatLQ0/v3vfwOwevVq+vTpw5QpU1iyZAl9+vThySef5Mwzz2zzOS6//HKWLVuG2+3m1ltvJS4ujoaGBmw2G7XeWopqi/BqgQOEbTy8kUW7FuH06N38HRYHV/a+kqGpwa370ZZosmOziTKf3NHAuzo5341xUnJXVX3KvOoiKN6ot+KX7dTv9/XUt71fbIZe4Kc0dtNP76/fn2+N7tDp8o7HoUOH+PTTT3nuuef8y1q7WFhdXR1wsbA1b731Fj/96U/p2bNnxJ7vlZWVrF69mlGjRqEoCs8//zzl5eWMHDmSESNGMHv2bM4++2w0TWPOnDn+8bYKCgq44IIL+POf/8x11113TM81e/ZsLr30Uv/AgCDvM0Y5rtxNpubbBpp6ChzJ627sLVAceHGg5lBjb4FD+oXGptsIVFW/jaB8r96bqK3eAvb4xtsIsloZWyBDH3ugC5034Xq+t/uvgaIoTJo0KeiKrwgdPp+PysrKsJt6ItRJ7saI9Nzbuu8zLi6Os846i6qqKvbs2cPs2bNJT08HoLCwkE8++YRXX32VGTNmMH/+fK688kp27txJYmJiq8+zePFi3G43ixcvpq6uDp/PR3l5Od5oLzW+4Nb9hbsWsqV0i3/ZoJRBXNX7KmJtgd14zYqZ9Oh0ku3JHRtMmIr0890oHZK7z6O3zJV8D0Wbmlvxa0va3sdi1wv85HzIGAjpgyE+E6wxnXpvblsXFpu8/fbbnHHGGeTn5wcsP/Ji4Y8V/KAXoXfeeScQuee7x+PhT3/6Ez/88ANWq5Vhw4bx0UcfkZCQwOOPP87tt99OTk4OAKNGjeKpp54C4LXXXmPv3r3cd9993HfffQD06NGDLVv09+IxY8Zw9tln8+c//xnQL8i8+eabQQP4RWruRuvw3C02faaBhO6tr9c0fbDPI28hqC1uvoXAWdbYS0Br7i3gLIW6w/otR631FjBZITatccDBNsYWsIXO7Xvher536CVgt9vNN998Q1FREf369eOUU07pyIcXQgjxI9q67/PCCy/0L+/ZsyeDBg3izjvv5N1338XhcJCXl8fUqVMBmDhxIv/85z/55ptvuPjii9t8LpvNxrhx4xg6dCj9B/XHFG/CZrZhsuh/JDVNY2PpRhbtXES9V2+pjLZGc2X+lQxJHRLUuh9vbZyGTwYRE+GmacC9yv1waAMUb9WnzavYFziXd0uKAnHZ+n34af0gczCk9NWnzLPYDRuV+2gXFpvMnj2badOmBe175MXCH7Nx40a2b9/OhAkTOvRn6GrS0tJYs2ZNq+vS09OZN29eq+v+9re/8be//a3Nx12yZEnA94qisHfv3hM+TtHFKYo+1ocjEdLaGMTR59FnGTjyFoKmCwM1RY09k47oLVCxT++11FZvgajY5osCrfUWiE6RGUTa6biL/qVLl/LOO+/w2GOPkZqa6l++bds2rrzySv89ogBXX301c+fO9Q/2IoQQonM13fd5JK/Xy86dOwEYPHhwUAF+PLdmud1uvtn8DWf8pPmCQ627loU7F7KlrEXrfuogrsoPbt23KBYyYzJJiPrxVj8hugTVB/XlcGgLHNoEh7/XB9xzlre9jy228T78vvpge5lD9Lm4rdEh1TX2xy4sbtmyha1btwbcE95Sy4uFp5566lGnkJs9ezZXXHFFmz2OhBCdzGyF+Cz9qzWapo9D0mZvgUN6bwHV19hbQAPNq79fOsugZEsbvQXM+vthbMtBB9MCewzYYmWKwqM47mr89ddfZ+fOnQEFP8D111/Pzp07mTJlCiNGjOCjjz5iwYIFPPvsszKYnxBCdIKj3ff57rvv8pOf/ITc3Fx27NjBo48+ygUXXADoo3FrmsbMmTOZNGkSCxYsoKioKGAQriZ79uxh69atnH/++SiKwlPPPcWBggMMGt58v+6Gkg0s2hXYun9V/lUMSRsS9HhJUUlkRGfINHyia3PX6/NtF21onDZvO1Tt1z/YtkYx6V1sU3o3F/jJvbrkdHlHXlicPXs2l112GUlJSUfdz+PxsGvXrjaLflVVeeuttwIGBxRChDhF0e/xt8dDahvTtqu+o/cWqD0ErjqgxW0EXi9UHtB7S7XVW8AW3WKKwozgWwhi0gwb5yQUHPdPvmbNGq655pqAZevWrWPdunVcf/31zJgxA4A77riDc845hzfffFOKfoOZTCbsdntY3ZfSFUjuxojk3I923+cPP/zAb37zGyoqKkhJSWHChAk8/PDDAFitVhYuXMjUqVO544476NevHx988IG/da1pzufp06ejaRoPP/ww1157LSazifz++Tz/1vOkZqZSUVXB+9vf5/vy7/3HNCR1CFfkXxHUum8z2ciOzSbGGtM54YSpSD7fjWSqLyemeA3m7dv1e/HLdoGruu0d7An6ffhp/fVu+hmD9amyjnO6PKMd7cIi6IX6m2++GTCAHwRfLJw+fToFBQUBvQaO9Nlnn+HxeBgzZox/mZzvxpDcjRG2uZvMehf+uIy2t3HVHjFF4RFjC9SW6LcN0KK3QEMV1FfC4W2t9xZQTBCTekRvgZbjC6RBVHzY5n7cRf+hQ4fo3Tvwys3HH3+MoijceOONAcvHjh3LX//613YdoGg/s9ksXeMMILkbI5JzP9p9nw888AAPPPBAm/sOHTqU1atXt7quZUtbr169WLlyJaX1pRQ7i/3LNxzeEHDvfow1hit76/fut6SgkGJPIS06DZMSXn9QjRDJ53un0jR9oKpd/4XdyzEf2kicz9P6tiYLJPaA1D76aPpZp0Bijt6K38W7nh7twiLA8uXLqa+v59JLLw3Yr+XFQrPZzODBg1m8eDGZmZkAPPLII6xYsSLgHvPZs2czceLEgFtE5Xw3huRujIjOPSpW/0ru1fp6VdVvBzjyFoKWvQUaagiYolD1QlUhVBXQZm8Bqx1zbCaJcZlwyaNA+PRCPO6iPzY2FqfTGbDsyy+/xGQyBXUFTUxMxOdro2ub6DSapuHxeLBarTKFYieS3I0huZ98xXXFlDaUAnren+3/jM/2fwYaoMDQtKFc0Su4dd9hcZAdk43d0rVaN0OZnO8nkaqC8zDsWg67l+sD8DUW+hr6dQBFASU6FVLyIX0AZJ6iT5flSArLQaeOdmER4LzzzqO0tDRoedPFwrY0jR7f0qxZs4KWyfluDMndGJL7UZhMjff0pwGDW9/G7Tyit8ARUxTWlTS+p7foLeCqQXPXojnLUEwWwin14y76BwwYwMKFC7n77rsBqKio4IsvvuDMM88kNjbwA96BAwf8V3GFcbxeL+Xl5aSkpGC1dq17Bbsyyd0YkvvJo2kaRXVFVLgqAFA1lcW7FrOqaBVoEEUU4/uOZ0hGYOu+CRNp0Wmk2FPkg0sHk/O9g6k+/UPhnhWw53MoWh88ur7VgZp9OlUpw4jrPRJrcl6nTpcXyeR8N0Yk5z569GhWrVrl73Fy9tlns2TJEt544w1++ctfYrc3X8TeunUrubm5ANTU1HDvvffy/vvvo6oqV155ZasXsgAuu+wyVq9ejcvlon///jz11FP89Kc/jejcO4QtGmx5kJTX+npV1adNPaK3gFpdRL3LS5TXG1a5H3fR/9vf/parrrqKMWPGcOaZZ/Kf//wHp9PJ7bffHrTtxx9/zPDhwzvkQIUQQhhH1VQKawqp9uj3LXtVL+9tf49NhzcBkBiVyLU519I9JXD+31hrLNkx2TINnwhdPq/edX/vCr3YL1oP3obAbSx26HYa5J8HvUajmh24y8ogKQUscm4LEc5effVVJk2aFLR89OjRfPrpp63uc/PNN5Obm8vevXtxOBxs3ry5zcd/7LHH6NevHxaLhUWLFnHVVVdRXFzc5vaig5hMEJ2sf6UP8C9WPR5qy8qIMvDQTobjLvqvuOIKHnvsMR5++GGWLl2Kw+HggQceCJqaZdWqVaxatYpXX321ww5WCCFE5/OpPg7UHKDOq8+r7fK5eHPrm+ys1Kf8S49OZ0r/Kdgamls7ZRo+EdK8br3Q3/91c6HvqQ/cxhIFWcMg/1zoda7+wbCJp437+YUQEW/r1q189913zJ07F7NZv83naI2gTTNYaJqG2Wzm8OHDVFdXEx0d3SnHKyLDCc1b8Lvf/Y57772X0tJS0tPTW+2uecopp3D48OHIHYBCCCHCgFf1sr96P/U+vSCq89TxxuY3KKwtBCA3PpfJAydjV+zUN+jbJNgSyIzJxGKK3KlxRAjy1OsDP+1fBXu/1At9d13gNmabPvBez1F6oR+b3uUH3xNCtM+9997Lvffey7Bhw3jiiScYOnQooDdwpqSkkJGRwV133eWf6Wb16tX06dOHKVOmsGTJEvr06cOTTz7JmWee2eZzXH755Sxbtgy3282tt95KQkICHrm4KDrQCX8iM5vNZGS0PdWCw+HA4XCc6MOLDmQymYiKigq7qSdCneRuDMm943h8HvZW78Wt6vc0VzZU8vrm1ymt1wfq6pvUl+sGXIfNbEP1qURFRZEbn0uCQ1r3O4uc7z/CXQfOcr3QP7AKCteCuzZwG7MVMoZA3tl6oR+f+aOD8EnuxpDcjRHJuT/22GMMHDgQs9nMs88+y5gxY9i2bRvnnHMOmzdvJjc3l9WrV3P11VeTlpbG+PHjKSws5JNPPuHVV19lxowZzJ8/nyuvvJKdO3e22Ri6ePFi3G43ixcvpq5OvxgZybkbKVxzl2aYCGA2m0lKSjL6MCKO5G4Myb1jNHgb2Fe9D6/mBfQR+2dsnkG1W7+nf1j6MMb3GY+5sThKtCfSLa2bTMPXyeR8P4KmgasG6iugYA0cWAmF68BVHbidyQIZg6DHzxoL/ezjGoxPcjeG5G6MSM79jDPO8P/7vvvu4/XXX2fVqlVceOGF/uUjR47krrvu4v3332f8+PE4HA7y8vKYOnUqABMnTuSf//wn33zzDRdffHGbz2Wz2Rg3bhxDhw7l1FNPZdCgQRGbu5HC9XyXoj8CaJqG1+vFYrHIyNmdSHI3huTefk6Pk/01+/Fp+pSr+6v3M3PLTOq9evf9n3X7GWN6jvEX+ElRSWRGZ+Lz+lAsiuTeieR8Rx+B2VWtF/oH18GBb6DwO2ioCtzOZIb0gZDzE+g1GhJzwHpiPRIld2NI7saQ3JuZTCY0TTvq8sGDBwfldDy5eTwedu3axcCBAyV3A4Tr+S5NMhHA6/VSVlaG1+s1+lAiiuRuDMm9fWrcNeyr3ucv+H8o/4HXNr3mL/gvyruIS3te6i/4U+2pZMdm4/P5JHcDROz5rvr0bvtlu2DrB/D5/4MFv4Llj8Ku/zYX/IpJb9E/7SYY9xpc8TSMnAZpfU+44IcIzt1gkrsxIjX3yspKli1bhsvlwu128+9//5vy8nJGjhzJxx9/zOHDhwFYu3YtzzzzDFdeeSWgj+qvaRozZ87E5/Mxb948ioqKGDlyZNBz7Nmzhw8//JCGhgZcLhdPP/00BQUFnHHGGRGbu9HCNXdp6RdCCAFAlauKwtpCNPTWivUl63lv+3tomoaCwtg+Yzk983T/9pnRmaQ4Uow6XBFpfB69mK+vhOLNULAaCtfoxX9LignS+kG3EdDrHEjMhagEfXomIYQ4Rh6Phz/96U/88MMPWK1Whg0bxkcffURCQgLLli1j8uTJ1NXV0a1bN/7whz8wceJEAKxWKwsXLmTq1Knccccd9OvXjw8++MB/P3/TgH/Tp09H0zQefvhhrr32WsxmM4MHD2bx4sVkZmbKQH6iQx130T9v3jwuuugi4uPjT8bxCCGEMEBZfRmHnIf8339d+DWLdy8GwGwyM7HfRAal6tMKKSh0i+0m0/GJk8/rai70D3+v36df+J0+3V5LigIpfaH7CMg7C5LywJ4IZmnbEEKcmLS0NNasWdPquieeeIInnniizX2HDh3K6tWrW103ffp0/7979erFypUr23egQhyD4/5reO2112IymfjZz37GZZddxqWXXsqAAQNOxrEJIYToBCXOEg7X60WUpmks27eM5QeWAxBljmLyoMn0TOgJgAkTOXE5xNpiDTpaEfY89Xqh76yA8p3NhX5tcfC2Kb2h++nQ40y90HckgSWq0w9ZCCGECGXH3dft8OHDzJw5k5ycHB577DEGDx5Mr169+PWvf83SpUtxuVwn4zhFOyiKgtVqDavBKLoCyd0YkvvxKaot8hf8qqbywc4P/AV/jDWGW4be4i/4LYqFHvE9Wi34JXdjhE3urlqoKoTirfo9+d++DB/fB5/9A374KLDgT86HoT9n9Bwv9psWEDvmr8QOvIAxP78JLFEsXLiQ/v37k5CQQFZWFr/5zW/w+fQxKlasWEFsbKz/KyYmBkVR+O677456eCtXrsRkMvHwww8DYZR7FyO5G0NyN0Yk5z569Gjsdrv/vXrMmDEAR31/B6ipqeGXv/wlycnJJCYmMnny5DafY8WKFZx22mnExcVxxhlnsGnTJiB8cz/ulv7ExEQmTpzIxIkT0TSNb7/9lo8++oiPPvqIF154AYfDwbnnnsvll1/OpZdeSk5Ozsk4bnEcLBYLKSly321nk9yNIbkfG03TKKgt8E/B51W9vLPtHbaUbQEgyZ7ETYNvItWRCoDVZKVHfA+izK23okruxuiyuTdNrddQpX9VF+j36Bd8B9WFwdsn5eld97uN0Fv3HUlg/5BXX32VSZMmBWw6YsQIVqxYQVpaGhUVFVxzzTVMnz6dO+64g7PPPpva2lr/tu+88w5//OMfOfXUU9s8VFVVuffeezn99ObxLLps7l2c5G4Myd0YkZ778b6/A9x8883k5uayd+9eHA4HmzdvbvWxy8vLGTt2LK+++ipXXnklc+bM4aqrrvKP3xCOubfrZjdFURg5ciQjR47kwQcfpLi42H8B4A9/+AO33347AwcO5PLLL+fGG2+kX79+HXXc4jhomoaqqphMprC7ahXKJHdjSO4/TtVU9lfvp85bB0CDt4E5389hd+VuADJjMrlx8I3E2/SxW6JMUfSI74HVbG3zMSV3Y3Sp3Jum1muohIZqqCnSu+0XrIGqA8HbJ+Y2FvqnQ3LPxkI/8UcH5OvWrVvA9yaTiV27drW67ezZs5k0adJRs3v55ZcZOXIkVVXNUwB2qdzDiORuDMndGJJ7sKO9v2/dupXvvvuOuXPnYjabARg+fHirj/P111/To0cPrr76agCmTJnCww8/zOeff875558flrl36FC2GRkZ3HTTTbz33nuUlpby6aefcskll7Bo0SLefffdjnwqcRy8Xi+HDx8Ou6knQp3kbgzJ/ei8qpe9VXv9BX+tu5ZXN73qL/jz4vO4Zegt/oLfYXGQl5B31IIfJHejhHzuTVPrle+G4k16a/6GuXq3/U/+AlsWBBb8Cd1h4Fi46CG4+FEY8UvoNRpS8iE6Oajgv/fee0lLS+PCCy9k48aN/uVffvklCQkJJCcns2HDBm6++eagQyspKWHp0qXccMMNbR5+WVkZTz31FA8++GDA8pDPPUxJ7saQ3I0R6bkf7/v76tWr6dOnD1OmTCElJYWf/OQnfP31120+vqZpQd9v2bIlbHM/acPaWiwWzj33XM4991wef/xxmXZCCCEM5vF52Fe9D5eqj71S3lDOjE0zKGsoA6B/cn+u7X+tv8CPtcaSE5eDSZGpzsRxaJpar6FK78LvLNNb9AvXQPme4O3js/Vu+91Pg8Q8vUU/OvlHB+R77LHHGDhwIGazmWeffZYxY8awbds24uLiOOuss6iqqmLPnj3Mnj2b9PT0oP3nzp3LaaedRt++fdt8jvvvv5977rnHP9WWEEKIk+9E3t8LCwv55JNPePXVV5kxYwbz58/nyiuvZOfOnUHv4T/96U/Zu3cv7733HmPHjmXWrFns3r2buro6A37aztFpn+Ss1qO3EgkhhDh5XD4Xe6r3+Av+Q3WHeGnDS/6C/9SMU5k0cJK/4E+wJZAblysFvzg2XhfUFMPh7VC8GQ5t0lvxl/8LPv4jbHovsOCPzYD+l8EFf4eLHoYRN0HPcyBjIMRnHdMI/GeccQaxsbE4HA7uu+8+4uLiWLVqVcA2PXv2ZNCgQdx5551B+8+ePfuogzytW7eO1atXc8sttxxzDEIIIdrvRN7fHQ4HeXl5TJ06FavVysSJE8nKyuKbb74JevyUlBQ++OADHnvsMTIzM/n000+54IIL6N69e6f8fEaQCWyFECLMOT1ODtQcwKvpXdX2Ve9j5paZNHgbADi7+9lckneJ/961FHsKmTGZhh2v6CLczuYWfW+9fp9+4Xf6V9kOfbC+lmLS9On1up2m36/vSNRb9aPioQPumzSZTEHdNUHvIrtz586AZdu2bWPjxo384he/aPPxPv/8c3744Qf/PaRVVVVYLBZ27drFyy+/3O7jFUIIcWyO5f198ODBQffgH+2e/HPOOYfVq1f7H6d3794BA7aGGyn6I4CiKFgslrAajKIrkNyNIbkHqnXXcqDmACoqAN+Xfc9b297Cp+pT3FzS8xJGdR/l3z4jOsM/Yv/xkNyN0em5u2qbC32fS+++f3Cdfq9+6Q7Q1MDto1P1bvvdRkBiD7DHH/OAfEdTWVnJ6tWrGTVqFIqi8Pzzz1NeXs7IkSN59913+clPfkJubi47duzg0Ucf5YILLgjYf/bs2Vx66aVHHaF52rRpTJw40f/93XffTc+ePfnjH/8o57tBJHdjSO7GiNTcT/T9ffTo0WiaxsyZM5k0aRILFiygqKiIkSNHtvo869atY8iQIdTV1fHXv/6VkSNHMmDAALxeb1jmLkV/BLBYLKSmHv+HeNE+krsxJPdmVa4qCmsL0dCvjq8tXsv87fPR0FBQGNd3HKdlnAaAgkJWTBZJ9qQTei7J3RgnPXf/1HqVeku+6gF3HRxcq0+vd3hbcKHvSGq8R3+EPtWeLQYcyXrL/o8MCHmsPB4Pf/rTn/zTKw0bNoyPPvqIhIQEfvjhB37zm99QUVFBSkoKEyZM4OGHH27xI2m8+eabPPnkk0GP+8gjj7BixQqWLFlCdHQ00dHRzT+Ww0FsbKz/3lA53zufvM8YQ3I3RqTmfqLv71arlYULFzJ16lTuuOMO+vXrxwcffOB/z7711lsBmD59OqC/3y9duhSz2cy4ceN49dVXgfDNXdFa6yshjsmWLVsYPHgwmzdvZtCgQUYfTptavsThdtUqlEnuxpDcdeUN5RTVFfm/X1GwgiV7lgBgNpm5rv91DEgZAIAJE93iuvlH7D8RkrsxTkruqgquxtb8hmrQfHpX/qL1+vR6Jd/ry1pyJOrd9ruN0KfXszj04t+RBFZ7xxxXCJHz3RgnI/cHlz/Ioh8WMSJ7xHHtt7F4I6dmncrzlz3fIccRyuR8N4bkboyukvvx1qHtaunfv38/+/fv56yzzvIv27BhA0888QQul4trr72WsWPHtucpRAfwer2UlZWRkpIiAyp2IsndGJI7lDhLOFx/GND/eH2892NWFKwAIMocxZRBU8hLyAPArJjJjcsl2hrd1sMdE8ndGB2Wu88LrmqorwB3rd5676mHog2Nhf4Wffq9lqLi9UK/+wh9Sj2zTe+270iCqNh2/VyhTs53Y0juxpDcjSG5GyNcc29X0X/XXXdRW1vLp59+CkBxcTHnnnsubrebuLg45s2bx3vvvce4ceM65GCFEEIcXVFtEeWucgBUTWXBjgV8V/wdAHG2OG4cdCNZsVkAWBQLPeJ7YLeEX0usOAZed/P9+e5aQANvAxRt1KfXK96iT7/XUlQsZJ+m36ef2hdMFr34dySBPaFDBuQTQgghRMdqV9H/7bffcvfdd/u/nzVrFvX19WzevJmePXtyySWX8H//939S9AshxEmmaRqFtYVUuasA8Pg8vL3tbbaVbwP0EflvGnITyfZkAGwmGz3ie2Az2ww7ZmEAT0Nzoe9pnI/Y69Kn2CtYA8Wbggt9WwxkD9dH3k/tCyYz2OIgOlkv9E3mzv85hBBCCHHM2lX0l5eXk56e7v9+8eLFnHPOOeTn5wMwbtw4/vznP7fvCIUQQhyVqqkcqDlAracWgHpvPbO3zGZv9V4AsmKyuGnwTcTa9C7XDrOD3PhcLCYZyzUi+KfWq9Rb8gF8br0lv2CN3oXf5w7cx+rQC/1uIyC9v96ib3Hohb4jqcMG5BNCCCHEydeuT3xpaWns27cP0KdXWLVqFf/617/8671eL16vt31HKNpNURRMJlNID0YRjiR3Y0Ra7l7Vy/6a/dR76wGocdcwY/MMDtUdAqBXQi8mDZzk78IfY4khJy4Hcwe3zkZa7qGizdxdNS2m1mss6H0eKNnaXOg3XQBoYrFD9rDGQn+AXtibbS0G5HN0ys/UFcj5bgzJ3RiSuzEkd2OEa+7tKvovuOACnnnmGeLj41m+fDmqqgYM3Ld161ZycnLae4yinSwWS0CPDNE5JHdjRFLuHp+HfdX7cKkuAMrqy3h98+tUNFQAMChlEL/o/wt/i368LZ7usd1Pyh+ySMo9lPhz17TmIr+hCtTGC+6qVx9tv2CNPvq+p/6IB4iCzFP0wfgyBumFvmLWR+N3JEFUXGf/SF2CnO/GkNyNIbkbQ3I3Rrjm3q6i/1//+hfbt2/nd7/7HTabjf/7v/+jZ8+eALhcLt59912uu+66DjlQIYQQzVw+F/uq9+FR9fuvD9YeZMbmGdQ13qd9eubpXNX7KkyKCYDkqGT/AH4iTDRNrVdfqbfsN02jp3rh8A/Nhb67LnA/sw0yh+j36GcO1r9HAXs8OJJlQD4hhBAizLSr6M/IyOCrr76iqqoKh8OBzdY8IJSqqnz22WfS0h8CPB5PWE49Eeokd2NEQu713nr2V+/Hq+mtuXuq9jBryyxcPr3Ff3TOaC7scaG/RT/NkUZ69Mm9ah0JuYcETWueWq+hGo/XQ1mNi5QYC9aKnXqhf3Bd42j8LZitkDFEH3U/c6jewg9gi9ULfUeiDMh3HOR8N4bkbgzJ3RiSuzHCNfcOGcUpISEhaJnD4eCUU07piIcXQgjRqM5Tx/7q/aioAGwp3cLcH+bia5xD/bJel/Gzbj/zb58Vk+UfsV90Ya6axkK/Rdd9TYXS7cTsXoW5bBO4awL3MVn0LvvdR0DWKfo9+6APyNd0n75FZm8QQgghwl27i/79+/fzyCOP8L///Y+SkhIWLlzIqFGjKC0t5R//+Ac33XQTw4cP74hjFUKIiFblqqKwthANDYA1h9awYMcCNDQUReGavtcwPF1/v1VQ6BbbjYSo4Iuyootw1eoj7tdXQuNtHGgaVOyFgtVQuAZzfSV2nwpm/TYOTGZIH9hc6FujG5db9SI/OlkG5BNCCCEiTLuK/q1bt3L22WejqiojR45k586d/tH6U1NT+fLLL6mrq+O1117rkIMVQohIVd5QTlFdEQCapvFFwRcs3bsUAKvJynUDrqNfcj8ATJjIjc8lxhpj2PGKE+R2Nhb6FYHT6FUXwoHVerFfdzhgF00xoaUP1O/Rzx4GtsbXXTHr9+dHJ8uAfEIIIUQEa1fRf99995GYmMiqVatQFCVopMPLLruMd955p10HKIQQke6w8zAl9SUAqJrKkj1L+KrwKwAcFgeTB02mR3wPACyKhdz4XBwWac3tMjwNjV33KwOn0ast0e/RL1itF/0tKQqk9EXtdhoVcYNITknBbDbRPCBfEkQlgMnUmT+JEEIIIUJQu4r+L774gr/+9a+kpaVRVlYWtD43N5fCwsJW9hSdyWq1kpmZafRhRBzJ3RjhlvuhukOUNejvrz7Vx/wd81lfsh7Qp+C7afBNZMRkAHqLf158HjZz59+nHW65n3Ret17o11eAt8U0evUVUPid3qpfsSd4v6SekHM6dDsNHElYgAxoHJAvCeyJYO6Q4XrEUcj5bgzJ3RiSuzEkd2OEa+7t+mSgqirRLjVzDQAA5P9JREFU0dFtrj98+DBRUVHteQohhIhImqZRWFtIlbsKALfPzdvb3uaH8h8ASHWkcvPgm0m0JwJgN9vJjc/FagqfkWbDjs+j359fXwGeFtPouWr0EfcPfAtlO/T79luK76Z33e8+AmJb9KjzD8iX2DwavxBCCCHEEdpV9J966ql8+OGH3H777UHrvF4vc+fO5Sc/+Ul7nkJ0AK/XS3l5OcnJyVgs0gLUWSR3Y4RD7qqmUlBTQI1HH43d6XEya+ss9lfvB6BbbDduHHyj/579aEs0uXG5mA2cci0ccj8pfF59xP36isDR9T31ULRBL/RLvgfNF7hfTJpe6Oecrhf9TcxRepHvSAKrQ3I3iORuDMndGJK7MSR3Y4Rr7u36Sf70pz9x+eWXc9tttzFx4kQAiouL+fTTT3nkkUf4/vvvee655zrkQMWJ0zQNVVXRjmw9EieV5G6Mrp67T/Wxv2Y/Tq8T0Efsn7F5BiVO/Z7+/MR8Jg2cRJRZb9mNs8bRPa47JsXYe7e7eu4dSvU1F/quGmicbQGfGw5t0rvuF2/SW/5bciRCtxGQcwYk9tDv24fGkfcT9ULfFjg4o+RuDMndGJK7MSR3Y0juxgjX3NtV9I8ZM4Y33niDu+++m5dffhmASZMmoWka8fHxzJo1i1GjRnXIgQohRLjzqB72V++nwacP5lZaX8rrm16n0lUJwJDUIUzoNwGLSX/rToxKJDsmG6WpOBTGUVVwtSj0NbVxuVdvyS9YDQfXBw7UB/q9+N1O1Qv9lN7QdPFGMTcX+jLyvhBCCCHaod19Fm644QbGjRvHsmXL2LFjB6qqkp+fz8UXX0xcnHxQEUKIY+H2udlbvRdP43zshTWFzNgyA6dHb/EfmTWSK/Kv8Lfop9pT/QP4CYNoGriqG0fer2ou9DUVSnfohX7hd+CuC9zPYofs4Xr3/fT+0HgRB8WkT7HnSIKo+OaWfiGEEEKIduiQGxViYmIYO3ZsRzyUEEJEnHpvPfur9+PVvADsqtzF7K2zcTfO035e7nmcn3u+v0U/IzqDVEeqYccb0TRNb8lvqNQH5Wu6F1/ToGJvY6G/Rl/XktkKmUP1Qj9zMPhnWGicYs+eqH/JFHtCCCGE6GDtKvo//fRT/vvf//LII4+0uv7+++/n/PPP57zzzmvP04h2slgsZGRIi2Bnk9yN0dVyr/PUcaDmAL7G4nFT6Sbe3fau//sr86/kJ9n6gKgKCtkx2f4R+0NJV8v9uLlqm1v01Rb34lcVQMEavdivOxy4j2KGjIF6oZ91ClgdTSv0Lvv2RL0LfzsGYAz73EOU5G4Myd0YkrsxJHdjhGvu7Sr6H3roIXJzc9tcX1hYyMMPPyxFv8Hkfl9jSO7G6Eq5V7urKagpQGsc6O2bom9YuHMhACbFxIR+Ezgl7RT9e0x0j+tOnC00b5vqSrkfM7ezsdCv1Afha1Jbohf5Bauh+mDgPooCqX31Qr/bqfo9+02sMc1T7Jk7ZmrFsMy9C5DcjSG5G0NyN4bkboxwzb1d/Qg3bdrEyJEj21x/+umns3HjxvY8hegAXq+X0tJSvF6v0YcSUSR3Y3SV3CsaKvwFv6Zp/Hf/f/0Fv9VsZcqgKf6C36JY6BHfI2QLfug6uf8oTwNUF+mD75X+AHUlesFfXwE7lsH/HoFP/gJbFwYW/Mk9WWkaienXa3j4uwToOQqXZuOW+x4hbegFpA69kEn3/oMazQ5mK16vl2uuuYacnBwURWHv3r1HPay8vDyio6OJjY0lNjaWW2+9FQij3LsYyd0Y4ZL76DdGozyooDyo0OfZPj+6/TlvnOPfvv9z/TvhCAOFS+5djeRujHDNvV0t/S6XC7fbfdT1TqezPU8hOoCmaXi93rCbeiLUSe7G6Aq5l9aXUuwsBkDVVD7c/SErD64EwGFxcNPgm+ge1x0Aq8lKj/ge/in6QlVXyL1NXpd+D359BXjrm5e7aqBwrd6iX7ZDv2+/pfhu+qj73UegOlK498obOX3YIL27fmwmz82Zydrvd7Pth+3YbDbGjx/PI488wqOPPgrA2WefzW9/+1vOPffcYzrMTz75hLPOOitgWZfOvQuT3I0RjrnvLN/Jl/u/5Kzcs1pdv6t8Fyv2rejkowoUjrl3BZK7McI193YV/YMHD2bBggX85je/CVqnaRrvv/8+AwcObM9TCCFEWDlUd4iyhjIAvKqXedvnsfGw3iMqISqBmwbfRHp0OgBRpih6xPfA2kFdwUULPo9e5NdXgqfF6PqeeihaDwdWQ8nW5hH5m8SkNRb6p0N8tn/xy7MXMPKMM6hyuiA2HeKz2HugkEsuuYSUlBQAxo4dy9KlSwH9nsG77777JP+QQohQNiB1AN+Xfs+MdTPaLPrfWP8GGpp/WyGEOBHt6t7/61//mq+++ooJEyawadMmvF4vXq+XjRs3MmHCBFauXMmvf/3rjjpWIYTosjRNo7C20F/wu3wuZm2Z5S/406PTufWUW/0Fv8PiIC8hTwr+juTzQl0plO6E4s1QXagX/D63PrXequnw0e9gzQx9fVPB70iC3hfCuX+Gix6GgVfpBb/JAtEplJHEUzPe5cF/PdE8/R4wZcoUvvzyS0pKSqiqquL999/noosuOuHDv+aaa8jIyODqq69m37597U1DCGGw8QPGE2uL5b2t7/mnZ21J1VRmbpiJWTFzw9AbDDhCIUS4aFdL/6RJk9i1axcPPfQQ77//PqbGqYZUVUVRFP7yl78wZcqUDjlQIYToqlRNpaCmgBpPDQBOj5M3trxBQU0BADlxOUwZNIVoazQAsdZYcuJyMCkyfVu7qT59xP36Cr27fuOgiahevSW/YA0cXA/ehsD9bLHQ7TTIOR1SekPTa6GYwJ6gXwiIigdF4f4/3so999xDYmJiwEP07t2btLQ0MjMzAbjooov89+Ifr7feeotTTz0Vt9vNAw88wFVXXcXatWtP6LGEEKEhxhbDhIETmLF+BvO2zmPyKZMD1n+2+zMOVB/gsj6XkRWX1epjFFQXMH/rfJbsXMKO8h0crDmI1WSlT0ofru5/NXePvJu4qMDxYNYVreMnr/0Et8/N+z9/n6sHXB30uJMXTGb2xtkMzxzOF5O/6LgfWghhiHYV/QB/+9vfmDRpEgsWLGD37t0A5OfnM3bsWPLz89t9gKL9LBYLaWlp/osyonNI7sYItdx9qo/9NftxevVWnEpXJTM2z+CwU5/erW9SX64bcB22xnnbE2wJdIvt1uVGjw2p3FUVXI2FfkM1/kJfU6F0u951/+BacNcF7md1QNYwvet+ev8WrfYK2OMbC/0EaPEzrlu3jtWrV/P8888HHcbtt9+O1WqlqqoKVVW55ZZb+P3vf8+TTz553D/SmWeeCYDdbufJJ58kISGBXbt2+S8shETuESSkzvcIEo653zTsJmasn8GM9TOCiv4Z62f4t6lx17S6/1OrnuKJlU/gsDjIjM1kSPoQyurL2HBoA2uL1vL25rf58qYvSXIk+fcZnjWcxy54jHuW3sPURVM5Lfs0chOaZ+OatWEWszfOJsYaw9xr5hJjj8Fhc4RV7l1BOJ7vXUG45t7uoh/0Iv93v/tdRzyUOAkURcFsPvF5oMWJkdyNEUq5e1QP+6v30+DTW5FLnCW8vul1qt3VAJySfgrX9LkGc+M87Sn2FDJjMg073vYwPHdN01v0Gyr1/zd1zdc0qNjTOMXeGn1dS2YrZA7V79PPGNxiKj0FouL0Qt+eoA/O14rPP/+cH374gW7dugFQVVWFxWJh165dbNiwgSeffJK4OL2VbfLkydx///3t/lEVRUFRFDRNMz73CCW5GyMccz+7x9n0Tu7N53s/Z0/FHnom9QSgsqGSBdsWkOJI4Yp+V/DWprda3X9M7zFc0fcKzso9y/+3BOBA1QHuXHIni35YxJ8++xPTL58esN/dP7mbz/Z8xn+2/4fr5l/H8huXYzFZ2F62nTs+ugOA5y99nr4pfQHCLveuIBzP964gXHPvsEsYtbW1HDhwgP379wd9CWN5vV7KysrCbuqJUCe5GyNUcnf73Oyt2usv+A9UH+ClDS/5C/4zs89kQt8J/g9pGdEZXbbgB4Ny1zS9Jb9iHxzapBf39RV6l/6qAtiyAJbeD8v/BTs/ay74FTNkDoERU+GyJ2DkryB7uF7w22IhIUe/AJCSD9HJbRb8ANOmTWPnzp2sX7+e9evXc+WVV3LHHXfw73//mxEjRjB79mycTid1dXXMmTOHIUOG+Pd1uVw0NDQE/ftI+/fvZ+XKlXg8Hurq6vj9739Pbm4u+fn5IXO+RxrJ3RjhmvuNp9yIhsYb69/wL5u7eS4N3gauH3K9vydYa87vdT7n5J0TUPAD5CTk8Pb4t7GarLy56U18qi9o3xlXzaBbXDe+OvAVf1/+d9w+NxPnTaTWXcv1Q65nyjD9Ft1wzT3USe7GCNfc29XS39DQwIMPPshrr71GWVlZm9v5fMFvNKLzaJqGx+MJu6knQp3kboxQyL3B28C+6n14Nf0Pxo6KHcz5fg4enweAC3tcyOic0XqLLQpZMVkk2ZOO9pAhr1Nzd9U2dt2v1O/Nb1Jb0tiivxqqDwbuoyiQ2k/vut/tVLDFNK+zRoM9UW/Vt7T94bo10dHRREdH+793OBzExsaSmJjI448/zu23305OTg4Ao0aN4qmnnvJv269fP/+AfP3763NvN+XXdO//9OnTqampYdq0aezZsweHw8FPf/pTFi1ahNlsxuPxGH6+R6JQeJ+JROGa++RTJvPX5X9l1sZZ/H3031EUxd+1/+bhN//o/tWuat7Z/A5fF3zNwZqDOD1Of0YmxUStu5Yd5Tvon9o/YL+U6BTeHPcm5886n0e/fJQ1B9ew7tA68pPyefGyF/3bhWvuoU5yN0a45t6uov/2229n5syZjB07lrPPPpukpK79oVUIIdrL6XGyv2Y/Pk2/2Lnh8Abe++E9VE1FQeGq3ldxRtYZAJgw0S2uG/G2eCMPuWtwO5sLfZ+7eXl9RXPX/Yq9wfsl99IL/e6n6YV9E3OUXuQ7ksBq77DDfOONN/z/Tk9PZ968eW1uu3dvK8fbaPr05q64gwYNYtOmTR1xeEKIEJSTkMP5Pc9n2e5l/G/v/8iMzeTbwm8ZnjmcUzJPOeq+X+z7ggnvTaCkruSo25U5W2+cOyfvHP4y6i88+PmDLN21FKvJytxr5gYN/ieE6NraVfS///77/PKXv+Sll17qqOMRQoguq8ZdQ0FNASr6/eQrD67kP7v+A4BZMfOL/r9gcOpg//c5cTnEWGPafLyI56mH+kq9sPe5mpe7avQp9grWQNkOvZt/SwndGwv90yEmtXm52dbcom+LRgghQsVNw25i2e5lzFg/w3+r103DbjrqPtWuaq559xoOOw9zfs/z+eNZf2RoxlCS7En+6V5z/53LgeoDeFRPm49zYa8LefDzBwE4Lfs0RmSP6KCfSggRKtpV9CuKwqmnntpRxyKEEF1WZUMlB+sOoqGhaRqf7f+M/+7/LwA2s40bBt5AfqI+o4lFsdAjvgd2S8e1MIcNr0sv8usrwVvfvNzj1KfWK1gNJd83D9TXJDa9udCPz25ebrI0F/pRsZ3wAwghxPG7esDVJNoTef/794m1xWIz27huyHVH3eejHR9x2HmYnPgc/nPtf3BYHQHrNU2joqHiqI9R1VDFpAWTAP1WgFUFq3jmm2e4a+Rd7fuBhBAhpV1F/1VXXcWnn37Kr371q446HnESWCwWUlJSsFg6ZLIGcYwkd2MYkXtpfSnFzmIAVE1l0a5FfFv0LQAx1hhuHHwj3WL10d1tJhs94nscdWCmrqhdufs8jYV+hV7cN/G69AH6ClZD8WZ9u5YcSdB9BHQ/AxJz9fv2QR+oz57QWOjHNS8PQ/I+YwzJ3RjhnLvdYucXg37BS9+9hNPj5JqB15ASnXLUffZU7AHg9G6nBxX8AJtLNlPrrj3qY0xbPI29lXs5o9sZ/P7M3zPhvQnct+w+zulxjv/WgnDOPZRJ7sYI19zb9dM88MAD/PznP2fatGn86le/Ijc3t9UpDpKTk9vzNKKdFEXBarX++IaiQ0nuxujs3IvriiltKAXAq3p554d32FK6BYAkexI3Db6JVIfexdxhdpATn4PVFH7nxXHn7vPq9+fXV0DLD6WqF4q36oV+0QbwHjGifVQsdDtNL/RT8kFpnIRGMUFUPDgSISoBwmx+3bbI+4wxJHdjhHvuvzrtV+ws3wnAHaff8aPbR1v125SKaopaXf/4148fdf9XvnuFd7e8S3xUPG+Pf5teSb244/Q7eH7180ycP5E1t6whxhYT9rmHKsndGOGae7uK/j59+gCwbt06XnvttTa3k9H7jeXz+aiuriY+Pj4s550MVZK7MTord03TKKorosKld510+VzM2TqHXZW7AMiIyeCmQTcRH6UP0hdjiSEnLidoWqVwcUy5qz592rz6Cv2+fBrvxddUOLxdL/QPrgV3XeB+Voc+pV730yGtf4sp9BS9Jd+RpLfsh2m2RyPvM8aQ3I0R7rkPzxrOp5M/PebtR/UYBcDKgpW8/N3LTDttGqBPGfvQ5w/x5qY3sZltuFsOftpo6+Gt3LP0HgCmXzadXkm9APi/i/6PFftXsLF4I3ctuYvXrnot7HMPVZK7McI193YV/X/9619RwrjbZLhQVRWXy4WqqmF18oY6yd0YnZG7pmkU1BRQ7akGoNZdyxtb3uBgrT5NXG58LpMHTva3wsRb4+ke1z2s3y/bzF1V9Rb9hkpoqKa50NegYg8c+FYflK+hKvABzTbIGqoX+hmDwdziqrsttrHQTwRzeHW/O17yPmMMyd0Yknug4VnDuWHoDczeOJtfLf4VD37+INlx2ewo20GVq4qHz32YV9a+wr6qfQH7NXgbmDhvIk6PkxuH3ci1Q671r7Nb7Lw9/m1GvDyC19e/zoX5FzK+33jJ3QByvhsjXHNv16elv//97x10GEII0TX4VB8Hag5Q59Vbo8sbypmxeQZl9fp0SP2S+3Ft/2v99+wnRSWRHZvd5uOFJU1r0aJf3TzonqZBdaFe6BesAWdp4H4mM6QPgpzTIesUaDnQoTW6udC3hNd4CEIIcaJmXDWDwemDeX3d6+yu2E29p55Ts07lrpF3Mbb/WF5Z+0rQPr9Z+hs2lWyib0pfnhvzXND6gWkDefqSp5m2eBq3Lr6VUzNOJR6ZWlaIrqxDm0iqqqqIjY0Nq6siQgjRxKt62V+9n3qfPqr8obpDzNg8gxp3DQDD04czrs84fxf+NEca6dHphh1vp3M7odIJ3lrQWtzWVVusd90/sBqOvPdUMUFaP31AvuxTwdZiCkOLvbnQt8pMB0KI8LD8xuXHvc+Nw27kxmE3Bi03m8zc97P7uO9n97W639579gYte+GyF3jhsheO+ny3nHYLt5x2CwAej4eysrLjPmYhROhod9G/Zs0a/vKXv/DFF1/gdrv55JNPOO+88ygtLWXq1Knce++9jB49ugMOVQghjOPxedhbvRe3qt8bua96H7O2zKK+cVq5s7qdxZieY/xd+DOjM0lxHH3k5bBRXwGVRVBbAYodzCZwlkPhGr3Q///s3Xd8FHX+x/HXbMnuJpveG703kWpHsaIiFhALtrOf5bD8rGc99TgL9oJ6irSzoIgFxAaICjZEem8pQEjfJLub3Z35/THJJksSamCSzef5e/g7dtrOvmczyWfmO99v6baG6yR01u/oZw7Un8evZY6oK/QjIo/YRxBCCCGECFeHVPT//PPPDB8+nMzMTMaNG8fbb78dnJeUlERZWRmTJk2Sot9gFouFhISEsBt6oqWT3I1xOHL3+D1sK9+GX/MDsK54HdPXTMev6q/P6ngWJ2XpHSopKGQ6M4m1xTa5vbDhKQPXTvBVYdE0EqzVWLb+ohf7hRsaLh+bXVPoD4KopLrpJote5Dvi9d75xX6T84wxJHdjSO7GkNyNIbkbI1xzP6RP88ADD9CzZ0+WLFmCy+UKKfoBTjnlFN57771D2kFx6BRFISJCnoE90iR3YzR37lW+Kra7thOoaa7+564/mbl+JhoaCgoXdL2AQWmDADBhIjs6G2dEmBeunvKaYr+ml/2SbShrvyBi54q65/drOVP1zviyBkNMet10xazf4XfE6z3wh3Enh4eTnGeMIbkbQ3I3huRuDMndGOGa+yEV/b/99hv//ve/sdlsVFRUNJifmZnJzp07D+UtRDMIBAK4XC6io6Olv4UjSHI3RnPm7qp2kevKRUUvZH/M+5E5m+cA+nOUl/W4jJ6JPQGwKBayo7ODPfaHJW+F/kx+dc35vjwPVn8O+UvRgICqYTYpKI54vcjPHgyx7eoKesUEtpi6Ifak0D9kcp4xhuRuDMndGJK7MSR3Y4Rr7odU9FutVlRVbXJ+Xl4eTmeY3/FqBVRVxePxEBUVFVZf3pZOcjdGc+Ve5i0jryIPDQ1N0/h629cszFkIgM1s48reV9IxtiMAVpOV9jHtsZltzfIZWpzqSv3OvlcfopCKXbDmc71zPk0fgk9TzLgzBmHvciLW5C56gQ+Aot/Jr31O32Rq9C3EwZHzjDEkd2NI7saQ3I0huRsjXHM/pKL/mGOOYebMmYwfP77BvMrKSt59912GDRt2KG8hhBBHXJG7iJ1VeislVVOZtWEWf+z6AwBnhJOre18dHIbPbrbTLqYdVpO1ye21WtVV+p392mK/shDWfgnbF9c141dM0O5YAl1HUKk6sUfb9WkR0XV39M3h9VycEEIIIURrckh/iT322GMMGzaMc845h0svvRSAv/76i82bN/Pss8+ye/duHnrooWbZUSGEOBIKqgrY7d4N6EP0/W/t/1hTtAaABHsCf+vzNxIcCQBEWiJpF90uOERf2PB59GLfU6q/dpfCujmwdRGo9YbiyxoMPUdCdBoEVKgCojMgOgnMYXgRRAghhBCiFTqkdpZDhw5lzpw5bNy4kSuvvBKAu+66ixtuuIFAIMCcOXPo169fs+yoEEIcbjsqdgQLfo/fwzsr3wkW/GlRadx41I3Bgj/aGk37mPbhVfD7vVCyFXav0Qt+rwtWfARfPwibF9QV/Bn94dSHYcj1esEfEa0PwRebqffKLwW/EEI0u0cXPIrymMLVn15t9K60CJKHEPvvkNtcDh8+nHXr1rFs2TI2bNiAqqp07tyZgQMHBserFsYym83ExcWF1XMprYHkboyDyV3TNHIrcimv1puxV1RX8M7Kd9hZqTfx7xDbgSt7XYndYgcgzhZHRlRG+Jzj/NVQsROqigFNf4Z/wzew6Tv9QkCt1D7Q6zyI76C/tkbpRb89BrOqEme2y/f9CJPzjDEkd2O09tyVxw7ud8a7o97l6v5XN+/OHIDWnntL8sKSFyj1lHJ1/6vpENdhr8tK7sYI19wPqeifMmUKJ510Eh06dKB///70798/ZP7WrVv54Ycfgq0AhDFMJhN2u93o3WhzJHdjHGjuqqayvXw7lX59+LlidzHvrHyHYk8xAD0Te3Jpj0uxmPTTZaI9kbSotObfcSMEfHoHfVVFgAZ+D2z8DjZ8DT533XJJ3aDXKEjqqr+2OPRi3xEXXES+78aQ3I0huRujted+fPbxjU7/KecnALomdCUlKqXB/NSo1MO6X/vS2nNvSV5Y8gLbyrZxcoeT91n0S+7GCNfcD6nov+aaa5g6dSodOnRodP4vv/zCNddcI0W/wQKBAJWVlWHXC2VLJ7kb40By96t+tpdvxx3QC9z8inwmr5pMRc2QdANTB3JB1wsw1fREnxqZSpIj6fB+gCMh4K+5s1+kd8gXqIZNC2D9V3XD8QEkdIRe50NyD314PYu9ptiPb7hJ+b4bQnI3huRujNae+49/+7HR6bUtAB448QFD7+g3pbXn3lpJ7sYI19wPqejXaoZqakplZSUWi/TabDRVVamqqsLhcITVl7elk9yNsb+5+wI+tpVvw6vqzde3lG1hyqopeAP662HZwzij/RkoioKCQkZUBnH2uCPxEQ6fgB8qC6Byd02x74OtP+qd9HnK6paLzdbv7Kf11Yt9cwREp+vFfhOPNMj33RiSuzEkd2NI7saQ3I0huRsjXHM/4I78li9fzpQpU5gyZQoAixYtCr6u/99LL73Es88+S7du3ZptZ5988kkURaFPnz4N5v3888+ccMIJREZGkpaWxu23305FRUWD5bxeL/feey8ZGRk4HA6GDh3KN99802z7KIRo+bwBL1vKtwQL/jVFa3hn5TvBgv/sTmdzZoczURQFEyayo7Nbd8GvBqB8BxSshopddcX+Nw/BX/+rK/ij02DIDTD8QUjvpxf7sdmQ0gsiE5os+IUQQhxZATXA84ufp9/r/Yh8MpL4/8Rz7oxz+SP/j0aXv/rTq1EeU3h0waOUecq495t76f5KdxxPOujwQoeQZZfkLmHszLFkTszE9oSN9OfTueTLS/hk7SeNbnvB1gUojykNttPU+zfm283fcvrU04mdEEv0v6MZ+vZQ3lv2HgAdXuiA8pjCgq0LDkseJe4S/jH3H3R8sSO2J2xkTszkhs9vIN+V3+i6+9qfxvKYvGwyymMK28q2AXDKe6egPKYE/5POCMXhdsC34WfNmsVjjz0GgKIoTJo0iUmTJjW6bFxcXPDiwKHKzc3lqaeeIioqqsG8ZcuWceqpp9KzZ08mTpxIbm4uzz77LBs2bGDu3Lkhy1599dXMnDmT8ePH07VrVyZPnszZZ5/N/PnzOeGEE5plX4UQLVeVr4ocVw5+zQ/A7zt/Z9aGWWhoKIrC6K6jOTr1aADMipl20e2ItEYaucsHT1X1u/qVBaD69bv7ub/Bms+hoqBuuahkfei97CGgmMBkAWcqRCaB6ZAGeRFCCNHM/Kqfs2eczdebvqZLQhe6JXZjbeFavtzwJd9v+Z6FVy9kcObgRtctqipi0FuD2FS8iR5JPeiV3At3vT5cnl/8PHd9fRcaGvH2ePql9iO/PJ+FuQtZmLuQOZvm8O6od4OPvTWH1357jVvm3AJAnD2O7ondyS3P5erZV7OiYMU+1z+UPErcJQx5ewibijfRM7knPZN6srJgJW8tfYvZ62az8OqF9EjqccifMTUqleOzj+f3/N/xBrz0SelDrC02OL9bYvPdJBWiMQdc9N9www2ce+65aJrGkCFDePzxxxkxYkTIMoqiEBUVRefOnZutef/dd9/NMcccQyAQoLCwMGTeAw88QHx8PAsWLCAmJgaADh06cP311/P1119zxhlnAPDrr7/y/vvv88wzz3D33XcDcOWVV9KnTx/uuecefv7552bZVyGEfq74/PPPqayspH379jz11FOMHDkSr9fLrbfeyqeffoqmaZx11lm8/vrrREdHh6x/0003MWnSJHJycsjKymr0PT744AMeeughdu7cSbdu3XjppZcYPLjxX+yg98qf48pBRQXgh9wf+GrLVwBYTBYu73k53RO6668VC+1j2gd77G9VVBWqCvW7+qofNA3y/4Q1n0F5vTsXjnjocS60P1Yv9BUzOFP0iwDhNBShEEKEkQ9XfUhmTCa/X/87AzMGAlBYVcio90fxc87P3P3N3Sy8emGj677+++v0SenDulvX0TVR75y1tuifv2V+sOB/+KSH+edJ/8RqtuLz+XhryVvcsfAOpvw1hb4pfbn7uLub5bMs37Wcf3z1DwD+77j/44nhTxBhjgDgvWXvcd3n16Gw91Zmh5pH+7j2rLh5Bb1TegOQU5bD6I9G82ver4ydOZalNyw95OF5R3QdwYiuI+jwQge2lW3j5REvc3KHkw9pm0IciAO+TJeens7AgQMZNGgQ8+fP54YbbmDgwIEh/w0YMIDu3bs3W8H/ww8/MHPmTF544YUG88rLy/nmm28YN25csOAHvZh3Op18+OGHwWkzZ87EbDZzww03BKfZ7XauvfZaFi9eTE5OTrPsb0tjNpuJjY0Nq+dSWoO2nvudd97J1q1bKS8v55133mHcuHEUFRXxyiuvsHTpUtauXcuWLVsoKCjgqaeeCll36dKlrF69eq/b37lzJ1dddRWvv/46ZWVl3HDDDYwePbrJ3Mu8ZWx3bUdFRdVU5myeEyz47RY71/a9Nljw20w2OsV2an0Fv6ZBZSGL58zAFJPGExNfh50rmPzYDQy48HZibvyC9g8vZ8L3xdBvLJzxBHQ8UW/G70xlwuQvUGLS+fHnxY1uvqCggEsuuYT09HTi4uI47bTTWLNmDSDfd6NI7saQ3I0huet8qo8p508JFrgASZFJvDziZQAWbVtEWf1+Wuoxm8x8esmnwYIfwGF1APDEoifQ0Di769k8dspjWM1WfR2zmWsGXcP/Hft/APznp//grT+c6yF4bvFz+FU/p3Y8ladPfzpY8ANc1f8q7jr2Lnyqb6/bOJQ8fKqP985/L1jwA2THZvPh6A+xmCws37Wc2etmH8pHPGjyfTdGuOZ+SFX5sGHDmms/mhQIBLjtttu47rrr6Nu3b4P5K1aswO/3M2jQoJDpERER9O/fnz///DM47c8//6Rbt24hFwcAhgwZAuiPCWRnZze6HwUFBezevTtk2saNGwHw+Xz4fHUnJLPZjMlkQlVVAoFAyDomkwmz2XzQ8zRNw+/37/c8RVGwWCzY7Xb8fn/Idmvn7W292s+3J6tV/0Xg9/sbdOhosVhQFOWg5wUCAVRV3e95tXkf7LzDdZwCgQAWi4VAIEAgENiv43Sw86DlHafOnTsH9ysQCFBdXU1eXh5btmzh9NNPD/4cjhw5MtivRm1Wt912GxMnTuSYY44JvseeeW/bto3ExEROPfVUVFVl7Nix3HjjjRQWFhIfHx/cb1VVKawqZFflLv091ACzNs9i2e5laGjEWGO4utfVpEamovpVHFYHHWI7YFbMDTJtscfJ50OrKobKAlSfh/EP/4fBvTvBlkVoP/9GVfluXhrTjsFdksiJO5ZzH/qAdiuqGdvJimpLAGcKeTt2MeN/75Oenh48Fnse3/Lyco455hgmTpxIQkICL7zwAueddx6rV6/GbDbjcDhQVbXBZzTqvNfijtNhOO9pmhZynqk/z4jzXls6TrW5K4rS6n4/tebjZLfbm/XnyaSZgneS6/+7loqKpmgN5kUoESia/rq5jlPtz3FTx6J2v/ul9uOYjGNCclUUhaPTjsZmtuENeFm3ex1Hpx0dnFdreIfhZEZlBtetPU6V1ZUs3KrfDb990O3B+VarFZPJhNVq5bbBtzHhpwkUVhXy07afOKXTKcFMa9Wut+cxrN33QCCApmnB9b7aoF98v7rf1fh8vgbH6Zp+1/Cfn/4T8vlr59W+b9+UvsHhEOsf3z6JfYJ5bCrZRP/U/sF1apcZlDGI47KPa3AsMqIyOL/7+cxcM5Mv1n/ByC4jaUxjx6mxPBpbb3/+xqg9z6iq2iLOe/szLxzOe7Xf+5Zy3mtsXlPfraYc8q34efPm8d///pfNmzdTUlLSYCcVRWHTpk0Hvf033niDbdu28e233zY6f8eOHQDBP1TrS09PZ9GiRSHLNrUcQH5+4x12ALz22mvBvgz2VFZWRlFRUfB1dHQ0UVFRuN1uXC5XyLJRUVFER0fj9XopKwu96hgZGUlMTAzV1dWUlpaGzLPb7cTFxeHz+SguLg6ZZ7PZiI+Px+/3h+wH6F+uuLg4XC4XVVVVmOo9n2uxWEhK0ocf23M9k8lESkpKo/MA0tL0ccqLi4sb/NCmpurjyZaWljb4YUhOTsZsNlNWVtbgy5qYmIjVaqW8vByvN/QKckJCAhEREbhcLjweT8i8uLg47HY7lZWVVFVVhcyLjY3F4XBQVVVFZWVlyLzDfZwKCwvxer3YbDZMJtM+j1NiYiKqqjaY15qP03333ccHH3yAx+PhrLPOom/fvowePZr777+fdevWERERwUcffcQFF1wAQFVVFe+88w4dOnSgffv2AMHjvedxysrKomPHjsybN4/jjz+e119/naOOOgq/309ubi5xcXHEx8eTX5ZP/m79Z9un+vh4+8dscG3AZDaRaEvk0uxLifHF4C5zY7fYiY+Lx2Ky4PP5Wsdx8pZRlr8FX7We07R332ZgcjWVpaXgdhNQ07nuxCzc7YZT2v5knBY7o0YUsviv9Zx9yfV4fAEoKWP8+PGMHz8+eJ5r7OcpMzOT8ePH43K5KC0t5dJLL+WBBx5gw4YNZGVloSgKmqY1WM+I816LO06H6bxXXl5OaWlp8DwDxp732spxUlU1eH5PSkpqdb+fWutxUlWVyMhIoqOjm+3nyel3EmnS+22JrPm/+sopx4MHBw6iqOtXqp21HXb01mDNdZwqKipwuVxNHqfaz9QlvkuTxyklKoWc8hxyd+fSztoOIKTlbceojiHr1h6njcUbCWh6wZRhyQguk5aWRiAQIC8vD6vVSlpUGnkVefyx7Q9O6XRKcL9BL3Zr19vz56l2391uN36/H6vVSs7uHAqq9P5l2tvaU1RU1OA4xRJLdEQ0rmpX8LtQ+/PkduuPJbR3tsfr9Tb685ToSCS/Ih+X1xVynGr3p0d8j+DrPY9T11i9NcTqgtUhme1Z1O95nOr/rO95nOrfyNjXz1P984zFYjH8vFdfa/y7fH/Pe7UXDzIyMoCGebeU30975rAvh1T0P/PMM9x3332kpqYyZMiQRu/EH4qioiIefvhhHnroIZKTkxtdpvYH3mazNZhnt9uD82uXbWq5+ttqzN///nfGjBkTMm3jxo2cf/75xMbGkpiYGJxe2xzE4XAQERERsk7tH2U2my1knfrzIiIimpxXe1JvbJ7FYmkwT1EUVFXF7XYTHx8fcuKvf+W3sfWamldfQkJCk0M3xsXFNZhX/4/Sxq5iAcTExDR6RRHqTgj11eYdFRWFw+FodF5kZGTwOO8573Adp/j4eIqLi4O57+s41a5/IMeipR+nt956izfeeIOFCxeyZs0aFEWhb9++pKen069fPwBOP/10br31VkD/5fnKK6+wYMGC4OepPW6NHadx48Zx4YUX4vV6iYuL46uvvgrm7nA42FGxg9JAKY5YB1X+Kt5f+z7bKrehmBQynZlc2ftKokz69ykmIoYMZ0brOU6eUkyF6yDgJdYGmruAkl8/5O0pn/LzXT248+MiMJkx9RiB1vl0IiOi9D9l7XH89Odaxl1xJdFxCUSpKgsXLsTlcnHFFVfw+OOPA/v38/Tdd9+RmppK165dUVWVkpIS4uPjG6xnxHmvdhnDjxOH97xX+8dT/fO7kee9tnKc/H5/8Pxee+emNf1+aq3Hye/3U1paSlRUVLP9PFVYKqhS9cKkiirchP4tWNsHjBs3HuqKm+2+7cSid8TWXMfJ6XQG+7dp7DjV/g0bbYtu+jjVdLAXExMTXCbkWMQkhqxbO89V7Qqu3z2re8g6tXeZ4+PjyYzJJK8iDy2iLnun0wno39vabe/581S77w6HI3gsFFvde2SnZpMYndjocYqxxeCqdgXn1f481f5MxUfFB7e/589T7SMKGlrIcapdPiM2I/h6z0yz4vX+hCp9lY3+nQ+NH6fIyroLR00d+/35eWrsPNPYNutrSb+fWut5rzb32u+90ee9+vta/zjFxsY2WHdvDqnof/HFFxk+fDhz5swJ+TI2l3/+858kJCRw2223NblM7ZdpzytQoN8hrP9lczgcTS5Xf1uNSUlJCV652ZPVam3085tMppA7680xT1GUJrNuap7P5ws2N2ls/t62Cex13t76bTjYeWazucnnaA7HvMN1nCwWC2azuUHuB3MM9zUPWu5xslqtnHnmmbzyyit069aNadOmYbPZKCsrQ1VVrr/+eu69914mTpzIv/71L2666SYyMzOD26r/C7J+3l9//TWPPvoov/zyCz179mTmzJlccMEFrFy5EpPJxC73LirUChRFwRVw8e7qd9lVuQtFUegc15lxvcZhM+u/9BPtiaRFpYV8hhZ7nNyl4NoJ/po/TMvzsKz+HPKX8vD72xh/SgoJ0Q6ISYcuQzD1uVDvvMUeB9HpTHz5NYpLSrnqqquCzdr+7//+j6lTp4bs276Ob0lJCbfccgsTJkzAarUGrzzXNgNtzJE87x3KPGi5P09NzWvs/G7Eea8tHafa3Gv/yGtNv59a83Gqzbu5fp5URUVDC/67KXvOq9aq0RR9veY6TvX3rbFM6y+7r7+7m/qbz2w2Nzo9OkK/2KBqKiXVJaQ6UxusZ7FY2Fm5E4A4R1zdsTDrx0JDa7Dt2uPkDriD26ldL9ZRV7R4VE+Tv4OCFyRqPn/tvNr59XPc35+n2v/dXbW7wbxatfOibdEN/o4Dgo8p7PmZPYG6i0OH+vNUm3v973tr+/10oPNawnlvbz+H+7NNOPzH6UBr70Mab6OkpITRo0cfloJ/w4YNvPnmm9x+++3k5+ezdetWtm7disfjwefzsXXrVoqLi4NN82ub+de3Y8eOYNMM0JvxN7UcELKsEKJ5+f1+Nm7cyF9//cXVV19NdHQ0sbGxXHnllXz33XcAzJ8/nwkTJpCWlhZsAjVgwAC++uqrBtv766+/OOWUU+jTpw9ms5mxY8fidrtZu24tu927Ka8uB6DQXcik5ZOCz/T3TurNVb2vChb8KY6UBgV/i+Qph93roWSLXvBX7ILf3obvHof8pfyZU8Vv2yu5/rKL9A764juCxQ62GEjqDgkdmf7hxzz//PN8+eWXwYucr776KieccAJ9+vTZ711xuVyMGDGCsWPHctVVVx2uTyyEEOII6ZLQBYtJLyhWFqxsdJkSdwl55XkA9EzqGZweFaHf6S2oLGh0PYD1ResbTIuzx5ESpd9Q+2vnX42ut6VkC+Xe8v34BAdn1e5V+5xX/7MCRFn1z1v7d8WeGvusterfCRbiSDqkon/IkCGsW7euufYlRF5eHqqqcvvtt9OxY8fgf7/88gvr16+nY8eOPP744/Tp0weLxcLvv/8esn51dTXLli2jf//+wWn9+/dn/fr1lJeHnjx++eWX4HwhxKErKytjxowZVFRU4Pf7+eijj5g/fz4nnXQSgwYNYurUqcHnuaZNmxZ8NOi7775jxYoVLFu2jGXLlgF6vyGnnHJKg/cYOHAgCxYsYO3atWiaxscff4zH48GcaMbj16+y51XkMemvSZR4SgAYkj6ES3tcGvzDJiMqg+TIxh8dajG8LijcAMWbwFcJVUWwdAp88wjk/Kr32K+YWFiUyLoilczrppB27MV88PnX/Of1KVxz9xMQEcns2bO56667+Oqrr+jYsWNw8/Pnz2f69OnBCy05OTmMGjWKt956q9HdcbvdnHvuuQwcOLDBqAtCCCFap6iIKIa11zvonrhkYqPLvPTbSwS0AEmRSQzJHBKc3iWhCwoKHr+HP3f82WC9n7b/xPJdyxvd5lldzgLgv3/+t9H5by1t/HdRc/k171eW5C5pMH172XY+XfspAOd0OydkXu3IB4tzGo5041f9e93nSKve9L/KV9XkMkIcDodU9L/22mt88sknzJgxo7n2J6hPnz7MmjWrwX+9e/emXbt2zJo1i2uvvZbY2FhOO+00pk2bFtLpw9SpU6moqAh5Dn/06NEEAgHefPPN4DSv18u7777L0KFDm+y5v7Uzm81ER0eH3dATLV1bzl1RFN566y2ysrJITExkwoQJzJgxg/79+/PMM8/g8XjIzs6mXbt2eL1enn32WUB//qm2+Ky905+cnBx89m7EiBHBQnP48OHcfffdnHXWWcTExPDwow/z7FvPYo+xExEVwebyzby1/C0qfXpHMcPbDWdU51GYFL0H5uzobOLt8Qaks5+qK6FwIxRthOoKvVn/sv/B1/+ErT+CVtPUNGswnPYIN9w/gY0/fcay72ez7NfFnHfeKG655Raef/55vvvuO6699lo+++wzevfuHfI2kydPZvXq1cELLRkZGbz77rtcfvnlDXbJ5/Nx0UUXkZGRwWuvvRYyry1/340kuRtDcjeG5H54PXjigygozNkwh0cXPIovoD+2ZTabmZc3j2cX67+r7zv+PmyWuj6y4uxxnNj+RAD+8dU/KKqq68Dszx1/cuWnV2I1Nd4q+K5j78JisvDN5m+4/9v7g+8JMPWvqTz787NNrtscrCYrV316FWt2rwlOyy3PZezMsfhUH31S+nBe9/NC1jmvm/76v3/+l/lb5genl3vLuf7z69lYvLHJ9+uS0AWA7zZ/t899k++7McI190N6pn/s2LH4/X6uuOIKbr75ZrKyshoEpCgKf/3VeJOdvUlKSuL8889vMP2FF14ACJn35JNPctxxxzFs2DBuuOEGcnNzee655zjjjDM466yzgssNHTqUMWPGcP/991NQUECXLl1477332Lp1K//9b+NXGMOByWRq0MmGOPzacu4xMTHMnz+/0XkpKSnMnDlzv7azZ0cmc+fODXl91113ceedd7KrahdFnro/MtZVrOODtR8EeyIe2Xkkx2YcC4BZMZMdnR1sntfiVFeBawfUNmf0umD9PNg8H+r9MUR6f+h1HsTqHQ1FWhxEZvYGe13HUk6nk7i4OJ544glKS0sZPnx4cPVx48bxxhtvEBcXF/L2ZrOZhIQEIiP1uxE33XQToI+k8vPPPzN37lwcDkfI0KerV6+mXbt2bfb7bqS2fJ4xkuRuDMn98Dql4yk8e8az3P313Ty28DFe+uUluiR0Id+VT55Lb9Z/Rb8ruOPYOxqs+8zpzzBs8jAWbV9E1vNZdE/sTpWvig3FGziz85kcm3Us01dMb7Bev9R+vHDmC9w691Ym/DSBSX9MCnnPO4+5k4/XfMy2sm2YleYvwm4edDNzNs6h92u96ZXcC6vZysqClfhVP8mRybx/0fvB1oG1rjjqCib9MYlf8n7h1Cmn0iGuA/GOeFbvXo3dYufZM57lH1/9o9H3u7LflXy27jMmLpnIp+s+JTM6E5Ni4qwuZ3HfCfeFLCvfd2OEa+6HVPQnJCSQmJhI165dm2t/DsqAAQP49ttvuffee7njjjuIjo7m2muv5d///neDZadMmcJDDz3E1KlTKSkpoV+/fnzxxRecdNJJBuz5kVHbe7/D4WiygwvR/CT3w8+v+sl15VLprxv25Zf8X5i9cTYaGmaTmTHdx3BU8lEAWBQL7WLa4bA03WmnYXxuvdj31AzBUl0FG7+Bjd+Bv95wOKm9oed5kFDTRN9ih+g0cIS2Wpg8eXLw301dgGnM1q1bQ16/8cYbwX8PGzasyd5m5ftuDMndGJK7MST3w+/OY+/kuOzjmLh4Ij9u/5FlO5cRbYtmePvh3DjoRi7uc3Gj6w3JHMKP1/zIIwse4aecn1hXtI7O8Z15+rSnuePYO7jus+uafM9bhtxCt8Ru/Oen//Br3q+s3r2ansk9efyUx/nb0X/jvb/eA/Re/JtbvCOeX6/7lUcXPMrsdbPZUbGD5Mhkzu56No+e/ChZMVkN1rGYLHx9xdc8vvBxZq6eSW55Lm6/m7G9x/LoyY+ytXRrk+93Ua+LeOe8d5j0xyRW717NlpItaGh0iOvQYFn5vhsjXHNXtKb+ghP7tGrVKvr06cPKlSsbNJltSWrHG68db1McGZL74VXlqyK3Ihefqt/99qt+vtz8JUvyl6AGVOwRdsb1GkfXeP2iZIQpgvYx7YkwR+xts0ee36sX++6Smtce2Pg9bPga6j/zl9QNeo2CpJqLrOYIiE7Xi/0W0DGQfN+NIbkbQ3I3xuHI/bEFj/HZus8YlDHogNZbvms5A9IH8Oo5rzbLfrRkRn7fd1fuJuXZFBQUSu4tIdZ+YMOUNeXqT6/mvb/e45Fhj/DoyY82yzabm5xnjNFacj/QOvSQ7vQLIYQRij3F7KzcGRxmqdxbzoy1M9hevh2ASEskV/e6mvbx7QFwmB1kx2Qf1ucCD5i/ul6xr0GgGjYvgPVfgbeibrn4jtB7FCT31It7k1W/sx+Z2CKKfSGEEOJweXvp24D+GEBzFfxCtEWHXPSXl5fz2muvMX/+fAoKCpg0aRJDhgyhuLiYyZMnc95559GlS5fm2FchRBunaio7K3dS4i0JTttStoX/rf0fFdV6oZzpzOTC9AtJi9Y7AoyyRJEdnY3Z1EI6ZAn4wLVT74UfTX+97SdY+2Vd037Qn9XvNQrS+tUU+xZwpkJkEoRRczMhhBBt24wVM4i3x3NG5zOCv6sDaoD//vlfHlv4GAC3DbnNyF0UotU7pKI/NzeXYcOGkZOTQ9euXVm7di0VFfof3gkJCUyaNIlt27bx4osvNsvOCiHaLl/AR44rB3fADeid/C3esZgvN38ZfM58cNpgzml/Dj6X3uQ/xhoT7CTHcAEfVOzSi31NBTUA25fA2i9qLgDUcKbpHfRlDgDFBIoZnCkQlQwt5cKFEEII0UyW7ljKc4ufwxnhpGtCV8wmM+uL1lNe06HtZX0v429H/83gvRSidTukov///u//cLlcLFu2jJSUFFJSUkLmn3/++XzxxReHtIPi0NX2QhlOnVG0BpJ786moriCvIg+/5gegOlDNrI2z+KtAHxnErJg5r8t5DE4bjKZq4NA758mMzkQxugl8wA+VBVC5Wy/2NRVyf4M1X+gXAWpFJkGvkZA1RC/uFZNe6DtTW0WxL993Y0juxpDcjSG5G+Nw535x74sp9ZTy4/Yf2VK6hYrqCuLscZzR+Qyu6X8NY3uPNf53uQHk+26McM39kIr+r7/+mjvuuINevXpRVFTUYH6nTp3Iyck5lLcQzaB2vElxZEnuzaPQXciuqrriuNhdzLQ109hZuROAmIgYLu91OdnR2QAoJoWMhAxSo1IN2d8gNQAVtcV+ADQNdiyD1bOhPL9uOUc89DgH2h+nN+FHgagk/Y6/ufV0uyLfd2NI7saQ3I0huRvjcOc+JHMIQzKHHLbtN2by+ZOZfP7kI/qeB0q+78YI19wP6S9Kt9tNcnJyk/NdLtehbF40E1VV8Xq92Gy2sLtq1ZJJ7ocmoAbIr8in3FcenLaueB0frPsAT80Qdp1iO3FJj0twRjgBMGEiLTINO3ZUVTUmd1XV7+xXFNQV+7tW6cV+6ba65Wwx0H0EdDwJzFZAgcgEvdi3tLARBvaDfN+NIbkbQ3I3huRuDMndGJK7McI190P6JL169eKHH35ocv6nn37K0UcffShvIZpBIBCgrKyMQCBg9K60KZL7wfMGvGwp2xIs+FVN5fvt3/PeqveCBf8JmSfwt75/Cxb8EaYIOsZ2xGlxGpO7quqFfsEqvVd+LQC718LCp+Hnl+oK/ogo6HMhnPkkdDlVL/gdCZDSE+LatcqCH+T7bhTJ3RiSuzEkd2NI7saQ3I0Rrrkf0p3+8ePHc9VVV9GvXz/GjBkD6FdHNm7cyGOPPcbixYv5+OOPm2VHhRBtQ5m3jPyKfFRUADx+Dx+u+5C1xWsBsJqtXNT1Ivol9wuu47Q6yXJmYTaZ8fl8R3aHNQ0qC/Xn89Wa9y7aBGs+g4I1dctZ7ND1dL3Qt0bq0+xxEJ0OVvuR3WchhBBCCNFmHFLRP27cOLZt28Y///lPHnzwQQDOOussNE3DZDLx1FNPcf755zfHfgohwpymaeyq2kWRp65/kJ2VO5m+ZjpFbn1aoiORcT3HhTyvn+xIJiUypcH2DjtNg6piqNgJgWp9Wul2vRn/zhV1y5kjoPNw6HYG1LRKwBajF/sRkUd+v4UQQgiDbS3dSscXOwKgPaIZvDdChL9D7iXqwQcf5IorruDjjz9m48aNqKpK586dufDCC+nUqVNz7KMQIsz5VT+5rlwq/ZXBact3L+fjDR/jC+h3z3sk9ODi7hdjt+h3xc2KmQxnBjERMUd+h6uKwbUTAl79dXk+rP4M8pfWLWOyQKdh0O0ssMfq0yKiIToNbM4jv89CCCFajMKqQl759RXmbpzLusJ1VPoqibfHkxKVQu+U3pzU7iRGdh9Ju9h2Ru+qOAhXf3o17/313n4tO3nUZK7qf9UBbT/flc/ExROZt2kem0s24wv4SI5KZmjmUG4adBNndD7jYHZbhLFm6Rq6Xbt23HHHHc2xKXEYmEwmIiMjw6ozitZAct8/Vb4qcity8dU0jVc1la+2fMWPeT8Glzmt/WmcnH0yJkXP0maykR2Tjc1sa7C9w5q7u0Qv9mv6FaCiANZ8Drm/6nf+ARQzdDgeup+td8wHYI3Si327ARcojhD5vhtDcjeG5G6McMl9Se4Szp1xbrAVW5ozjc4JnQmoATYWb2TV7lV8uOpDSjwl/POkfxq8t+GT+5HULbEbx2cf3+T8PFceW0u3AnB8u8aXayr33/J+44xpZ1DqKcWsmOkY3xFnhJPNJZuZtXYWs9bO4v+O+z+ePv3pZvs8bUm4ft+bdTyotWvX8tFHH7Fjxw569OjB1VdfTUxM+P6R21qYzWY5DgaQ3PetxFPCjsodaOgFc0V1Be+vfZ/NZZsBcFgcXNz9YrondA+uExMRQ6YzM3gBYE+HJXd3aU2x79ZfVxXB2i9h28+g6X0PoJggeyj0PBeiakY1sTggJr3uTn8Yk++7MSR3Y0juxgiH3CuqK7jwgwspchcxNHMoL494mcGZg4PzVU3l9/zfeX/l+8Tb4w3c0zrhkPuR9sCJD/DAiQ80OX/U+6PYWrqVk9qfRJeELo0u01jumqYxbtY4Sj2lDMoYxPsXvU/nhM4AeP1enlz0JP/64V888/MznNP1HIZ1GNZ8H6qNCNfv+wEX/a+88govvfQSP//8M0lJScHpn3/+OWPGjKG6ujo47aWXXmLJkiUhy4kjT1VVqquriYiICLurVi2Z5N40TdPYUbmDEm9JcFqOK4fpq6dTXq332J8Wlca4nuNIcOh3yxUUUiJTSHLs/XzSrLl7yvVi31fz2IGnFNbOha2LQPXXLZc1CHqM1At80Dvti04DR8v4g+1IkO+7MSR3Y0juxgiH3OdsmMOOih2YFTOzxs4iPTo9ZL5JMRkybv3ehEPuLcmuil3M2TAHgGuPvrbJ5RrLfU3hGtYXrQfgzXPfDBb8ADaLjcdPeZyvNn7Fb/m/8fn6z6XoPwjh+n0/4E/y2Wef0blz55BC3u/3c91112E2m3n33XdZsWIFEyZMYNu2bTz55JPNusPiwAUCAUpLS8Nu6ImWTnJvnC/gY0v5lpCC/9cdv/LmX28GC/7+Kf256aibggW/RbHQLrrdPgt+aKbcvS4o3ADFm/SC3+uCFTNh3oOweX5dwZ9+FAx/CIbcoBf8ZhvEtYfkHm2q4Af5vhtFcjeG5G6McMh9U/EmAJIikxoU/HuztXQrymMKymMKAJ+v+5yTJ59M/H/icT7l5Ji3j2H68ul73Uapp5THFz7OwDcHEjshFvsTdrq/0p27v76bgsqCJtcLBAJMXzqdkf8bSeqzqUT8K4LUZ1M5//3z+WFb00N3a5rGW3+8xcA3BxL5ZCSJTycyYvqIva4D0OGFDiiPKSzYuqDR+Qu2LkB5TKHDCx32uu7KgpVc/NHFpD2bhv0JOz1e6cG/Fv4rOPSvUab8NQW/6ifGFsPoXqObXK6x73uVryr4766JXRtdr3Z6bZ9IAC6viy4vdUF5TOEfc//R6Hrv/PkOymMKMf+OCX5P26JwOM805oCL/tWrV3PMMceETJs/fz67d+/mjjvu4KqrrqJ3797cc889XHzxxcyZM6fZdlYI0bpV+irZXLYZd00zeb/q55MNn/Dpxk8JaAEUReHczucyptsYIsz6ePUOs4NOsZ1wRhyBzu+qK6FwIxRthOoKqK7Se+Of9yBs+Bpqf4Gm9IKT74djb4G4bDBZITYbUnrqz/EryuHfVyGEEK1OjE1vNryrchcbijYc1DZe/uVlznv/PFYUrKBLQhecEU5+yfuFcbPGcfvc2xtd56+df9HntT48suARlu9aTnJkMl0Tu7K1dCvPLX6O/m/0Z2XBygbref1exn48luu+uY45G+egaRp9UvrgV/3MXjebkyefzLM/P9voe149+2pu+OIGlu5YSoIjgU7xnfhp+08Mf284n6z55KA++/76JfcXhr49lM/WfUZmTCbZsdmsK1rHwwseZvh7w6msrtz3Rg6Td5e9C8AlvS8h0npgo/j0SOoRXGfRtkUN5nv9Xn7J/QWAY7Lq6rVoWzTvj34fq8nKS7++xOfrPg9Zb23hWm6bexsAb5z7RkgLAhEeDrh5f1FREdnZ2SHTvvvuOxRF4YILLgiZfvzxx/PJJ4f3h1oI0ToUugspqCoIPr9f6i1l+urp5FXkAeCMcHJZj8voENshuE6cLY70qPQmn99vNtVV4NoBXr2lAX4PbPxeL/TrXVUnsQv0Ph+SuumvTRZwpkJkEoRREzAhhBCHx4iuIzB/ZSagBThj2hncd/x9nN31bLJjs/e9co27vr6Lh056iIeHPYzFZEHTNCb9MYlb5tzCy7++zIntTmRM7zHB5YvdxZz7v3PJc+Vx/YDreXL4kyTX9D1T5inj9q9uZ8pfUxj94WhW/n0lFlNdeXDHvDuYtW4W3eO7M+ncSQzrVNdcfPry6dzwxQ3c8809DM4YHNKU/L9L/8uUv6ZgMVl4d9S7jOs3DgC3z83tc2/nvm/vO+gM98dD8x/ijM5nMOWCKcFWgz9u/5ELPriAxbmLueebe3j1nFdD1nlq0VPBZvcH6se//bjvhYDFOYtZU7gGgGsHNN20vynOCCePDHuEe7+9l2tmX8NzZzzH6Z1PxxnhZPXu1Tw8/2E2lWzi1I6nMrbP2JB1B2UM4t+n/pu7v7mba2Zfw7KblpEVk4XH72HszLFU+aq4uv/VXNb3sgPeL9HyHXDRn5qays6dO0OmLVq0iMjISI466qiQ6REREURERBzaHgohWrWAGiC/Ip9yX3lw2qbSTby/9n0qa56VbxfTjst6XBa8A6KgkBaVRoI94fDunM+tF/uespqdrYbNC2H9V3qT/lrxHaDXKP0Ov6LoPfQ7UyAqRYp9IYQQ+61TfCeeP/N5xs8bz9bSrdz05U0ApEalMiB9ACe2O5HRvUY32XQb4OQOJ/P4KY8HXyuKwk2DbmLpjqW8tfQt/vXDv0KK/omLJ5Jbnsuo7qN4c+SbIduKtcfyznnvsGLXCv7c+SefrPmEi3tfDMC6wnVM+mMSMbYYpo6YSv/s/iHrXt7vcnZV7uKur+/iPz/9J1j0a5rGv3/8NwC3Dr41WPADOKwOJo2cxKLti1hXtO4gEtw/tXe267cSPKHdCbx41otc/snlvLX0LR4e9jCpztTg/PVF6/kp56fDtk8A//3zvwD0Tu590P023HP8PWRGZ/Ls4mcZN2tcyLykyCQmnjGRW4fc2ugNkzuPvZPvtnzH3I1zufyTy/n+yu+5a95dLN+1nG6J3XhlxCsHtU+i5Tvgv1YHDRrEe++9h8ul/0G8atUqfv31V84880wsltBrCGvXriUrK6t59lQcNJPJhN1uD6vOKFoDyR28AS9byrYEC35N0/gh9wfeWfFOsOA/Jv0Yrut7XbDgtygWOsR0OOiCf79y93mgZCvsXqsX/KofNi+Ar/8JKz6qK/hjMuGYv+tN+VN7g8ms39lP7a131NeGj+2e5PtuDMndGJK7McIl99uG3sav1/3KZX0vIzoiGtCb+8/dOJcHvn+A7q9055rZ1zTZBH38MeP3On1FwQpyynKC0z9Y9QEANw26qdH1zCYzo7qPAuC7zd8Fp89cPRNVUzmr81l0Te7aaO4X9bwI0J+zD6j6M9Dri9azqUR/Jvz2oQ0fNzAppkanN6drj7620ccCx/YeS5ozDZ/q4+tNX4fMm3z+ZLRHtIP6b39UVlfy4aoPg/u3L0193wNqgM0lmymsKsSkmGgf256jUo/CGeGksKqQ9/56j1/zfm10m4qi8N7575HuTOeHbT8w8n8jee3317CZbXww+gOiIqL267OEs3A5z+zpgO/0P/LIIwwePJiuXbvSu3dv/vjjDxRF4f7772+w7KxZsxg+fHiz7Kg4eGazmbi4OKN3o81p67mXV5eTX5FPQNP/CPAGvHy8/mNWFurPDJpNZi7ocgEDUgcE14m0RJIVnYXVZD3o991r7n6v3hu/uwTQQA3A9iWw9gt9GL5azjToNRIyB+pD8SkmiEzUp5ubdaTTsNHWv+9GkdyNIbkbI5xyH5gxkOkXTiegBlhZsJKlO5Yyf+t8vtzwJcXuYiYvm8zuyt18cdkXDdbtk9Kn0W12T+yOxWTBr/pZU7iG7NhsKqsr2Vi8EdCbvD/xwxONrrurchcAOeV1Fwv+2vUXAEvylnDuJ+c2ul7tI3tuv5sidxEpUSmsLVwLQKQ1ko7xHRtdr3dy70anN5emMjKbzPRI6sHOip3BZvZHykerP8JV7cJqsoa0fmhKU9/3iz68iNnrZjMwfSDzxs2jV3IvQO+474UlL3DPt/dw6pRT+fFvPzIoY1CD9ZOjkpl24TROn3o6czfOBeDp05+mf1r/Q/p84SKczjP1HfBfr3379uX777/nySefZPPmzRxzzDHcfffdDBw4MGS5BQsWEBkZyZgxY5rYkjhSNE3D5/NhtVpRpIOxI6at5q5pGgVVBRR6CoPTCt2FTFs9jYIqvXfgeHs8l/e8nAxnRnCZBFsCaVFph5xVo7n7q6FiJ1QVAxpoKuT+Dms+h4pddStHJkHPcyF7qH5XH6Wm2E8FizyqtDdt9ftuNMndGJK7McIxd7PJzFFpR3FU2lFcc/Q1lHnKuGb2NcxaO4svN3zJktwlIR2ygf4oQFPbSnQksqtyF66aFmulntLg/N/zf9/n/tTvHb7Eo4+ys71sO9vLtu/3uq5q/b1TolKaXLZ+s/rDoamM6s9z1X+M7wh45893ADiv+3nBPhX2prHv+5frv2T2utlYTBY+GvNRyEUVq9nK/x3/f6wpXMO7y97lofkPMffyuY1u++i0o0mKTKKgsoBIa+R+XYRoK8LxPAMHUfQDHHfccXz55Zd7Xebkk09mxYoVB7VTonn5/X6Ki4tJTEzEaj34O6jiwLTF3P2qn1xXLpX+uiaJq4tW89G6j/AGvAB0ievCJT3qeqw1YSI9Kp04e1zz7EP93E3oRX1lIXqxr8GOZbD6MyjPq1vJEQ89zoH2x+md8wE4EvQm/BZbs+xXuGuL3/eWQHI3huRujLaQe6w9lndHvcvsdbNRNbXRon9X5S7axbZrsG5ADVDk1lutRdv0xwbqN3HffPvmJu+8N6Z23QdPeJBbe9+637nXPrKwt2EAd9W/4L6H2kJL0xpvNr8/Pe/XtlzY27zajGodzo78NhRtYNF2vbf9/WnaD41/32uHO+yW2K3JY3lWl7N4d9m7/Jb3W5PbvvazaymoLMCkmKjyVXH959fz8cUf79d+hbtwPc9IO1UhRLNw+93kuHLwqfqwdqqm8u22b1mQsyC4zMnZJ3Na+9OCnctYTVayo7NxWBzNuzNqAMp3QHWJfldf02DXKn34vdJtdcvZoqH7COg4DMw1J3Z7HESng9XevPskhBBC7IdYeyzJkcnsqtxFdaC6wfxVBasaLfrXFa3Dr/oB6JnUM7it7JhscspzWL5r+QEV/X1T+vLp2k9ZUbACDqA1fo+kHoB+539r6VY6xHVo+Bl2r2py/Sir/lx5U4X7+qL1+9yHVQWNbz+gBlhXqHcgWJtR/e0ero78au/yZ0ZncmaXMw96O+Xe8n0vVMPj9zQ6/bXfXmPW2lnE2eOYfclsRr0/ik/WfMIbv7/RZL8PovULrx4KhBCGKPGUsKVsS7Dgr/JVMWXVlGDBH2GO4PKel3NGhzOCBb/T6qRTbKfmLfhVFSoKoDQHqnbrBf/udfDDM/DzS3UFf0QU9L4AznwSupymF/y2GEjqDgkdpeAXQghxWBRWFaJq6l6XWVe4LniXvFtitwbzX/zlxUbXe3GJPr1vSt+QIQBre+OfuGRisLO9/TGm1xgUFOZunMu6kv3vab9bYjc6xXcC4OVfXm4wX9M0Xv614fRatSMXLM5Z3GCeX/Xz1tK39rkPb//5dqMtAj5c9SE7KnZgNVk5vfPpIfMOV0d+ATXAlOVTALi6/9WHNAxx96TugH6BYkvJlkaXqX1Ov/biS30rdq3grq/vAuCtkW9xUvuTeGuknued8+5kZcHKg9430bJJ0S+EOGiappFfkU9+ZX6wM5/8inxeXfYq60v0K/HJkcnc0v8WeifV3SZIsifRPqZ9yFjAh7gjehP+gtX6s/sEoHgz/Pg8LHoOivROjLDYoce5erHffYT+OiIaErtCYmeIiGye/RFCCCEa8f7K9+n9Wm9eXPIiueW5IfM0TWPexnmMen8UGhrtY9tzZueGd4W/3/I9jy98PHhXX9M03vrjreBwcA+e+GDI8vcefy8Z0Rn8sO0HLvrwIjaXbG7wvr/l/cb4r8aHNAnvm9qX6wZch0/1cemXl/Llhi8bNLnPd+Xz2m+vMeHHCcFpiqJw3/H3AfDyry8zY8WM4DyP38PNX97MpuJNTWZ0XrfzAH14u/lb5genl3vLuf7z64MdE+6Ny+visk8uo8RdEpz2c87PjJ83HtCb2Kc50/a5nebw1cavyHflo6BwTf9r9rl8bnkuHV7oQNdXu/LHrj9C5o3pNQaHxYFf9TPmozGs2V3XGaEv4OOZn55h8rLJgH6Bob4qXxVjZ47F4/dww4AbGN1rNACje43m+gHX4/a7uWTmJbh97kP7wKJFkub9bYDJZMJms4Xd0BMtXbjn7gv4yKnIwe2v++Xw564/+WTjJ8E7Cb0TezO6+2hsZv25eLNiJsOZQUxETPPtiLsUXDugphmbqTyHuOWzsBStrlvGHAGdT4GuZ+hN+gGsURCTXvdaHJJw/763VJK7MSR3Y4RD7goKawvXMn7eeMbPG0+aM43M6Ex8qo+cspxg53npznRmjZ2Fw9qwNdxzZzzH7V/dzou/vEjn+M7klOews2InAH8f9HfG9hkbsnxyVDJfXf4Vo94fxex1s5m9bjad4juRHJlMla+KzSWbg8Pont/j/JB1Xzn7FSqrK5mxcgYXfHQB8fZ4Oid0BvSCP9+VD8BVR10Vst51A65jwbYFzFgxg8s/uZx7v72XdGc664rWUVldydOnPx2847ynK466gkl/TOKXvF84dcqpdIjrQLwjntW7V2O32Hn2jGf5x1f/2GvO/zrlXzz+w+NkTMygd3Jvyr3lbCjeAMDQzKE8ffrTe12/Ob2zTG/aP6zDsGB2e+NX/Wwr01smqiY15PueGZPJO6Pe4apPr+KPHX/Q5/U+tIttR5w9jk3Fm4KdKF7c+2JuHnRzyHb/MfcfrClcQ6/kXrxw1gsh81446wV+3P4jq3av4o55d/DGuW8cykdu1cLhPNMYKfrbALPZTHx8vNG70eaEc+6VvkpyXbn4Nf0ug1/1M3fLXBbn603xFBTO7HgmJ2aeGOyQx2aykR2THbwAcMiqK6E8H6or9NdVRbB6NuacXzDX3okwWaDjSfpdfXusPs3i0Iv92teiWYTz970lk9yNIbkbIxxyv3HQjfRL7cc3m79h4baFrC9az/Jdy1EUhXh7PMM7Duecrudw/YDrG3Q0V+u2obfRPq49ExdPZNnOZfhUH0Myh3Dr4Fu54qgrGl2nb2pfVty8greWvsWstbNYWbCSbaXbiLRG0jmhMye2O5Hze5zPCe1OCFkvwhzB9Iumc83R1/D20rdZnLuYFbv0jrrTnGmc3+N8RnYbyXndzwtZT1EUpl4wlZPancSkPyaxpnANldWVHJt1LA+c+ADtYts1WfRbTBa+vuJrHl/4ODNXzyS3PBe3383Y3mN59ORH2Vq6dZ85D80ayi/X/cJjCx/jh20/UOoppVtiNy7rcxn3HH9PoxdTDofCqkI+X/c5AH/r/7cDXj86Ohqz2Rwy7ZI+l3BU6lG8+MuLLNi6gO1l28krzyMxMpGT2p/EVUddxZjeoaOnfbjqQ97+823sFjvvX/R+g88faY3k/dHvM/TtoUz6YxJndD6DC3teeMD7Gw7C4TzTGEVrqmtMsU+rVq2iT58+rFy5kt69D+94o4dC0zT8fj8WiyWshp5o6cI190J3IQVVBcHm/OXV5fxvzf/YVq5flXZYHFza41K6xHcJrhNjjSHDmYHZZG50mwfE5wFXPnjKal5XwbqvYNN3EPChAZpiQml/PEqPcyAyQV/OYtd743eE34m8JQjX73tLJ7kbQ3I3xuHI/bEFj/HZus8aHc98b5bvWs6A9AG8es6rzbIf+7K1dCsdX9Q74dvXM+TNrTV93zu80IFtZduYf9V8Tu5wstG7c0haU+7hpLXkfqB1qNzpbwP8fj9FRUVhN/RESxduuauaSl5FHuXVdT3Hbivfxow1M4LNyTKcGVze83Li7XWFdYojheTIfY9Hu08BH7h26nf00UD1w+aFsPbLurv9gJp+NEXtRhCf1g6r2QRmW12x34JP3q1duH3fWwvJ3RiSuzEkd2NI7saQ3I0RrrlL0S+E2CdvwEtOeQ5e1QvoV0GX7FjCF5u/CHbqMzB1IKO6jAp2zmdRLGQ6M0PGCD4oqgoVu6CyoG74vfw/YdUnek/9tRI6Qd/RqHGdUF0eMFkhNgMiE6XYF0IIIYQQbZYU/UKIvSqvLie/Ip+ApnfO5wv4+HTjp/xZ8CcAJsXEeZ3PY3Da4GAzKIfZQVZ0FhHmiIN/Y03T7+q7dkLNUIAUbYIVM6F+r79Ryfrwe5kD9eJeM4EjEZI7QcQhvL8QQgghhBBhQIp+IUSTdlXuotBTGHxd7Clm2upp7KzUewmOiYjhsp6X0S6mXXCZ2IhYMpwZhzQO7Z498lNRAKtmQV69oWsiovTh9zoN0zvsQ9EvADiSoLhE7u4LIYQQQgiBFP1tgqIoWK3WFt0ZRThqzbn7VT95FXlU+OqelV9fsp4P1n4QHKKvQ2wHLutxWbD5voJCamQqiY7Eg39jb4Ve7Nc+o+916c/sb1kINcMAYrZC5+HQbQREROrT7HEQkwEWG4rf32pzb81a8/e9NZPcjSG5G6Mt594hrsMR78CvVmvKfev4rUbvQrNpTbmHk3DNXYr+NsBisZCYeAiFmDgorTV3t99NjisHX02TelVTWZizkG+2fRNc5vjM4zmrw1nB3vgtioXs6GwirZEH96Z79sgf8MGm72HdHPC565bLHgq9z9ef0wewRkFspn7Xv0Zrzb21k9yNIbkbQ3I3huRuDMndGJK7McI1dyn62wBN01BVFZPJFHZXrVqy1ph7qaeUHZU7UFEB8Pg9fLT+I9YUrQHAarJyYbcLOSr5qOA6DouD7OhsrKaD6OE04NPv7FcVA5reUV/Or7D605ppNZK7Q5/REN9ef222QUx6o8Pvtcbcw4HkbgzJ3RiSuzEkd2NI7saQ3I0RrrlL0d8GhOvQEy1da8pd0zR2Vu6k2FtXaO+q3MW0NdMochcBkGhP5PJel5MWlRZcJsGWQFpU2oGfFPfskR+gYC2snAml2+uWi06HvhdBal/9GX3FrA+/F5Xc5DP7rSn3cCK5G0NyN4bkbgzJ3RiSuzEkd2OEa+5S9AvRxvkCPnIqcoLP6gOsKFzBzPUz8QX0Jv7dE7pzcfeLcVgcAJgwkR6VTpw97sDeLNgj/w5Q/fq08nxY+THsXFG3nC0Gep0H7Y8Hkxm9k74kcKaBWU5bQgghhBBC7C/561mINqzSV0muKxe/phfgqqYyb+s8FuUuCi5zartTOaXdKcHe+K0mK9nR2cELAPvNXQLlOyDg1V97SmH157Dtp7q7/eYI6HYmdD0dLHZ9mj0WojPAaj+UjyqEEEIIIUSbdAhjagkhWrNCdyHbyrcFC/6K6greXflusOC3W+xc2ftKTm1/arDgj7JE0Sm2034V/DfccAPp6enExMTQt1d3Pv/gPQh4mT3na3occwaxvU8n/fJXuPPjbQQ0BTqcAGc8wQfrI+h28iXEdD+RQSOv5ee1Oxst+BctWoTT6Qz+FxUVhaIoLF26tBlTEkIIIYQQonWTor8NUBQFi8USVp1RtAYtNXdVU8lx5bCrahca+vA/ua5cXl32KptKNwGQGpXKLf1voUdCj+B6ifZE2se0x2LavwZCd952C1uXfk/52oW88+xDjLv9nxQt+4pBJZ+z6OZMyp45mtUP9uav3SbeKBgIA65kp8vPVXc8wusvv0BZuYsbbryJ0aNHN7r9E088kYqKiuB/77zzDh06dGDAgAEtMvdw11K/7+FOcjeG5G4Myd0YkrsxJHdjhGvuUvS3ARaLhaSkJCwWeZrjSGqJuXsDXjaXbqa8ujw47fedvzNp+STKvPpweUclH8XNR91MokMfrsSEiSxn1v532BfwQel2eiRq2DQPaBpK8UaqPW7yFk0j0+ElOdoKce3gmL9jim/Hpl0uUEzklaskJiZx6jkXoCgK48aNY8eOHZSUlOzzbadOncq4ceOwWq0tLve2oCV+39sCyd0YkrsxJHdjSO7GkNyNEa65h9enEY3SNC3473C7atWStbTcy6vLya/IJ6AFAPCrfj7f9Dm/7fwNAAWFszudzXEZxwX312aykR2Tjc1s2/cbqAGoKAjpkf/vdz7Au598g8cX4OxesfTNcIAjgR/9fTjn75Mod31CcmI8z094AlJ60T+lD126dGHevHmcdtppTJ48mUGDBhEf33BovvoKCgqYN28eEydObHG5txWSuzEkd2NI7saQ3I0huRtDcjdGuOYuRX8bEK5DT7R0LSn3XZW7KPQUBl+XecuYvmY6ua5cAKKsUVzW8zI6xnYMLhNjjSHDmYHZZN77xjUNKguhYmddj/xVRbB6Nq+dUMjLx/VnwQYXK3f5UfpeBJ2Hc4I5grKR49iys4yps78jpUs/MFsxA5dffjkXXnghXq+XuLg4vvvuu31+vvfff5+BAwfSrVs3fD5fi8m9LWlJ3/e2RHI3huRuDMndGJK7MSR3Y4Rr7tK8X4gw5lf9bCvfFlLwby7dzCt/vhIs+LOjs7n16FtDCv4URwrZMdn7LvjdJVCwBspz9YK/ukoffu+bh2H7EgDMZgunnn0+3xalMifHqffQb3FAQmc6DjiZ3v36c+uttwLw9ddf8/DDD/PLL7/g9Xp59dVXOffcc6mqqtrrbkydOpUrr7zyYCISQgghhBAirMmdfiHClNvvJseVg0/1AXpzpZ/yf2Lu5rnBDvyGpg/lnE7nBDvnMytmMp2ZREdE733j3goozwdfpf5a9cPmhbD2S6iuqFsuYwD0uQCcqfi1H9m4LR9i20FUYnARv9/Pxo0bAfjrr7845ZRT6NOnDwBjx47llltuYd26dRx99NGN7sratWtZvnw5Y8eOPeCMhBBCCCGECHdyp1+IMFTqKWVr2dZgwV8dqOb9de8zZ/McNDTMJjMXdbuIUV1GBQt+u9lOp9hOey/4fR4o2gRFG/SCX9Mg7w/49lFY/kGw4C+zZzOj4jgq+l6J357IR198y/yf/+Cks8fw4ZffsX37dgA2bNjAv//9b4YPHw7AwIEDWbBgAWvXrkXTND7++GM8Hg+dOnVqcpemTp3K2WefTWJiYpPLCCGEEEII0VbJnf42QFEUTCZTWHVG0RoYkbumaeys3Emxtzg4rdBdyLTV0yioKgAg1hbLuJ7jyIzODC4TGxFLhjMDk9LEdcCAD1w7oKoYaloJULQJVsyE4k11y0UlQ58LUaK78dbf7uLv/34HTVPo0rULM2bMoP+AAXz+5ZfceeedlJSUkJiYyJgxY3jiiScAGD58OHfffTdnnXUWRUVFdOzYkQ8++IDY2FgARowYwYknnsgDDzwQ/LzTp09n4sSJwV2Q77sxJHdjSO7GkNyNIbkbQ3I3huRujHDNXdHqd1EoDsiqVavo06cPK1eupHfv3kbvjmjjfKqPHFcObr87OG1N0Ro+XPch3oAXgM5xnbmkxyVEWaMAvcf+1MjU4PB8DagBqNgFlbuDPfJTUQArP4H8pXXLRTihxznQaRjUtBwgIhpiMiAistk/qxBCCNEcHlvwGJ+t+4xBGYMOaL3lu5YzIH0Ar57z6mHaMyGEaNqB1qFyp1+IMFDpqyTXlYtf03vPVzWV77d/z/fbvw8uc1LWSZzR4Yzg3XyLYiErOit4ASBEYz3ye136M/ubF0LNsH+YrdB5OHQbUVfcW+x6sW+PPWyfVwghhBBCCLF/pOhvA2QIM2McqdwL3YUUVBUEO+er8lXx4boPWV+yHoAIcwSju42mT1Kf4DoOi4Ps6Gyspkb2y10C5TugpnUAgWrY9D2smwu+ulYEtDsGeo2CyJpWAiYLRKfrrw1sEiXfd2NI7saQ3I0huRtDcjeG5G4Myd0Y4Zq7FP1CtFKqppJfkU9ZdVlw2o6KHUxbM40STwkASY4kxvUaR0pkSnCZeFs86VHpDZ9V8rr0Yr+2R35NhZxfYdWn4K7rI4DkHtD3Iohrr79WTPqz/M5U2NcQf0IIIYQQQogjSop+IVqh6kA1Oa4cPAFPcNqygmV8suET/DXN8Xsl9mJ0t9HYLXZAf34/PSqdeHt86MZ8bn34PW953bSCNbDyYyjdXjctJgP6XASpferu5DviIToDLBGH5XMKIYQQQgghDo0U/UK0Mq5qF3kVeQRqnqsPqAHmbpnLz/k/A3pxf3qH0xmWNSx4N99qspLlzCLSWq9TvYBPL/br38Uvz9eL/Z0r6qbZY6HnedD+uLo7+RHOmk76GukPQAghhBBCCNFiSNEvRCtSUFXAbvfu4GtXtYv/rfkfW8u3Avqz+pf0uISu8V2Dy0RZosiKzsJS26t+Yz3ye0ph9Wew7ee6aeYI6HYmdD1d75wPwGzTi31H3OH9oEIIIYQQQohmIUV/G2C1WklLSzN6N9qc5sw9oAbIrcilwlcRnLatfBsz1szAVe0CIMOZwWU9LyPBnhBcJtGeSGpkqn7Hv7Ee+f0e2PANrJ+nd9gH+jP67Y+HXiPBHqdPM1nAmQZRSYZ20rc/5PtuDMndGJK7MSR3Y0juxpDcjSG5GyNcc5eiX4gWzu13k+vKpVrVi3JN0/hlxy98sfkL1Jq78gNSBzCq8yisZr2XURMmMpwZxNpqhs2rKgbXzroe+dUAbPsJ1nwOnrqOAEnrqz+3H5NRM0HRO+mLTpNO+oQQQgghhGiFpOhvA/x+P8XFxSQkJGCxyCE/Upoj91JPKTsqd6CiF/e+gI/Zm2azdNdSAEyKiXM7ncvQ9KHB5/cjTBFkR2frHfh5Xfpz+r4qfYOaBrtWwMpP9Om14tpBn9GQ0qNumj1OL/4ttoPad6PI990YkrsxJHdjSO7GkNyNIbkbQ3I3RrjmHj6fRDRJ0zRUVUXTNKN3pU05lNw1TWNn5U6KvXWd7BV7ipmxZgb5FXqxHh0RzWU9L6N9TPvgMtHWaDKdmZgD1VC0KbRH/tJtsGIm7F5XNy0yAXpdANmD9Wb9ANYovdi3OQ94v1sC+b4bQ3I3huRuDMndGJK7MSR3Y0juxgjX3KXoF6KF8ak+cl25VPmrgtM2lGzg/bXv4/a7AegQ04FLe15KdER0cJlkRzIpEXFQlhvaI39VEaz6FHJ+qZtmdUD3s6HzcKh5JABzBESn6xcChBBCCCGEEGFBin4hWpAqXxU5rhz8mt7RnqZpLMxdyDdbv0FDv+J4XMZxjOg4AnPNM/ZmxUxmZBrR1VVQtqau9/3qKlg/FzZ9rw/PB6CYofPJesFvi66b5kwFZ0qL76RPCCGEEEIIcWCk6BeihShyF7GralewuPf4PXy8/mNWFa0CwGKycGHXC+mf0j+4jt1kI9tkJ6I0p65HftUPmxfC2i+gurLuDTIHQu8L9OIeAAUiE/W7+2Y5FQghhBBCCBGO5C/9NsBisZCammr0brQ5+5u7qqnkV+RTVl3Xi35BVQHTVk+j0F0IQLw9nnE9x5HuTA8uExPQyPS5Man6kH1oGuT9AatmQeXuujdI6Ax9R0Ni57ppthiIyQSr/dA+ZAsk33djSO7GkNyNIbkbQ3I3huRuDMndGOGauxT9bYAiTbYNsT+5VweqyXHl4Al4gtNWFq5k5vqZVAf0Ifq6xXfj4u4XE2mN1Ldb7SYlECDJFFG3oaJNeid9xZvqpjlToPeFkHF0XbN9a2RNJ311fQGEG/m+G0NyN4bkbgzJ3RiSuzEkd2NI7sYI19yl6G8D/H4/paWlxMXFhdXQEy3dvnJ3VbvIq8gjoAUA/Y7/11u/5ofcH4LLDG83nOHthmNSTOD3YqkqJsscSZTFoS9QsQtWzoL8pXUbjnBCz3Oh40lgqnnfNtRJn3zfjSG5G0NyN4bkbgzJ3RiSuzEkd2OEa+4mo3dAHH6apuH3+8Nu6In94fV6+dvf/ka7du2IiYnhmGOOYfHixQBMnjwZi8WC0+kM/rd9+/YG25gwYQKKovDjjz82+T7nnHMOKSkpxMbGMnToUBYvXrzX3AuqCtju2h4s+Ct9lUxeOTlY8NvMNq7odQWntT8NU0CF8p04ynfQyRqrF/xeFyz7H3zzaF3Bb7ZCt7PgzCf0XvlNFn0Yvuh0SO7ZJgp+aNvfdyNJ7saQ3I0huRtDcjeG5G4Myd0Y4Zp7+Fy+EKIRfr+fDh068OOPP5KVlcWHH37IyJEj2bp1KwAnn3wy3377bZPr5+Xl8b///Y/09PQmlwF4+umn6d69OxaLhc8++4xRo0aRm5vbYLmAGiCvIg+Xz1X3HhV5TF89nVJvKQApkSmM6zWOJFsCVOwGTynxlijSI9NRVB+smwvrvwKfPnwfigLZQ6HXKL1jPn2iXuRHp9cNySeEEEIIIYRoc6ToF2EtKiqKhx9+OPj6kksu4c4772TdunX7tf5dd93Fo48+yh133LHX5Xr37g3oVwfNZjO7d++mvLw8ZBmP30OOK4dqtTo47Y9df/Dpxk8JqPod/77Jfbmwy4XYqqugeDOKppJuTyTeGgU5S2DVbHAX1200pSf0uQji2tVNs8Xoz+1bHfv1GYUQQgghhBDhS5r3izZlw4YNFBcX06VLFwCWLFlCYmIivXr14o033ghZdsGCBRQWFnLBBRfs17bPPfdc7HY75557LjfddBOxsbHBeWXeMraUbQkW/H7Vz+yNs/l4/ccE1AAKCmd3PJtL2p+NrSwPKguwAB2i0okvzYX5T8Hv79YV/DEZcNztcPz4uoLf4tB76k/sLAW/EEIIIYQQApA7/W2CxWIhOTkZk6ltX+Nxu92MGzeO+++/n9jYWIYNG8bKlStp164dv/32GxdccAHJyclcdNFF+P1+7rjjDqZOnbrf2//iiy+orq7miy++oLKyEovFQlJSErs9uyn21t2dL/OWMWPNDHJcOQBEWaO4tPP5dLI4oWKnPs1sJ8vnx7JkEuxaWfcm9ljodR60Ow5MZn2ayVrXSV+Y9jh6IOT7bgzJ3RiSuzEkd2NI7saQ3I0huRsjXHOXor8NUBQFs9ls9G4YyufzMWbMGLp06RJs7t+xY8fg/KFDh3L77bfzySefcNFFF/Hqq69ywgkn0KdPnwN6n4iICC644AL69etH596dSeuchtvvDs7fUraFGWtmUOmrBCArKp3Ls04jFhP49WH7klSNlA1fo2xfDJqqr2ixQdczoevp+r9B76QvKgWcqRBmJ6ZDId93Y0juxpDcjSG5G0NyN4bkbgzJ3RjhmrsU/W2A3++nrKyM2NjYsBp6Yn+pqsoVV1yBoii89957TY6/aTKZgj11zp8/nx9++IGPPvoIgN27dzNq1CgmTJjA9ddfH7KeN+DF7XPjCXio8lfh9Xup9Fby+8rfOS7+OGxRNhSTws/5PzNn8xw0NNA0Bif0ZGTyQCyKXrCb/dW0z1uOY/NCCNQ896+YoMMJ0HOkfpe/liNBb+IvnfQ10Na/70aR3I0huRtDcjeG5G4Myd0YkrsxwjX38PkkokmapuHz+cJu6In9deONN7Jjxw7mzZsX8sP71VdfMXDgQJKTk1m6dCkvvfQSzz77LKAP5+fxeILLDh48mFdffZVhpwyjzFuG2+/G4/fgCXjYtnUbm9Zt4piTjkFRFD6Y/AG78nfR9+i+qD6Van81s7fM5q/df4GmYUZjVOqxDErooW9cVYnfsZzUTYswV1fU7XhaP+hzoV7c14qI1l9HRB7WzFqztv59N4rkbgzJ3RiSuzEkd2NI7saQ3I0RrrlL0S/C2rZt23j77bex2+0kJSUFp8+dO5dvvvmGK6+8ksrKSjIzM7n33nu55JJLAIiLi8On+oLFvWbSqLBUkFudC9Xw2F2PAfDIc4+gaRpvTnyTe264B7PZTJeeXXh1xqskpSSRV5DHx1s+psBdAKqfWLOdy7NPJysyBTSNyN0bSN34PY6qEoLtD+LaQd/RkNyj7oNY7HqxX/9uvxBCCCGEEELsgxT9Iqy1b9++ySt1J554Is899xyg96bv8XvYXbUbj19vpu/X/MFl5y2dF7LuI889Evx3dodspn81Pfha1VSKPcX8tus35myaQ7XmRlEDdIpK55Ls03BaHESU5ZO4/muiS3OxKjXPDUUmQK8LIHuw3qwfwGSp6aQvUTrpE0IIIYQQQhwwKfpFm6NqKm6/O/ifx+8JDqV3oDRNo9RbSl5FHrmuXHIrcslz5eH1e9DUAGoggEnRODH5KM5MHUKEp4yE1XNw7lxJhGLBrJj14fW6nw2dh9d7Rl8BZ20nfeHXmYgQQgghhBDiyJCivw2wWCwkJiaGVWcU+0vVVDx+j17cBzy4fW68qvegt1dRXUFuRa5e4LtyyavIC/bEr79hALSA3uu+BlEWG+dnnEj/qAziNnxH7LZfMGkqESYrJpMFOp0CPc6GCGfdNhzxEJ0BlohD+ORtV1v+vhtJcjeG5G4Myd0YkrsxJHdjSO7GCNfcw+vTiEYpioLVGv69vGuahifgqSvyazra0zi4jjjcfjd5rrzg3fvcilzKvGWNvLEKagBFU0m1x5PpSCbbkUKmI5lUawyJuUuJWzoTs8+DWTFhNVlRMgdC7wv0u/m1Ipw1nfRFHWQCAtrO972lkdyNIbkbQ3I3huRuDMndGJK7McI1dyn624BAIEB5eTkxMTFhNe5kY0PlqagHta3qQDX5FfkhBX6Ru6jpFTSNxIgYsuyJZNkSyIpMJt2eSISp5iSh+onasYq4Df8jwqt30mcxmbEmddc76UvoVLctsw1i0vU7/OKQhev3vaWT3I0huRtDcjeG5G4Myd0YkrsxwjV3KfrbAFVV8Xq9qKraar+81YHqkGfwPQEPAS1wUNvyq352Vu4MNs/PrciloLJgry0CYm2xZDkzybInkWWNIdPqxG62NVjO7CkjJucPYnKXYqquRFP1PvmsMRlY+oyGjKPrOuRTzBCdBlHJ0klfMwqH73trJLkbQ3I3huRuDMndGJK7MSR3Y4Rr7lL0ixan/lB5tf9bvyf9A6FqKgVVBXUFviuXnZU793rBIMoaRVZ0FpnOTLKd2WTaYnEGVPBV6k3596RpOIq3EJPzG5EF61BqRgvQANXqxNpnFJYup+g98QOgQFQSONPALD+CQgghhBBCiMNHKg5hqNqh8mqL+z2HyjsQmqZR5CkKeQ4/rzIPX8DX5Dp2i51MZyaZzkyyorPIis4iNiIWxe8Fbzl4ysHdyHP8gMnnwZn/F7E5v2Gt1B8FUBQFk2JCje+A1mEYFXH9scdFg6lmCD57LMRkgqVhKwEhhBBCCCGEaG5S9IsjJqAG9B70m2movPLq8uAwebV38j1+T5PrWE1WMpwZwbv4WdFZJNgTMCk1BbnfD94yqNwGgaZ7+I9w7SJm+29E71yBJeDXC32TFZPZipI9FDqdDPHt8QVUFFfN/lij9E76bM4mtyuEEEIIIYQQzU2K/jbAYrGQkJBwRIeeaO6h8ip9lcFh8nIr9AK/orqiyeVNion0qPSQAj8lMqWuwA/uqAqeMvC69Ob7TbBoGvG7NxK9/VciSraiKCZMoDfZj0qGTsOg/fEhPe9bTAoJsdFY4jIhKvGgP7s4MEZ834XkbhTJ3RiSuzEkd2NI7saQ3I0RrrmH16dpwbxeLzfffDPffvstpaWl9OrVi+eff55jjz2WyZMn89JLL7Fx40bi4+O5+eabue+++4Lr/vbbb1x77bVs3LiRwYMHM2XKFNq3b9/gPQoKCrj99ttZuHAhbrebQYMG8fLLL9OzZ08iIg7fmO97DpXn9rvxBrwHPVSex+8hryKPvIo8clw55FfkU+IpaXJ5BYXkyGSyo7ODBX5aVBoWUxNfbw2ortCb7jfynL5ZMWE3ReAw23BUu4nK+RXztp/1iwOg98ynKJDaV7+rn9pLnxayERuKM4UIR0Jd035xRCiKcli/76JxkrsxJHdjSO7GkNyNIbkbQ3I3RrjmLkX/EeL3++nQoQM//vgjWVlZfPjhh4wcOZKtW7fi8Xh45ZVXGDx4MHl5eZx11lm0a9eOyy67DK/Xy4UXXsgjjzzCuHHj+Ne//sW4ceNYtGhRg/eoqKjgmGOO4cUXXyQpKYmJEycyatQo1qxZg8vlIjo6ull6oawdKs/td+MOuA9pqDxfwEd+ZX7wOfxcVy6F7sK9rpNoTyQzuuYZfGcW6c50bI30pN/wzTx1z+nXdORnQsFmthFptuEw27CbIrCZrLB7HWxeADuWhV4UiHBCh+Oh4zC9M749WRzgTAFHPAFVxVVe3my5i/0TCASa9fsu9o/kbgzJ3RiSuzEkd2NI7saQ3I0RrrlL0X+EREVF8fDDDwdfX3LJJdx5552sW7eOm266KTi9Q4cOXHjhhSxevJjLLruMBQsWYLPZuO666wB48MEHef7559myZQsdO3YMeY9OnToxfvz44Ovbb7+de++9l927dwf34UC/vPWHyqu9g3+wQ+UF1AA7q3aGFPi7KnfttUVATERMsIO9LGcWGc4MIq2R+/+mfr9e6HvLUQLV2E1W7JZIvcA3R2A3RaDUDpfnc8PWn/Ri37UjdDvxHfUm/FmDwWxt+D4RTr3Yt8cGJ6mqisfjOajcxcGT3I0huRtDcjeG5G4Myd0YkrsxJHdjhGvuUvQbZMOGDRQXF9OlS5cG83744QfGjRsHwOrVq+nXr19wXmRkJJ07d2bVqlUNiv49LVq0iNTUVBITEykqKtrnPvkCPtwBN26fO9hc/1CGyttdtTs4TF5uRS47KncQUJu+YBBpjSTLmRXyHH50RPRBvLkKXhc2nweH6sdutuGwJ2I3WRs+0w9QngebFkDOEvDX63fAbNWL/E4nQ3yHxt/LFgPOVOmgTwghhBBCCNEiSdFvALfbzbhx47j//vuJjY0NmTdx4kSKi4u56qqrAL3JfkxMTMgyMTExVFQ03YkdQGFhITfeeCMTJkxodL5f9Qd70K+9i38oQ+WVeErIqcgJ3sXPr8inOtB0z/w2s00fKq9eM/04W1zdXfcDZFUsONQADp8Hh78au8mG2eZoegXVD/nL9Lv6hetD50Um6Xf1Oxyv38FvQAFHnF7sW/fyHkIIIYQQQghhMCn6jzCfz8eYMWPo0qVLSHN/gOnTp/P888/zww8/4HDoxaTT6aS8vDxkufLycpzOpu8su1wuRowYwdixY7nqqqvw+Xx4/V6KPEX4PD7cfjc+temx6/el3FseMkxerisXt9/d5PJmk1kv8J2ZwTv5iY7Exu+67weLYsFhceCwOLCrGg6fG4u3XC/kUWBvz/e7S2Drj7Dlh7qO+aCmY74+NR3z9W7YMR/o0xwJejN+y370ISCEEEIIIYQQBpOi/whSVZUrrrgCRVF47733Qu5qz549m7vuuovvvvsupNl+r169eO2114Kv3W43mzZtonfv3o2+h9vt5txzz2XgwIE89dRTAJjNZrw2Ly6v64DvpFf5qhoU+K5qV5PLK4pCelR6sHl+pjOTtKi0gy7wzYpZL+7NdhxWBw6zA6um6cV7RSH4PfveiKbpd/M3L4D8P/fomC9KH2qv0zB96L1GP5RZ77QvKrnx5/mb2nezmbi4uLB6Hqg1kNyNIbkbQ3I3huRuDMndGJK7MSR3Y4Rr7lL0H0E33ngjO3bsYN68eSFjP3733Xdce+21zJkzp0Exf/LJJ+N2u3nnnXe4/PLLefLJJxk4cGCjz/P7fD4uuugiMjIyQi4UmEwmrBFWfP693933BrzkV+ST48oJFvh7GyoPICUyJVjg1/ak3+RQeftgwoTdYg/exXdYHESYa4bMUAP6nfnybbCXiw4h/B68G3/k5kdf5tuVuyh1++mV5uD5C7M5dkDNXf2sQfg1E0efeRluj4eNP30GwKJfljJi3G16CwAUNE2jqqqK33//nYEDBzb5losXL+b444/n8ccf55///Cd2u/2gshAHz2QySe4GkNyNIbkbQ3I3huRuDMndGJK7McI1dyn6j5Bt27bx9ttvY7fbSUqqG+pt7ty5PPHEE5SWljJ8+PDg9HHjxvHGG29gs9mYNWsW1157LbfccguDBw9m2rRpweVqe/5/4403+Pnnn5k7dy4OhyOkH4AVK1bgifCgRqiYasaM96t+dlTsCLmLX1BVsNfPEG+PD969z47OJsOZsX9D5TVCQdHv3lscwULfZraFtkTQNL3Qd5fo/6vt57CA5fn6Xf3tS/BXVtIhOsCP47uTlRjFh1tiGfnfn9l609s4o/RRAF55azqx0U7cnppWA2YbJ54xigrXNVCT1wcffMB9993HgAEDmnxbVVW54447GDx4MKAP+VFZWRl2vX+2dJK7MSR3Y0juxpDcjSG5G0NyN4bkboxwzV2K/iOkffv2aFrjQ9PNnz9/r+sOHjyY5cuXNzrvjTfeCP572LBhjb6Hx+thwYYF5Kv55Fflk+vKZWfVzib3ByA6Ijp49z4zWn8W/4CGytuD3WwPFve1xX6TjxpUV4G7GNylsL99D6h+2PGX3gt/4brg5CibmYcv6qc3329/HJfYorlz2hms27SVgf16sWt3EW9O/4Rn/jmefzzyHMS1B0d8zR3+OlOnTmXcuHF7fTzizTffZOjQoZSV6X0FqKpKVVUVDocjrE4aLZ3kbgzJ3RiSuzEkd2NI7saQ3I0huRsjXHOXoj/M7azcyd0L7qbcUw4mGi1aHRZHyB38TGcmMbaYRra2fyJMEcECP9ISid1i3/cz/f5q/Y6+u3j/ntOv5SmFLYtg6yL9IkF9tR3zpfUJdsy3YfN2ikvL6dIhG4B7n3yJB+74O1FpnfTn9SMTGrxFQUEB8+bNY+LEiU3uRlFRES+88AJLlixh/Pjx+7//QgghhBBCCHEYSdEf5pIdoZ3TRZgjyHBmBHvRz3JmEW+PP+ih8iyKJVjY1zbV3+9n+muf068q3v/n9EFv9l+0QW/Cn/cnaIG6eRFR0P446DhM72W/Hrfbw7jb/8n9t15DbEw0i5dvZENOAe/edBcLFy5s8u3ef/99Bg4cSLdu3Zpc5sEHH2T8+PHExcXt/+cQQgghhBBCiMNMiv4wZzaZGd11NKVlpWQlZZEanXpIQ+XVL+4dFgdW0/73Zg/oBbvXpd/RP5Dn9EFvAbB9CWxeCOV5ofPi2kGnUyBrUKPD6fl8PsbceA9dOmTz8AP3okalcPtDN/Daa6/t84LH1KlTueaaa5qc/+eff/Lbb7/x6quv7v9nEUIIIYQQQogjQIr+NuC8LuexYfcGfCbfft/R32tP+gejuqqm+X7J/j+nX6t8R03HfItDm/6bLHqR3+lkiO/Y4Dn8WqqqcsXtD6NYbLw34yMUexRlpaUsXbqUkSNH6rtXXU15eTlpaWmsX78+2BHi2rVrWb58OWPHjm1y9xYuXMi6devIzMwEoKysDIvFwsaNG3n99dfD6nmg1sBsNhMbGyu5H2GSuzEkd2NI7saQ3I0huRtDcjdGuOYuRX8bYDKZiLBH4Pf7G52voOgFvtkRvItvtzTDUBUBn950310CfveBrasG9I75Ni+A3WtD50UmQseToMMJYIve+3YUMzc++Cw7SqqY9/XXWGqG4IiNjSUvr661wM8//8xdd93F4sWLiY6u2+bUqVM5++yzSUxMbPItbrjhBi655JLg63/84x907NiR++67D4fDsf+fWTQLk8kkuRtAcjeG5G4Myd0YkrsxJHdjSO7GCNfcpehvAwKBAJ4qD5pFw2QyYTPbQpro2832g36mvwFV1TvXO9Dn9Gt5yvRO+bYs0i8W1JfSCzqfAml9gx3zNclkhahkthVW8vaUGY0OlXjiiScGXyckJGA2m0lLSwtO0zSN6dOnN9qB31NPPcWiRYuYO3cukZGRREbWjWzgcDhwOp1ER0fjcrmIjIwMu6uFLVkgEKCqqkpyP8Ikd2NI7saQ3I0huRtDcjeG5G6McM1div42QFVVbH4bSXFJOO3Og36mv0mH8px+7fpFG/W7+vlL9bv8tayResd8nYaBM3Xf2zLb9A78HAlgMtE+mr0OTVjr5JNPZuPGjSHTFEVh69atjS7/wAMPNLmtyZMnA3o/ApWVldjt9rA6abR0qqpK7gaQ3I0huRtDcjeG5G4Myd0YkrsxwjX3Zq7+ms9vv/3GrbfeSu/evYmKiqJdu3ZcfPHFrF+/vsGya9as4ayzzsLpdJKQkMAVV1zB7t27GyynqipPP/00HTt2xG63069fP/73v/8diY9juEhrJA6Lo3kLfp8byvJg1yoo3qTfmT/Qjvm2/ADf/wt+eAZyf6sr+GOzYcAVMOI/0O/ifRf8FgfEtYeUnhCVBKYW+9UWQgghhBBCiCOmxd7p/89//sNPP/3EmDFj6NevHzt37uSVV15hwIABLFmyhD59+gCQm5vLSSedRGxsLE899RQVFRU8++yzrFixgl9//ZWIiLrO5x588EEmTJjA9ddfz+DBg5k9ezaXXXYZiqKEPJMt9iLg04v7quIDf06/lmtnXcd8vnrbMFkgc6DeMV9CpyY75gsR4dTv7NtjD25fhBBCCCGEECKMtdii/84772TGjBkhRfvYsWPp27cvEyZMYNq0aYD+bHVlZSV//PEH7dq1A2DIkCGcfvrpTJ48mRtuuAGAvLw8nnvuOW655RZeeeUVAK677jqGDRvG//3f/zFmzJiwasLRrGqf03eXgLf8ILcRgJ3L9WK/YE3oPEdCXcd89pj9254tRr/7b3Me3P4IIYQQQgghRBvQYov+4447rsG0rl270rt3b9asqSsaP/74Y84999xgwQ9w2mmn0a1bNz788MNg0T979mx8Ph9///vfg8spisLNN9/MZZddxuLFiznhhBMO4ycyjtlsJjo6+sAuahzqc/q1POWw9Ue9Gb+7OHReSi/9rn5aXzDtz74p4IjTi31ry+9V86ByF4dMcjeG5G4Myd0YkrsxJHdjSO7GkNyNEa65t9iivzGaprFr1y569+4N6HfvCwoKGDRoUINlhwwZwpw5c4Kv//zzT6KioujZs2eD5Wrnh2vRbzKZiIqK2r+Ffe66YfZU38G9oaZB8WbYPB/y/tijYz6H3jFfx2EQndb0NupTTHprAGcqWCL2vXwLcUC5i2YjuRtDcjeG5G4Myd0Yhyt3VVNxeQ9sxKGAFtj3QmFCvu/GkNyNEa65t6qif/r06eTl5fH4448DsGPHDgDS09MbLJuenk5xcTFerxebzcaOHTtITU1tMDRd7br5+fl7fe+CgoIGnQPW9vbu8/nw+eoKZLPZjMlkQlVVAoHQXwomkwmz2XzQ8zRNw+/37/c8RVEwmUxUVVURERER8vkVRcFisaD5q/FXFOmFfs1z+gpgMeud4fkCDe/yW2vm+QMqwb7x/V6U3F8xb12IUpaLBnXzYrJQOw7D3G4oitWur7fHdi0mBUVRCKgqqgYoZohMhKhkLBE2fV4ggKqGrleb98HOO1zHyefz4Xa7cTgcweOwt+NksVgOeh4Q8h2sZbVaAfD7/Q1GMbBYLCiKctDzGst0b/OO1HHSNA23201kZCRWq7XZf57kODU+T1GU4Pl2z30x4rzXVo6Tz+ejqqoqeJ6pP8+I815bOU615xmHw4HVajX8vHcg81rzcdI0jerqaiIjI4PHobH1DuTnyaSZiLfHs6NiB2bMmPbo4zpAABW1wbw4axxm9DuB4X6cVFXF5XKFnGeg9fx+aq0/T/XPMyaTyfDz3v7Oa+3HSdM0PB4PTqczmE19LeX3U2Pb35tWU/SvXbuWW265hWOPPZarrroKALdbL1BtNluD5e12e3AZm80W/N+9Lbc3r732Go899lij88rKyigqKgq+jo6OJioqCrfbjcsVeuU4KiqK6OhovF4vZWVlIfMiIyOJiYmhurqa0tLSBvsZFxeHz+ejuDi0mbzNZiM+Ph6/3x+yH6B/uWJiYigvL0dV1eAXEVXFonpIcmjgKaeorCpkPZOikBKrN6EvcnmpV74DCmlx+rziCi9U7MKe+yP2/F/0iwa1FwQ0E96Uo/Bkn4g/tiMoCslmG2agrMqHb48fzMRoO1azQrlHw2t26s/tV5ugupSEhAQiIiJwuVx4PJ6Q9eLi4rDb7VRWVlJVFfo5YmNjcTgcVFVVUVlZGTLvcB+ngoICXC4X0dHRWCyWfR6nxMREVFVtMM9isZCUlKQfiz3mmUwmUlJSGp0HkJamt6YoLi5ucHJNTdVHRCgtLW1w0kpOTsZsNlNWVtbgpJKYmIjVaqW8vByv1xsyryUcJ7/fj8vlIjU1lYSEhGb/eZLj1PhxcjgcuN1u/H5/g/OpEee9tnKcXC4Xu3btCp5nwNjzXls5TrXnmejoaFJSUgw/79UK9+Pk9/sxmUzYbDbKy8ub5efphKQTGJIxBMWioHpUtOo9Llo6TChWBdWronlD50U49NaH4X6cAoEAO3fuJCoqqu7vSFrP76fW+vNU/zwTERFh+HmvvnA+TrV/x9RebDH6vAeNH6c9c9gXRdufQcwNtnPnTo4//nh8Ph9LliwhIyMDgN9//53BgwczZcoUrrjiipB17rnnHp555hk8Hg82m41zzz2XNWvWsGnTppDlqqqqiIqK4r777uPf//53k/vQ1J3+888/nz///DP4yAG0nCtVoF9x0jSNwsJC4uLisATc4C4FbxmKpmIxm/T11NCvwT7v9JuAHctRN82H3aEd8ymOeJSOJ+FvdzyaLbRjvtq7+SEtBGrnWe0o0akEbHGoB3C1saVeUfR4PBQXF5OQkIDFYmk1d1L2d15LvfLr9/spLi4mMTERu93eau6k1O57az1OqqpSUlJCfHw8pj2GzGzJd1Iamwet5zh5PB6KioqC55n681rynRRo3cep9jyTkJCA3W43/Lx3IPNa83Hy+/2UlpaSlJQU/PumsfVa6u+nA5nXko5T7Y2M+ucZaD2/n1rrcap/nrFarYaf9/Z3Xms/TrW5p6SkYLFYDD/vNTVv1apVHH300axcuTKkDm1Ki7/TX1ZWxogRIygtLWXRokXBgh/qmubXNvOvb8eOHSQkJATv7qenpzN//nw0TQtpmlS7bv3tNiYlJSV45WZPVqs1eBDrM5lMDf74PdR5iqI0+l57m+fz+VA8ZViKd2NVan4YTMH/p69nbnp4vNqm/IDeud/WH2HLQqgqDm0El9KzpmO+fmAy7/XLZam/TYtDH3bPEQ+KghloqusMs9ncZMcaBzvvcB0ni8WC2WzGYrGEHJeDOYb7mgfsdV79X9LNNe9wHIvmOk71t9PcP0+HMg/C9zjV/nIzmUxNfsYjed47lHnQ+o7TnueZfa3XUn4/Hco8MPY41eZe/7GK1vL7qTUfp9q8W8J5r1ZbOE5NnWeg5f9+qtUaj1Nt7vVzbG2/nw50Xks4TrX71lLOe43N29v2G133gJY+wjweDyNHjmT9+vV8++239OrVK2R+ZmYmycnJ/P777w3W/fXXX+nfv3/wdf/+/Xn77bdZs2ZNyHZ++eWX4Pyw5qsCixZsen9ANA1KtujD7eX+Dmq9q1pWB7Q7FjoNg+iGfSvsVYRTL/btsQe+T0IIIYQQQggh9ukgKsAjIxAIMHbsWBYvXsxHH33Escce2+hyF110EV988QU5OTnBad999x3r169nzJgxwWmjRo3CarXy2muvBadpmsYbb7xBZmZmo0MEhguTyUSUPQKT0vTd/Eb5vfpd/flPwoIJsH1JXcEfkwlHj4MR/4GjLjmwgt8WA4ldIalrWBf8JpPe+2dTVx3F4SG5G0NyN4bkbgzJ3RiSuzEkd2NI7sYI19xb7J3+u+66i88++4yRI0dSXFzMtGnTQuaPGzcOgAceeICPPvqIU045hX/84x9UVFTwzDPP0LdvX6655prg8llZWYwfP55nnnkGn8/H4MGD+fTTT1m0aBHTp09vsolJODCbzURH2qB6P3t5rCiAzQth+89QXa+TDcUMmQP0JvyJXeCALiIo4IjTh92zOg5gvdbLbNbH+RRHluRuDMndGJK7MSR3Y0juxpDcjSG5GyNcc2+xRf+yZcsA+Pzzz/n8888bzK8t+rOzs1m4cCF33nkn9913HxEREZxzzjk899xzDXrrnzBhAvHx8UyaNInJkyfTtWtXpk2bxmWXXXbYP4+RVFXF6/Vh0zRMpiYKdU2FnSv0Yn/XytB5jnjoeCJ0OPHA78wrJnAk6MW+JeLgPkArpeeuD2EWblcLWzLJ3RiSuzEkd2NI7saQ3I0huRtDcjdGuObeYov+BQsW7PeyvXv3Zt68eftczmQycf/993P//fcfwp61PoFAgLJKD4kODRN7FP1eF2z7CTb/AFWFofOSu0OnUyC9H5gO8KuimCEqCaJSwNxiv2aHVSAQoKysjMTExLA6abR0krsxJHdjSO7GkNyNIbkbQ3I3huRujHDNvW1WY21dsGO+hZD3OwTqNfu32Gs65jsZYg6wYz4AkxWikvWC3xS+j0wIIYQQQgghRGsgRX9bEqiGnD/0XvhLt4XOi8nQC/12x+iF/4Ey2/Se+CMTD/BZfyGEEEIIIYQQh4sU/W1BeT6Ra2di3vGTPnRfLcUMmUfXdMzX9eCKdWukfmffES/FvhBCCCGEEEK0MFL0h7udK7HMvpWo6ioURdOn2WOh40l6x3yOuIPbboRT75zPHtNsuxpuTCYTkZGRYfU8UGsguRtDcjeG5G4Myd0YkrsxJHdjSO7GCNfcpegPdym9UKJTMZds04fZ6zQMMvofeMd8tWwxEJ0GEVHNupvhyGw2ExMjF0WONMndGJK7MSR3Y0juxpDcjSG5G0NyN0a45i5Ff7gzmVBPuBNflQtrdAKmg2qCr+gtApypYHU09x6GLVVVqa6uJiIiIuyuFrZkkrsxJHdjSO7GkNyNIbkbQ3I3huRujHDNPXw+iWhSIH0AJaYEAqp2YCsqJohMgpReEN9BCv4DFAgEKC0tJRAIGL0rbYrkbgzJ3RiSuzEkd2NI7saQ3I0huRsjXHOXO/2iIcVcM+xeMpjlKyKEEEIIIYQQrZVUdKKOyVo37J7JbPTeCCGEEEIIIYQ4RFL0CzDb9Of1IxNk2D0hhBBCCCGECCNS9LcBJpMJe4QFk+IPnWGN1O/s2+Ok2D8MTCYTdrs9rDoBaQ0kd2NI7saQ3I0huRtDcjeG5G4Myd0Y4Zq7FP1tgNlsJs7pgOoKfUJEdE2xH37DUbQkZrOZuLg4o3ejzZHcjSG5G0NyN4bkbgzJ3RiSuzEkd2OEa+7hdQlDNErTNKp9ATRbDCR1g6QuUvAfAZqmUV1djaYd4KgJ4pBI7saQ3I0huRtDcjeG5G4Myd0YkrsxwjV3KfrbAL/fT7HmxB/TDiKijN6dNsPv91NcXIzf79/3wqLZSO7GkNyNIbkbQ3I3huRuDMndGJK7McI1dyn62wqTPMkhhBBCCCGEEG2NFP1HyOuvv86AAQOwWq08+uijwelPPfUUTqcz+J/dbic6Ojo4f9myZRx//PHExMTQqVMn3n777Sbf44YbbiA9PZ2YmBj69u3L559/fjg/khBCCCGEEEKIFk6K/iMkPT2dRx99lIsuuihk+gMPPEBFRUXwv2uuuSZkmSuuuIIzzzyT0tJSZs6cyR133MGaNWsafY8777yTrVu3Ul5ezjvvvMO4ceMoKio6rJ9LCCGEEEIIIUTLJUX/EXL++edz3nnn7bU3yOrqaj788EOuuOKK4LStW7dy6aWXYjKZGDBgAD179mTt2rWNrt+jRw9sNhsAiqJQXV1NXl4eJpMJm80WdkNPtHSSuzEkd2NI7saQ3I0huRtDcjeG5G4Myd0Y4Zp7eH2aVu7LL78kMjKSU045JTjttttuY9q0afj9fn799Ve2b9/OMccc0+Q2/v73v+NwOBg8eDDDhw+nb9++mM1m4uPjMZvNR+JjiBqSuzEkd2NI7saQ3I0huRtDcjeG5G4Myd0Y4Zq7FP0tyNSpU7n88stDriyNGDGCKVOmYLfbOe644/jPf/5Denp6k9t47bXXqKio4Ntvv+WMM85AURQ0TcPn84Xd0BMtneRuDMndGJK7MSR3Y0juxpDcjSG5G0NyN0a45i5FfwtRXFzMl19+yZVXXhky7ZxzzuGZZ57B6/WydOlS7r//fpYuXbrXbZnNZk499VS+/fZb5syZg9/vp6ioKOyGnmjpJHdjSO7GkNyNIbkbQ3I3huRuDMndGJK7McI1dyn6W4gPP/yQPn360KtXr+C0TZs2ERUVxejRozGbzfTr14/jjjuOhQsX7tc2/X4/GzduPFy7LIQQQgghhBCihZOi/wjx+/14PB4CgUDIv2tNnTo15C4/QLdu3aiqqmL27Nlomsbq1atZtGgRffv2bbD9srIyZsyYQUVFBX6/n48++oj58+dz0kknHfbPJoQQQgghhBCiZZKi/wh54okncDgcvP322zz55JM4HA6mTp0KwObNm/ntt9+49NJLQ9aJjY3lww8/5JFHHiEmJoYRI0Zw5513ctpppwHw1FNPMWLECEDvrf+tt94iKyuLxMREJkyYwIwZM+jfv/8R/ZxCCCGEEEIIIVoOi9E70FY8+uijPProo43O69SpE9XV1Y3OO/PMMznzzDMbnffAAw8E/x0TE8P8+fMbXU5RFKxWK4qiHNhOi0MiuRtDcjeG5G4Myd0YkrsxJHdjSO7GkNyNEa65S9HfBlgsFhITE43ejTZHcjeG5G4Myd0YkrsxJHdjSO7GkNyNIbkbI1xzl+b9bYCmaQQCgbAbeqKlk9yN0ZZzf/311xkwYABWqzWkZdGuXbsYOXIkKSkpjV65XrVqFSeddBIxMTH06tWLBQsW7PV95syZQ9++fYmKiqJz5878/PPPbTp3I0nuxpDcjSG5G0NyN4bkboxwzV2K/jbA7/eze/fusBt6oqWT3I3RlnNPT0/n0Ucf5aKLLgqZbjKZOPvss5kyZUqDdXw+H6NGjWL06NGUlJTw0ksvMXr0aIqKihp9j7/++otbb72Vt956C5fLxcKFC+nQoUObzt1IkrsxJHdjSO7GkNyNIbkbI1xzl6JfCCHCxPnnn895551HXFxcyPTk5GRuvvnmRjv2XLduHSUlJdx+++2YzWZOO+00jj76aGbNmtXoezz11FM8+OCDHHPMMZhMJrKyssjIyDgMn6b1ONgWFieffDJ2ux2n04nT6Qx2zNqYRx55hOzsbGJiYujatSvvvPPO4fgoQgghhAhDUvQLIUQbt2cTNk3TWLVqVaPL/vrrr+zevZsuXbqQnZ3N+PHj8Xq9R2I3W6yDaWFR6+2336aiooKKigrmzp3b5HLjxo1j7dq1lJeX8+WXX/Lggw+yYsWKZvsMQgghhAhfUvQLIUQb1r17d+Li4pg4cSI+n4+vvvqKhQsXUllZ2ejyeXl5zJw5k0WLFvHnn3/y22+/8fTTTx/hvW5ZDqaFxYHq2rUrUVFRAMFWA1u2bDnk7QohhBAi/EnR3wYoioLFYgm7oSdaurac+8E2d969ezfnnHMOUVFRdO/ene+++26v7zN58mS6du2K0+mkZ8+ebNq0qU3nfjCsViuffvops2fPJi0tjeeff56xY8eSlZXV6PIOh4PbbruN9PR0kpKSuPPOO5kzZ47kfpDuuOMOkpOTOf3001m+fPlel50wYQJRUVF069aNzMxMTjvtNMndIJK7MSR3Y0juxpDcjRGuuUvR3wZYLBaSkpKwWGSExiOpLed+sM2db7nlFtLS0ti9ezfPPPMMF198McXFxY0u++WXX/L8888ze/ZsXC4Xn3/+OQkJCW0694PVr18/Fi5cSFFREfPmzWPz5s0MGTKk0WX79OkT8ouw9t+S+4F7+umn2bJlC9u3b+f0009nxIgRuFyuJpe/7777qKioYMmSJVx00UVERERI7gaR3I0huRtDcjeG5G6McM09vD6NaFT953XD7apVS9aWcz///PMBfWi3+mqbO+/cubPBOhUVFXz66ads3ryZyMhIzjvvPPr27cvs2bO55pprGiz/+OOPM3HiRHr16gVAly5dgLadu9/vx+/3EwgE8Pv9eDwerFYrZrMZj8cTfPbe4/GgKAo2mw2A5cuX061bN1RV5dVXX0VVVc4666zgdn0BlYCq4QuojLl0HM+/8BJHHzcMTBb+/fSznHDyaazKK0PVNBQFTIqp5n8VFAUUQD8U9V8rwelKzXRqXysKpnrTa1YNeb3n+ijUvWdT221qfYO+J/UvrNxzzz288847LFmyhNNPP73JdRRFYejQoUybNo0333yTm2++OWSeODLa8nnGSJK7MSR3Y0juxgjX3KXobwP8fj9FRUUkJiZitVqN3p02Q3I/MBs2bMDpdIY0K+/bt2+jHcoFAgGWLl3KypUrufrqq7Farfztb3/jwQcfbNO5P/HEEzz22GPB108++STvvvsuV199NQ6HIzjd4XDQvn171m7YhC+g8sabbzN92hRUVeXEYcN56b8z2LDLhV/V+HXxT9x8xRh+WZcHwPBRl7Jy7UZOHHw0ZouFs0ZewLjrb8Pv9+N2leKIjkMz1/6SbD1j3O55ccBk2vMiAezrokWF10+5x0d+qTtkeqHLA8Bul3evFy1QFKq8fiq9/n1etPD5fGzYsKFNf9+NJLkbQ3I3huRuDMndGOGauxT9QogWoaKigpiYmJBpMTExjY4Xv2vXLvx+P19//TUrVqygtLSUM844g/bt23PJJZccqV1ucf750MPc9+BD9e7Ma/hVlZziKjbvrsAfUPGrGgFVQ9Ngw64KAG6+71/cfN+/Qrbl8akADBhyXLDgB71Y/cd9D/OP+x4OWV4NtO7xbGsv7Ov/qxFQ9f/dH36/n4Dfj9vro6KqmvzCciw1LSy8Hg9FRXqT/W27SlEUhQibjfKyUlb99ScDhx6Hoii8/95bFBYVk9ypN5t3N+xEceaM9zjjnFE4o2P4fcmPTJs+gwkvv8Xq/HKqK6twK5E4I1UiI8w4rGZMpvC5OyGEEEKIQyNFvxCiRXA6nZSXl4dMKy8vx+l0Nli29q71PffcQ1xcHHFxcdx4443MmTMn7Ir+2kLdr2r4Ayq+gBZsaq8X8HXTtIO4sa5pGtUBFY9PxesL4PGreHyBkH/r/6l4/AG8fhVPtT5PX6Zmns+Pz+fHZrViMZswmxTMJgWr2YTJpGA1KVjMCmZFCc631Fum/muL2VSzXOi04L8be22q2YZZaXTbpsPYRO/Nl57hjef/E3z91svP8q/nXmXUxZczuGtacPrgrmlkZGXz1eIV+P1+XpzwGFs3b8RisdC9d19efe9DomNiAfhy1oe8/cpEZn23BIAfvv2KF//9KD6fj/TMLO76578YdtpZqAE/AVWj3OOjwqd/ARQF7FYTjggLkVYzkTYzNov5sH1+IYQQQrRsUvQLIVqErl27UlFRQV5eHpmZmQCsXLmSK6+8ssGy8fHxZGRkNNqhXGvQHIV8QNWCBbm3tjhvpGD3+lTcvgBev/5vjz+Au7bAr1n3YC4W7ElD31dvQKU60PKa9SsKjVw4UDCb6qZZTYp+gaLeRQiLScFsDl0muE7NRYn+o67n7QtuDJlmMZnYWFDBZ3/m1Vy8qNumy+PDHh3HlM++D14c2dM5F1zMORdcHHz90jv/2+/PqmngrlZxV1dT2w2m2aQQGWHWWwJEmImMsDT6vkIIIYQIP1L0twGKomAymVpVURQO2nLuB9OhnNPpZNSoUTzyyCO8/PLLfPfddyxfvpxRo0Y1+h5XX301Tz/9NEcffTRlZWW8+eab/POf/zQs9z0Lef1/tWBTe7+qUu1XqaquvTNer1ivuZPu9e9xV92nhhb2/rp1/Ue4sDabFOxWE3arfte47t81/2sG1edFsdgIoOifud7jBLXZBOpN89Vc4PAH9HlqvWmq2rz7r2ng86v4mnezzUJRCGnBEHLBoeYCQ2gLBhNWs36BwqJo2PDTIVUhO9FJYlREo9/9gKrh8vhxeeoew7BZTTis+oWAKJsFm6Vtnq8ORls+vxtJcjeG5G4Myd0Y4Zq7omnNcY+nbVq1ahV9+vRh5cqV9O7d2+jdEaLFePTRR0M6lAOCHcrteRJt3749W7duBWD37t1cddVVLFiwgKysLF577TVOO+00AKZPn87/t3ff4VGVaRvA7znTJ2XSCYRiABEMWGjWCLKoQQRFQUFYUVFARV37pyIoCCgqFhBRUVlRkGJbXVFQF0VUrKAoLIEFCSGUQPr0c97vjzNzkkkjgSQnmdy/68rF5MyZyTtPhpN53vK8s2fP1gr7+Xw+3HbbbVi5ciViYmIwceJETJsWvs78RIWSVrdfRqnHj1KvjBJ3ACVeP1w+GWVe9ZiayFdO3stH3xt6VL0+rMHk3GaSYA0m6XazMSxht5mlYCJfftte4bbVLMFsbNodXhUR7CyQhdZponUUBGsVVOxQkGUF/rCOA7XTIazDoUIHg1yhg6bKOZV+bvlt9V9Zab5/Nq1mCe3j7GgXZ0f7eDvS4hxoE2uFqQ6/P4MBwdkApuBsAGOT/96JiIjo2OqbhzLpPwFM+omaL0UR8AYUuHwBuHwy3H41OS/x+FHqDaDUG0CZR0ap148yr4wyXwBuX/m/bl/5FPmmHlU3GQ3lI+gmI+xmNWGvnKjbgiPuVrOx0m31HItJatS17K1VqEOiuo6Disdq6lwILeVQhNCWcVTfKRHe4RDq+KjcKVHsCdRac9AoGZDqtCFN6whQOwVs5mOv8zebDHCY1U6AKKvaGRRpox9EREQtTX3zUE7vbwX8fn9Ebj3R3DHuja/UG0BugRv7ClzYV+DCX0dd2HukDAcLXfALCQLl+60KEcyLGit/N6A8ITeVJ+m2KqPpoXPUEfSwUfVgUn+io6uVt3nTblfZq7688dVtJRe8J/z5DKFnDH8+JRBAcWEBop3xkExGKELteAHUJFmI4L9Qfyfq92otAEVR7wPQ5DMhjodkMEAyGlCHnLnRKbIa96OyFXnFPuQWqv8f8oo8WmeVrAjkFriRW+DGD7uDDzQAydFWpMWpnQBp8WqHQIwt/FrlDwgUBfwocqsLI9Qigcaw+gCtsUggr+/6YNz1wbjrg3HXR6TGnUk/ETVriiJwuNQbTOzd2HvUhZyjLuQcdaPA5QtLJhFMIoUsw2AsT3hRcWCy0iGTUdIScXsoITeribjdbITNUp6Uq8dMcFjDjzmC66GNkiEsqQ79IEM1STUQvkd7xftD7atur3b1vlqSdJ1GYf1+Pww+MxJjrSf8R1JRhPY7DXUYiFAHgVA7DRQBIPi7r3guKnYmiPIOBlHp3Oo6HkLFCFtCx0NFZqMBJ8U50DmlfMtLWRE4VOJRO8UK1YQ/t9ANt09WTxDA4RIvDpd4sTmnUHtcjM0U7ABwaB0CidEWbcaIWiRQnQkT2kyzYpFAh9UEu9nIIoFERETNCJN+ImoWPH45OEqpjlTuPeLC3qMu5Ba64QsoYcmfxgBIBkCS1K3gUp02tI+zIdYkIzk+FtE2izol2WLS9i8PFS0LJfYWE9csNzeSZIAEfZNGRQnvdNDefzV0PIR3UIhaOx5qPLcBOx6MkgFtnXa0ddrRN3hMCIEClx+5Ba6wjoBCV3l5wxJPANvzSrA9r0Q7ZjVL4TMCKtUJCC8SqBbpVLcMVOsDOCzGOi0lICIiosbBpJ+ImowQAkfKfNgXnGqcE0zucwpcOFziDUuqtGn4wRHtUHIfZTWhQ7wdHRIc6JjgQPt4B9rH25HqtMFslCJ2WhY1LSk4Um3UsfOhuo4HVPre5w/giOKCPcqCAKRaa1AYDAYkRFmQEGVBr/Zx2vFSbwD7gx1uuYVu5Ba4cCj4/xEAvH4F/ztchv8dLtMeY5QMaBNrQ/vgsoB2wU6BUHKvFrJUUFCmdihIEmA3qx1udosRDrOxTsUFiYiI6MQx6SeiBucLKMgrKh+1zzkanJZf4ILbJ1cY3URYci8F14ybJQNSYqzokGBHpwQHOlRI7p12MwuJUatQl44HvxHw281IdNq0Ti5ZEfCGtnwM/usNqFtGVifaakK3NjHo1iZGO1bx/3BucFbA/iJ3WJ2A/YVu7C+sUCcAQFKMBWlxDqTF2bQlArF2MxQFasFMr6ydazFJWl2A0Ewc/t8mIiJqeKzefwJYvb/l27x5M2677Tb8/vvvSEpKwkMPPYSbbrpJu7+oqAinnHIKevbsic8//7za5/B4PLjvvvuwatUqBAIBTJw4EbNnz26ql6AbIQSK3QHkBNfa7ytQp+PvPaqOEsqKCJvWDCBs1N5gMMBmNqJ9vB0dg6P2obXE7eJsrbI4GFFjCu1o4Q2o20eGtpH01XErSVkROFyi1tfILSzvDHD55FofF20zqTMCgrsGpMXbkRRtrbKzRKhIYJTVqO0YcCLLb2q6vi9atAhz585Ffn4+4uLicMstt+DBBx+s9jm8Xi+mTJmCDz74AEIIZGVl4aWXXkJMTEy15xMRETUFVu8nqoe///3vGDVqFDZs2IDNmzdjwIABOO+889CjRw8AwPTp09G1a9dan2POnDn4/fffsW3bNvj9flx22WXo3LlzWOdBSxaQFRwoDhYECyb3fwWL6ZV4AjVOyQ8l9pIBSIpWR+07Vhi1T4uzIzHKoo1mElHjkiQD7MGR9YqECHYG+Kt2CFTsDAht/ZfqtFWtExDsAAh1CFSsE1BaQ52A0JKA0DaCqbE2rVAg4AOgbl9ZPhvABIfZWOdrRk3X96ysLIwePRpxcXHIy8vD4MGDcdppp2Ho0KFVnmPBggX45ZdfsH37dlgsFlx11VWYPXs25syZU7egExERNQNM+luBQCCAo0ePIiEhASYTf+UV7dmzB2PGjIEkSejduzd69OiB7du3o0ePHti6dSu+/fZbTJ48GcuWLavxOf79739j2rRpiI+PBwDceeedeOmll3D99de3qLiXeNQP7vuOBqfkByvl5xW6EQhVU684Jb/iWnuDAbZgsa8O8XZ0SIxC+3j1dlqco0qS0Zj4ftcH466Phoh7aNaNuh6/vA6GEAI+WV2b79WWCcjw+Ms7A8LqBKQ5tceWeQNhswH2FbpxqNgTVidg9+Ey7K6hTkBacFZA+3g7ArI6qwjwwhDcGtNhVTsA7LUUCazp+j5ixIiw8yRJwq5du2p8jqysLCQmJgIArrjiCnz22Wd8v9eiphkWX3/9NR555BH8/PPP6Nu3L9avX1/jc6xfvx6DBg2Cw+HQjq1ZswbnnHMO464Dvt/1wbjrI1LjHjmvhGokhICiKOBKjqpuv/12vPXWW3jkkUfwyy+/YO/evTj77LMBAHfccQeeeeaZGj8MVlQxtkII/PHHH80y7pW3v9tX4MZfR8qQc9SNQpcPCsKrkgMoH7WHAZIExDks2lr70Dr79vEOpMRYm8WofXOMe2vAuOujMeNuMBhgNRmDS23Ci2L6AuUdABVnByjBsgFRtdQJyK1QNDCv0A1/NXUCKkqMtpTPCAjVCbDJOBpcHiBJ0HYJqFgksLbr+7JlyzBp0iSUlpaic+fOGDVqVLUxGD9+PO655x4cOnQIVqsV7733HkaMGMH3ey1qmmHhcDgwceJE5Obm4pNPPjnm83Tu3Bk7d+4MO+b3+xl3HfD9rg/GXR+RGncm/dSqDRkyBNdddx1mzZoFAHjttdfQtm1bvPPOO0hJSUFmZuYxk/6srCw888wzOO+88+D3+/H888+jrKys1sc0NrcvtP2dmtznHA3f/q7y3vahUXtDcNTeYlSn8XasWCE/QR19i7GxIj5Ra2cxSbCYJMTYwo/7ZSW4VECGJ/ivN6AgIAtYTBI6JUahU2KUdr5WJ6DQVb6DQKU6AUdKfThS6sNv+4q0Y9E2E9oH6wOEZgRUrBNgNUvofd6FuOe2m6tc3wHg2muvxbXXXoutW7fi/fffr3GNfteuXZGcnIzU1FQAwMUXX4zJkydDUaoviki1z7Do27cv3nnnHb2bSETU6jDpp1br6NGjGDp0KF5//XWMGDECf/zxB7KystC1a1dMnz69xsJ9lT388MMoLCzEGWecAYvFgptvvhlHjhxp5NaHb38XSu5D29/ll3rD9xavZtTeKBkQbTOhQ7xDHbVPcGjbb6XG2ridFhHVm9kowWyUEG0N/3gRCHUGVKgX4A3IQABanQB0Us8VQqAwWCdAnRGgdliGtv8DgnUCDpRg+4FKdQKcakdAnOTFQ6OvxKNPzcfFlw7Hrh3bcOvfRyL1pJNxdr9+cFiNMBsl9OzZE5999hlmzJiBuXPnVnk9t956K8xmM4qKiqAoCm6++Wbcd999ePLJJxsngBGgthkW9ZGTk4OUlBQ4nU78/e9/x8MPP9wIrSU6ftHR0WHfu1wuPPXUU7jnnnvCjg8ZMgTr1q1DIBCo8bleeOEFPPPMM9rnyUWLFh2zphRRfTDpp1Zr165diIqKwsiRIwEAp512Gs4991z88MMP2L17N/r16wcAcLvd8Hg8OOWUU/Df//63yvPY7XYsWLAACxYsAAC88sor6N+/f4O10xdQtBEwda29ugXevgIXXD65aiG9SqP2kmRAm1hrhSJ6dm5/R0RNymSUYDJKiLKGH1cUEbatYKhDID7KgPgoC3pWqBPg8gXKtxAMXhOr1AnIL8Pu/DIU790OWbLgc086tn6ejbS4GLTrdjpW/Gst4jt0h91ihNlkgMNsQmGZB9t3ZENRRJUlSlu2bMG8efO0mQDXXXcdk89jqGkGXX10794dW7ZsQbdu3bB9+3ZcffXViIqKwh133NEYTSY6LqWlpdrt/fv3o2PHjrjyyivDzvnggw9QUlJS+aFhfvrpJ0ydOhUbN27EqaeeiunTp+PGG2/E119/3SjtptaJSX8rYDKZ0KZNG72b0ex069YNLpcLH374IYYPH45t27Zhw4YNmDx5Mvbu3audt2LFCrz77rtYvXp1tc+zb98+GI1GtGnTBps2bcLs2bPx0Ucf1SvuQggUuf3ho/bBfe0PFnmDI/aV9ravuNbeYECU1YgO8Y5gcm/XKuS3i7Of0LZXLQ3f7/pg3PXR0uMuSYbgevzw44oSKiIoazsLWM2SVisgxC+rnaIVCwbmFbrhSG4P2e/Fwd++gZxxHnZs344/fv0Bhh4X4YaHnkLXPuejc4d2wJHdePPFF3HdpDvwx/6i4O4GapFAh9WIvn37YunSpcjMzIQQAm+99RZ69erV4uPeWGqaQderVy/07t27zs+TmpqqLak49dRTMXXqVMyfPx933303464Dvt+PbdmyZTjnnHOQnp6uHfN4PJg6dSoWLVqEgQMH1vjYv/76C7169UKvXr0AqMuP5s+fz7jrJFLjzqS/FeBIbvWcTidWrlyJBx54AOPGjUNCQgLuvvtuDB48uMp5FosFKSkpAIC9e/fi1FNPxZ9//omOHTsiOzsb48ePR35+Prp06YIXX3xRu3BXFpAV5BV5tNGqfQUubUp+qTegJvUVp+RXGrU3ScHt7+Id6JgYPmqfGGXh7xp8v+uFcddHpMZdkgywSVUr82vbC1aoFxBrN+GkpChtxF8RAodKvPjU8SKWLZiDbe/Mhskeg44XjEJCt77Y/u6zWPPvl6H43LDEJKBtv0vxW+zZeOTDPxAbKMSK/xuFOW+vw+ndu+LGux/BE1PvRfv2HQADcEHmBVi06LmIjfuJqmkG3VdffVWvpL8ySZIghGDcdcK4H9vSpUsxZcqUsGNPPPEERo8ejfbt29f62Isuughz5szBr7/+ip49e2Lp0qW4+OKLGXedRGrcDSLSShM2oT/++AM9e/bE1q1bkZGRoXdzahQIBFBYWIi4uLiI2nqiuVIUgaMuH3KPlmHn/iMo8kvILfKGbX9X3ZT80Kh9aEuqtHi1Qn5ahQr5aXH2Jt3+riXi+10fjLs+GPdyFXcRCC0X8AbUHQWEECh0+5Fb4A6rFVCxTkB1LCYJ7eJsSItzIC3ejvZxdnRKciDKLMHnKkFMbCxMJrPWMatuY2oov6YbDGHX9srnGQyR9QGzqKgIHTt2xJtvvqnNoLvwwgvx9ttvY9CgQfD5fFi1ahVeffVVrF27FkajEWZz1eKw69evR5cuXdChQwdkZ2fjqquuwrXXXot7772X73cd8DpTu99++w1nnXUW8vLyEBcXB0AtaHnppZfil19+wYEDB9C1a9ca1/QrioLHH38cM2bMAAB06NABGzZsQGpqKuOug5byfq9vHtp8Xwk1GCEEAoFAxG09oZdQkamDJR4cLPbiYJEHh0o82F/owcESDw4VexGQFSjBuBuMRqjj9ZXW2hsMSHJYKo3aq8l9cnTz2P6uJeL7XR+Muz4Y93Kh7QVjbdVvL+gNKOicHKUtF5AVAZcvUN4REFwicLBCnQBfQMGefBf25Lu055MkICXGimijAovVCqPRGLymqzMUjMFkPnSdDx2XDAYYtY4AA4xS+G2jpBZYlSQDzJIEyVB+zGgwaLdNRvU+k1GC0WCAKXSO0VDhewkmI2CSpPLHSQaYjAYYDVLweYLHguc3RAdEbTPo1q9fjwsvvFA71263Y/z48ViyZAkAtSjamjVrkJmZiZ9//hljx45FYWEhUlJS8Pe//x333nsv3+86Ydxrt3TpUgwbNkxL+AHgrrvuwsyZM2Gz2Wp+YNDixYuxYsUK7Ny5E2lpaXj++edx+eWX4/vvv2fcazB37lzMnz8fhYWF6Nq1K77++mvExMRg06ZNuPPOO7F161bEx8fjueeew1VXXVXl8QcPHsRNN92ETZs24fDhw1W23o7EuHOk/wS0lJF+v9+PI0eOIDExsdoedQonhECxJ4BDxcGkvtiDg8Ue5BW5cSD4vS+gQKDCVHxA29c+NJpjAKDIMiwWM9Li7OiY6ECH4Kh9Wjy3v2ssfL/rg3HXB+N+/AKyEratYKh+gNsnI6/Ig9wCl9YRsL/IA3+gfJs+AQEhyzAYjTCg5XfQlhd+Le+sMEoGreMhVBTWWOl4qENCCnVYaJ0T0DoXQp0aJkmCFOzcUDsjAGOwc8NslMLOM1Z4PpPRoLVJKDKEz4UuacloFx/FYrRNhNeZmimKgg4dOmDRokUYNmyYdjw+Ph5Wq1q5VJZl5Ofno02bNvjiiy+q5Ay33XYbYmNjMWfOHABqgcCYmBgcPHgQiqIw7pW8+OKLePfdd7FkyRJ06NABv//+O0455RQUFBSgX79+ePXVVzF48GAUFhaiuLgYnTt3rvIchw8fxurVq5Geno4hQ4aEJfgt5f3Okf4WrLpeqy+//BIPPPAA8vLy4HA4MGbMGDz11FMwGqtO8RZCYOrUqXj99dfh8Xhw/vnn4+WXX0ZycrIOr6Z5K/MGgsm8F4dKPMgrCib1RepIvdtfXhU/bMs7QFtnL4WmaUpArM2M1Fgb2sbZ0CZW/Up0mGCRXejeMRV2m7WmphARkQ5MRgnR1WwvKCsC3dvK8PiDMwT8Clz+APYXeLRigfuOlqHE5YZkNEOBQdseVVYElGDRVfVfAUUREKhwn1J9e/QkBCALAVkBtB5svRmquymgyDKMxjwYDAaYjQYkRluRFG1Bcow1+GVDm1grUmNsSIyxwGYywigZ2DlAjeKLL76A3+/HkCFDwo7/97//hRL8z56Tk4NzzjkHmzdvRlJSUpXn6Nu3L+bNm4cpU6YgNTUVr776KtLS0hAfH98kW0C3JLIsY9asWdiwYQM6duwIQK0dAgDPPvssrr/+emRlZQEAkpKSqo03ACQnJ+OWW27BgQMHmqbhzQCT/mbixRdfxKeffoqNGzdqvVYWiwV9+/bFhg0bkJycjIKCAowcORKLFi3CbbfdVuU53nvvPSxduhSbNm1CamoqJk6ciHvuuQdvvvmmDq9IX26fjEMl5SP1+wvVhP5AsToVv9QrVztKH1I+Wq+OfERbTUiJtaKt0462TjWpT4mxItVpQ0qMrdp19mpPocz97omIWhBjDTsKdG8Tq80IcHl8KCw4gpi4BBiPY81neYeAejvUIRDqOJCrdCKEzlHrxoQ6FireryjhnQ2yUrUjQj0mwn9GhceGPV/wZ1XuyKju+cIeG3ouIWp8nYoiIAcL1ypCaLVuQh0kNQn9zQ7VxJH9Ql2eUeCu/gEGwGk3I95hRrzDgsRoC5KirUiMDnYQRFkRZTOGz1ioMHPBGJz9EJrh0Jw7DqobOMrJycE//vEP/Pjjj3A6ndizZ0+Nj3/77bcxadIk7XtZluH1enHo0CE4nc4aH9faLV26FKNHj66y9jtU/BlQq/gD0HakAICMjAw89NBDGDt2LMaPH49t27ahX79+KCsrQ48ePWrcMaq127dvH1wuF1avXo158+YhLi4O9957L26++Wb88MMPOP/889GrVy/k5+fj4osvxgsvvMD3bxCT/magtl6rtLS0sHMlScKuXbuqfZ49e/YgMzNTe46rr74aDz74IEwmE5KTkyFJkZN8+gJKMKn34ECRB7mFbhws9qoj9SUeFLsD5aP0QNWBi0pV8e1mYzCpt6FtrB1t42xIjrEiNdaGlFhblZGguojEuLcEjLs+GHd9MO5NR5IMwS39jIhzmJEaZ4ckSc06EWyJQp0acrCzQK7QqeGXFRSUeXGkzI+DxV7kFbpxIDhr73CJB0WeQPjfewEUufwocvmxB65qf57dYkS8w6J2DERZtNsJwdvRNpNanBHqZ4awpQcV6jBU7iAIXw5RXpOhMdQ0cGQ2mzF69GiMGzcO06ZNq/U5xo4di7Fjx2rfP/nkk1izZg2SkpIghOB1pgZ1GVg76aSTqhTx++OPP7TbkiRh7ty5mDt3btg5jHtVubm5KCoqwo4dO7Bnzx5kZ2fjb3/7G7p3747c3FwsXboUa9euRbt27TB+/HjcfffdeO211+r1MyL17yqT/magtl4rAPjmm28wdOhQFBcXIzk5Gc8++2y1zzNy5Ei888472L17N1JTU7F8+XJty4/qlgM0Z35ZQX6pV/ujvr/IgwPBP+6HSrwodPkggPLq99UJVk+WYIDFLCElxoo2ThvaBqfhp8ba0SbWipRYG2Jtpgb/4NYS4x4JGHd9MO76YNz1wbg3HkkyQIKhxg+oyTE1F0bzBmTkFXqwr8Clfm4oUgcH8ku9KCjzo9Dtq7K8wu2T4fa5sb+w+tkCJqMBccGZAqGvhCgz4hwWJERZEGc312tGX6iuQcWZBSfScVDbwNHJJ5+Mk08+Gd9//32d2xeydOlS3HXXXQD4fm9MiiJQ4g2g2O1HkduPYo8fxe4Aij1+lHgCMBoAh8UEu8UIh8UIu9kIh9Wk/hvshLSb1a/WUADabrcDAKZNmwa73Y7TTjsNo0ePxieffAK73Y6xY8eiW7duAICHHnoIl112Wb1/RqS+35n0NwO19VplZmbi/PPPR1FREXbv3o2lS5eGTRmqKDU1Ff3790fnzp1hNBpx2mmnYeHChQgEAigqKoLT6Ww2W08oiggm9R7sC069zyv0aKP3BS6/Om0QqHV5ocEAGIKFf5JiLEiNVZP5tk4r2sbZ0SbWhtRYG5x2c5NfDJtj3FsDxl0fjLs+GHd9MO76OFbcrSYjTkqKwklJUWHHhRDwBhS4fAHkFQV32yn24HCJFwVlPhS4fDjq8qPQ5YPXH94rEJAF8kt8yC/xVd8oAxBrMwU7AyyIqzRTIN5hCVsCqCjqZyD/cdZOCC0/DHUc5OXmoLTMhcVvLsPilxbA6XTi1jv+getvvEnrRPD4AxBQB1RCnQ21+fXXX7Fr1y6MGjVKjUGEvt9DyzwBlBdn1u4rr+uESsfKb1d4Hqi/V09AQUkwgS/xBFDi8aPYoyb1JR4/SrwBlISSem8AZd5AeN2o0HMGf4iiyJAkozYwZQirdVH+TWi751AHQVSoM8BiRLTVBJvFCIfZqN0f1mlgMcJhDu9YaK4dCN26dYPFYgkbqAvd7tmzZ7XH6ytS3++R80pasNp6rTIzM7Xz0tPTkZGRgSlTpmDlypVVnuexxx7Dn3/+iUOHDiE6OhoPPvggxo8fjxUrVsDv9zfp1hOhvepDU+/zitzBpF5N9I+W+bT1hrWq8MctMdqClBgbUp02tHXakBZM6tvE2pAYZWl2FyghRJPHnRh3vTDu+mDc9cG46+N4424wGGAzG2EzG5EQZUVGO3WNb2gHB7dPhscvw+MPoMDlx9FSP466fCh0+XA02ClQUKYeK/VU2mtdQB2ZdQfw15HqlxDYzJK2dEDtDDAHOwfU72MqLCE4dgyg1T/wQ+CvvftQUlyE//53Bz7ZuAV79+zCzaMvR0K7dPQ561wAwL4CDwKygu15JcF4IGynBanSbIKXX3sDWZcOQ8BoQ5HLD0UJoNjlgWR1wGiq8NmtmiQ49LsRCE+SIcqPqd9Wk3AfI8Gu/NzaM4Ul4JWeu4bkvS78sgKXT0apJ4AyXwClwSS9zCujzKfeDh0r9coo8wYgK/W8JlRabhqqJ2UwABajBJ9fUaetBhN8LS4CUFA+OCagLn8p8QRqXNZa+Wblt5y2uXSwAyE0g8BuUf/vOCp0FtgsRkRbgp0JoY4FsxFRFhOirOq/juC/Dfn5PCoqCiNHjsSsWbPwwgsv4H//+x9WrFiB1atXw+PxYOLEiRg3bhxSU1PxxBNPYOjQoTU+l8fjgdfr1W4bDAZYrdaIvb4z6W8Gauu1qiwQCGDnzp3V3rdlyxaMHj1aq9Z/00034bzzzmv4BkO98Ba5/dhXoE6JywsWyQv1nB8p9cEnK3UqAhy60MVHmbV19KmxNqTF29HWaUeqU03qWRCPiIiIGkp1OziEZgV4/Qrc/mBnQECGP6B+oPHLCgpcPhS6/OoMgTJfcLaAXzteOfHz+BXkFaozGqtjlMqXEKizBcpnCoSOm2v4DGQN7gM/6R/3w2a3o1uPnsgafiW++c86LemvrHy3hqof0mRZxrurVuKxp+ZrBRIVOQB3iQfFiguSsWWmDooQcPvkCol7AKU+GWWeAEp9FY5VSOorz/o4bsHPubE2E2LtZjjt5uDvVf3dOh0W7b5Ym3p/jM0EIxQcOXIEMc54BIQEt1+GyxeAK9hJ5fKpX25/AG6fOpPF7ZPh9qsdEC6f+jq083wyAnJwy2lA6ywI5bbq5h3qNy5ZhssrN8hrtxgltRMh2HlgN0taJ1zFWQc2kxF2q3pbnXkgVVjaYEKUxQiTUcKcZ57D7bdMQlJSEhITkzDt0cdw1jnnwWAA7vzHP3DeeefB7/cjKysLTz/9tNaU6OhorFmzRhtQDQ26hm536tSp1mKXLV3L/J8bYWrrtVq5ciXOPvtsdOzYEdnZ2ZgzZw4GDx5c7fP07dsXK1euxKhRoxAdHY3XX38dvXr1Oq42CSFQHEzq84rUNfWHir04EEzqD5d64QvULakP9WI67WpS3yZWHa1vFxytT42zIznaCouJST0RERHpp+KsACfK9+iWFQGPX02oUmKt8PjVnRwqDwYqQqDEE9BmCKizBdQOgVAnQeVkUlYEjpT6cKS0hiUEAGKCSwjio9SaAtrSgfj2MFvCt5o4kRpF3339HwghcM4Fg477ORpbqGOmzBtAmU+ulLAHk3lvICzBd/mq/q5ORGjafGhUO8pqQqzdhLgKHTWJUeU7RkRb6187yh98n1hMEqLM5rD34/HyBco7s8q8Abj9ameAS/sKhMW1/FjwMb4APD4Zbr8Cv7rHZ+1LcYX6M30BdRbCiTIbDbCajbAPuQ9XDvs/2EwS/ms2YuoHW2E1SbB1vRj3LMnSOhf+vb0Idksp7GYj/v3z/2AxS/htXyEcFhN2Hy7VZldIkrpY4kCRByZDM9xXtQEw6W8mXnzxRUyYMEHbU3LmzJnIzMzEzJkzcffdd6OgoACJiYkYNWoUHn/8ce1xFXutHnjgAezbtw89evSAz+dDnz59aq1YWez2IbfQUz5SH6x8f7jEi8MlXnjq08NpAKKtRiRH27St7VKDSX1anB0psVbYzJFXFIOIiIgin1EyIMqqJncVeQOy1gGgfimQDAY47WakI6ra53L5Aih0+cuXDrj8Wm2BApcPxe6qyZG6PjyAvUerLiFIyLgAN9/zMM4bdx+k0kP48L3VuGHafPy85yiizQIFJS41UfZ4IElSlU6Cij5+bwWGDL+qSdcyB2RFSzJLqxlxr+77gNxwGbzJaEC01RRM4oNfFiOibSZEWcqPR1tNiIsyIc5ugcNqUkewzZI2kt0SdvKwmCRYTBKc9hPvQPDLitZpUOoJoMQbgCu0DMIja7dDHQehjgVP8DFuv6xtg1rX36dfFvDLAZRWP2mmXkxGtYPPagrOPDBJsJqNiLFKcJoF7spKPPEf0owYRKQtWGhCf/zxB3r27ImtW7ciIyND7+ZUyxuQsSWnENkHS4KV7304XOJFfqm33tN2bBYJydFqtftUbc96O9Li7WgXZ4PDwj6kioQQCAQCMJkafmcAqhnjrg/GXR+Muz4Yd320lLirBd3UxMYTKO8QqLxzQE38soIitz+snkCoQyD0fcWp+X53CbatfApHd/wIs8OJk/42FmlnD4P7aB6+nT0m7Lnbdu+N6x5/AwlRFjx3y3CMuvF2DB95NeIdFvg9Lgw8sxuWvPsJTu11hvYYIQSEIsNQoaBcja89OI1eHS1W17pXHHGvnMCXehtwGj3UJaOhxD3aYiy/rSX06lr0KJtJ/ddqqjLT1GwyBBN6IyxGNUm2mtTkvinrR7WU9/uJUBSBQHAWTZkvoNZP8IaWXARnI1RapuCp0GHg9auzDryB+ncg1ESS1NnJKyed06zjXt88lEn/CWgJSX+Ry49xr22CN3DsPzZmkwHJ0VZ1Xb3ThjYx6tZ27Zx2dEiwI8ZmbtZvfiIiIqLmKjS12hucEeAJyPAFlHpPOw8tIQgrNBg2W8APt6/+67GjbSZtnXl8lAUJwZ0I4qIsEALhSXs1U+hDxxoys7CZJW3EvWLyHm0NJvSWCsl8sJhcXYoimowGNZE3SbCajLol9nR8ZEVoXwFFgaIAAUWBrAh4ZRllHll7T7qCnU8en6x1xnkDStgsnVAHgiegPkdClAUvjDlT75dZq/rmoRyajXCxdhOsJglurw9GoxGJwaQ+JZjYt41VR+nT4u1IirYyqW9AsiyjuLgYsbGxEbnfZ3PFuOuDcdcH464Pxl0fLT3uoanVqDC1WlHU9ela0cBgh0BtVeBDywecdjM6JVa/hMDjl8sLDVZaPlDgUreUq7wOu9SjjrTmHHWHHRcQEIoCgySFbRNXX6Fp9GFT6CuPxFeYVh9lNcF4Agl4xcQ+lNy3pMS+pb/fG5O680Tod1j32FTXSVDeeaD+a5IEbMIHWZYjKu5M+iOcwWDA45efCl9ZEbp1aAubreZ1XNSwFEWB1+uFoigRddFo7hh3fTDu+mDc9cG46yMS4y5JBm2v9IpC66U9fhneCqORdR1Ft5mN2jLM6gSCSwgKwmoLqJ0EoZ0JwqZJV/rBNU2jrzGZt6pT5Rt6cMlkNJSP0pskWI1GbZ19S0jsaxOJ73e9qZ0FoVhWH1O/348jR0oiLu5M+luBLslROCJ5YDS27IsfERERUWtgNkowGyXE2spnBYSq1ruD05RDU5OPZw2zySghMdqKxGhrtfcLIVDqDaidAiVu+D1lSIx3Itpurdc0+oYQSuxDhfOsxvLp+C09sSdqKkz6iYiIiIiauYrbCVYU0GYFhGYEqLdPZG29wWBAjM2MGJsZ7Z0WuEsE7DFRkIyNkzoYJUNYJfzQOnuLSTqhKf5EpGLST0RERETUQpmMEmKMEmJs5cdCswI8FToDPAEZ/oB+9buNUvlU/MpF9JjYEzUuJv2tgMlkQkJCQpPuuUqMu14Yd30w7vpg3PXBuOuDca+7mmYFyIqoUjTQ45drnRVgkIywRTthkI69vrm6xD6U3DOxrx++3/URqXGPrFdD1TIYDLBYWMCvqTHu+mDc9cG464Nx1wfjrg/G/cQZJbWCfrQ1PAXQCgYGyjsDfAF1r2eDwQCjqby2gCShvBJ+xeTeKMFkDN/zno4f3+/6iNS4M+lvBWRZRklJCWJiYiKqCmVzx7jrg3HXB+OuD8ZdH4y7Phj3xhOaFeBEeXIvKwIevwy31w93WSni4mJht5iZ2DcRvt/1Ealx5//aVkBRFHg8HiiKondTWhXGXR+Muz4Yd30w7vpg3PXBuDcto2RAlNUEp90Es0GGzcSR/KbE97s+IjXu/J9LREREREREFKGY9BMRERERERFFKCb9RERERERERBGKSX8rYDQaERcXF1HFKFoCxl0fjLs+GHd9MO76YNz1wbjrg3HXB+Ouj0iNO6v3twKSJMFms+ndjFaHcdcH464Pxl0fjLs+GHd9MO76YNz1wbjrI1LjzpH+VkCWZRQXF0OWZb2b0qow7vpg3PXBuOuDcdcH464Pxl0fjLs+GHd9RGrcmfS3AoqiwOVyRdzWE80d464Pxl0fjLs+GHd9MO76YNz1wbjrg3HXR6TGnUk/ERERERERUYRi0k9EREREREQUoZj0ExEREREREUUoJv2tgNFohNPpjLitJ5o7xl0fjLs+GHd9MO76YNz1wbjrg3HXB+Ouj0iNO7fsawUkSYLdbte7Ga0O464Pxl0fjLs+GHd9MO76YNz1wbjrg3HXR6TGnSP9rYAsyygpKYm4rSeaO8ZdH4y7Phh3fTDu+mDc9cG464Nx1wfjro9IjTuT/lZAURSUlZVF3NYTzR3jrg/GXR+Muz4Yd30w7vpg3PXBuOuDcddHpMadST8RERERERFRhGLST0RERERERBShmPQTERERERERRSgm/a2A0WhETExMxG090dwx7vpg3PXBuOuDcdcH464Pxl0fjLs+GHd9RGrcuWVfKyBJEqKiovRuRqvDuOuDcdcH464Pxl0fjLs+GHd9MO76YNz1Ealx50h/KxCpVSibO8ZdH4y7Phh3fTDu+mDc9cG464Nx1wfjro9IjTuT/lYgUvebbO4Yd30w7vpg3PXBuOuDcdcH464Pxl0fjLs+IjXuTPqJiIiIiIiIIhSTfiIiIiIiIqIIxUJ+J8Dr9QIAdu7cqXNLauf3+1FUVASn0wmz2ax3c1oNxl0fjLs+GHd9MO76YNz1wbjrg3HXB+Ouj5YS91D+GcpHj4VJ/wnIyckBAFxxxRX6NoSIiIiIiIhalZycHPTu3fuY5xmEEKIJ2hORCgsL8dVXX6FDhw6wWq16N6dGO3fuxBVXXIEPPvgAXbt21bs5rQbjrg/GXR+Muz4Yd30w7vpg3PXBuOuDcddHS4m71+tFTk4OBgwYgLi4uGOez5H+ExAXF4fLL79c72bUWdeuXZGRkaF3M1odxl0fjLs+GHd9MO76YNz1wbjrg3HXB+Ouj5YQ97qM8IewkB8RERERERFRhGLST0RERERERBShmPQTERERERERRSgm/a1AcnIypk+fjuTkZL2b0qow7vpg3PXBuOuDcdcH464Pxl0fjLs+GHd9RGrcWb2fiIiIiIiIKEJxpJ+IiIiIiIgoQjHpJyIiIiIiIopQTPqJiIiIiIiIIhSTfiIiIiIiIqIIxaSfiIiIiIiIKEIx6Y9gpaWlmD59OrKyspCQkACDwYAlS5bo3ayI9uOPP2LKlCnIyMhAVFQUOnbsiKuvvho7duzQu2mtzqxZs2AwGNCzZ0+9mxLRsrOzMXr0aLRv3x4OhwPdu3fHjBkz4HK59G5aRKjPdVxRFLz00ks444wzYLfbkZiYiEGDBmHLli1N2+gI8Mcff2DUqFHo3LkzHA4HkpKScMEFF+Cjjz7SzlEUBUuWLMHw4cPRoUMHREVFoWfPnnj88cfh8Xh0bH3L9ssvv2D48OFISEiAw+FAz5498cILL1R7bmFhIVJSUmAwGLB69eombmnLVJdrSn3f20VFRbj//vtx8sknw263o1OnTpgwYQL27t3bRK+qeavrZ8Prr78eBoOhylf37t2rfd5du3bh2muvRUpKCux2O04++WQ8/PDDTfGSWoT169dXG0+DwYDvv/9eO2/t2rWYMGECevbsCaPRiJNOOqna59u+fTvuv/9+nHHGGYiJiUHbtm0xdOhQ/PTTT030ik6MSe8GUOPJz8/HjBkz0LFjR5x++ulYv3693k2KeE8++SQ2btyIUaNG4bTTTsOBAwewYMEC9O7dG99//z0T0Cayb98+zJ49G1FRUXo3JaLl5OSgf//+cDqdmDJlChISEvDdd99h+vTp+Pnnn/Hhhx/q3cQWrz7X8RtvvBFvv/02rrvuOkyZMgVlZWX49ddfcejQoaZrcIT466+/UFJSgvHjx6Ndu3ZwuVx49913MXz4cLz88suYOHEiXC4XbrjhBpx99tmYPHkyUlJStPf/F198gS+//BIGg0Hvl9KirF27FsOGDcOZZ56JRx55BNHR0di1axf27dtX7fnTpk1jB2M91eWaUp/3tqIouOiii/Dnn3/i1ltvRbdu3bBz504sXLgQn332GbZt24aYmJgmfpXNS30+G1qtVixevDjs8U6ns8pzbt68GQMHDkRaWhruueceJCYmYu/evcjJyWn019PS3HHHHejXr1/Ysa5du2q3ly1bhhUrVqB3795o165djc+zePFivPbaa7jqqqtw6623oqioCC+//DLOPvtsfPrppxg8eHCjvYYGIShieTwekZeXJ4QQ4scffxQAxBtvvKFvoyLcxo0bhdfrDTu2Y8cOYbVaxdixY3VqVetzzTXXiEGDBokBAwaIjIwMvZsTsWbNmiUAiK1bt4Ydv+666wQAcfToUZ1aFjnqeh1fsWKFACDee++9Jm5h6xEIBMTpp58uTjnlFCGEEF6vV2zcuLHKeY899pgAINatW9fUTWzRioqKRJs2bcSIESOELMvHPP/3338XJpNJzJgxQwAQq1ataoJWtnx1uabU5729ceNGAUAsWLAg7NzXX3+d16Sgun42HD9+vIiKijrm88myLHr27CnOOuss4XK5Gry9keI///lPna4Nubm5wufzCSGEGDp0qOjUqVO15/3000+ipKQk7Fh+fr5ITk4W5513XoO0uTFxen8Es1qtSE1N1bsZrcq5554Li8USduzkk09GRkYGtm3bplOrWpevv/4aq1evxnPPPad3UyJecXExAKBNmzZhx9u2bQtJkqr8X6D6q+t1fN68eejfvz9GjBgBRVFQVlbWBK1rXYxGIzp06IDCwkIAgMViwbnnnlvlvBEjRgAAr/n1tGzZMhw8eBCzZs2CJEkoKyuDoig1nn/nnXdixIgRyMzMbMJWtnx1uabU571d298BALDb7SfU3khQ38+Gsixrca3O2rVrsXXrVkyfPh12ux0ulwuyLDd4uyNJSUkJAoFAtfe1a9cOZrP5mM/Rp08fREdHhx1LTExEZmZmi7jeM+knamRCCBw8eBBJSUl6NyXiybKM22+/HTfddBN69eqld3Mi3sCBAwEAEyZMwObNm5GTk4MVK1bgpZdewh133MHlFU2kuLgYP/zwA/r164eHHnoITqcT0dHR6Ny5M1auXKl381q0srIy5OfnY9euXXj22WexZs0a/O1vf6v1MQcOHAAAXvPr6fPPP0dsbCxyc3NxyimnIDo6GrGxsbjllluqrCNftWoVvv32W8ydO1en1rZO1b23+/bti6ioKDzyyCP48ssvkZubi6+++gr3338/+vXr1/ynPOukps+GLpcLsbGxcDqdSEhIwG233YbS0tKwcz7//HMAagdOKP4OhwOjR4/G0aNHm+w1tBQ33HADYmNjYbPZcOGFFzb4GvwDBw60iOs91/QTNbK3334bubm5mDFjht5NiXiLFi3CX3/9pf1BpMaVlZWFmTNnYvbs2fjXv/6lHX/44Yfx+OOP69iy1mXXrl0QQuCdd96ByWTC3Llz4XQ68fzzz2P06NGIjY1FVlaW3s1ske655x68/PLLAABJknDllVdiwYIFtT5m7ty5iI2NxZAhQ5qiiREjOzsbgUAAl19+OSZMmIA5c+Zg/fr1mD9/PgoLC7F8+XIAgNvtxr333ou77roLJ510Evbs2aNvw1uR6t7bSUlJWLFiBW6++eawDrFLLrkEq1evhsnEVKM61X02bNu2Le6//3707t0biqLg008/xcKFC7FlyxasX79ei2V2djYA4Oqrr0ZWVhYefPBBbNmyBXPmzEFOTg6++eYb1hOBOmPlqquuwqWXXoqkpCT8+eefePrpp5GZmYlvv/0WZ5555gn/jA0bNuC7777D1KlTG6DFjUzn5QXURLimXx/btm0TsbGx4pxzzhGBQEDv5kS0/Px8kZCQIJ5++mntGNf0N76lS5eKSy65RLzyyivi3XffFTfeeKMwGAxi/vz5ejct4tR0Hf/6668FAAFAfP/999rxkpISkZSU1CLWGjZX27ZtE+vWrRP//Oc/xdChQ8WIESPEgQMHajw/VOdi4cKFTdjKyNC5c2cBQEyePDns+KRJkwQAsWPHDiGEENOmTRNt27bV1tbWdd0uVVWfz4a1vbc3bdokLr30UjFr1izxwQcfiEcffVQ4HA4xcuTIRmh1y1efz4ahuC9fvlw7NmjQIAFAZGVlhZ07Z84c1hM5huzsbGG328Ull1xS7f21remv7ODBg6J9+/aic+fOVdb6N0dM+lsJJv1NLy8vT3Tu3Fl06NBB5Obm6t2ciDd58mTRtWvXsGI5TPob1/Lly4Xdbhc5OTlhx6+//nrhcDhEfn6+Ti2LTDVdx0PH09PTqzzmhhtuEGazWfj9/iZqZWS76KKLRL9+/YSiKFXue+edd4TBYBATJkzQoWUtX0ZGhgAgvvrqq7DjX331lQAg/vnPf4rdu3cLu90uXn/9de1+Jv3Hr66fDWt7b+/atUs4HA6xevXqsONLliwRAMQnn3zSkE1u8er72dDlcglJksJiP3ToUO3/REV//fWXACAee+yxBm93JBk9erSwWCzVdrjUNekvLS0V/fr1E06nU/z++++N0MqGxzX9RI2gqKgIQ4YMQWFhIT799NNatwChE5ednY1XXnkFd9xxB/bv3489e/Zgz5498Hg88Pv92LNnD9e5NYKFCxfizDPPRPv27cOODx8+HC6XC7/++qtOLWtdQteXyoW0ACAlJQV+v5+F/RrIyJEj8eOPP1bZX3vdunW47rrrMHToUCxatEin1rVsNb2PU1JSAAAFBQWYNm0a0tLSMHDgQO06H1pnfvjwYezZs6fW4n9Uf8d6by9ZsgQejweXXXZZ2PHhw4cDADZu3Ngk7WwJjuezod1uR2JiYthnmLr8X6GadejQAT6f77j/Lvp8Plx55ZX47bff8OGHH7aY7biZ9BM1MI/Hg2HDhmHHjh34+OOPceqpp+rdpIiXm5sLRVFwxx13ID09XfvatGkTduzYgfT0dNZUaAQHDx6stmKw3+8HgBor5VLDateuHVJTU5Gbm1vlvv3798Nms7X6fbIbitvtBqB+eA/ZtGkTRowYgb59+2LlypVcw3yc+vTpAwBV3sf79+8HACQnJ2Pv3r3YuXMnOnfurF3nx4wZAwC49dZbkZ6eXmvVc6qfury3Dx48CCFElb8F/DsQ7ng/G5aUlCA/Px/Jycnasbr8X6Ga/e9//4PNZqtSib8uFEXBddddhy+++ALLli3DgAEDGqGFjYNJP1EDkmUZ11xzDb777jusWrUK55xzjt5NahV69uyJ999/v8pXRkYGOnbsiPfffx8TJkzQu5kRp1u3bvj111+rjHouX74ckiThtNNO06llrc8111yDnJwcrFu3TjuWn5+PDz/8EIMGDYIk8c99fRw6dKjKMb/fjzfffBN2u137wL5t2zYMHToUJ510Ej7++GNuT3YCrr76agDAa6+9FnZ88eLFMJlMGDhwIB5//PEq1/mZM2cCAO6//368//773DWkgdT1vd2tWzcIIarsFBIqvNgQxdJaurp8NvR4PCgpKalyfObMmRBChBVjvfzyy2G1WvHGG2+EzWxZvHgxAOCiiy5qhFfR8hw+fLjKsS1btuBf//oXLr744uP6u3j77bdjxYoVWLhwIa688sqGaGaTYXd0hFuwYAEKCwu13r+PPvoI+/btA6C+cZ1Op57Nizj33HMP/vWvf2HYsGE4evQo3nrrrbD7x40bp1PLIltSUhKuuOKKKsefe+45AKj2Pjpx9913H9asWYPMzExMmTIFiYmJ+Pjjj7FmzRrcdNNNXNbSQOpyHX/wwQexcuVKXHXVVbj77rvhdDqxaNEi+P1+zJ49W8/mt0iTJk1CcXExLrjgAqSlpeHAgQN4++23sX37djzzzDOIjo5GSUkJLrnkEhQUFOC+++7Dv//977Dn6NKlCzt+6+HMM8/EjTfeiNdffx2BQAADBgzA+vXrsWrVKjz44INo165dtdeUuLg4AEC/fv14ra+jY11TJEmq83v7+uuvx9NPP41Jkybh119/RUZGBn755RcsXrwYGRkZGDFiRNO+uGaoLp8NDxw4gDPPPBNjxoxB9+7dAQCfffYZPvnkE2RlZeHyyy/Xzk9NTcXDDz+MadOmISsrC1dccQW2bNmCV199FWPGjEG/fv2a9PU1V9dccw3sdjvOPfdcpKSk4M8//8Qrr7wCh8OBJ554Qjvvt99+03Yg2rlzJ4qKirQdiE4//XQMGzYMgPqZcuHChTjnnHPgcDiq/B5HjBjRvDsdda4pQI2sU6dOWlXnyl+7d+/Wu3kRZ8CAATXGm//dmh4L+TW+TZs2iSFDhojU1FRhNptFt27dxKxZs1g4rgHV9Tq+a9cuMWLECBEbGyvsdrsYNGiQ+OGHH/RreAu2fPlyMXjwYNGmTRthMplEfHy8GDx4sPjwww+1c3bv3l3r9X78+PH6vYAWyufziUcffVR06tRJmM1m0bVrV/Hss8/W+hgW8qu/Y11T6vve3rdvn7jxxhtFenq6sFgsom3btuLmm28Whw8f1ucFNjN1+WxYUFAgxo0bJ7p27SocDoewWq0iIyNDzJ49W/h8virPqSiKmD9/vujWrZswm82iQ4cOYurUqdWe21o9//zzon///iIhIUGYTCbRtm1bMW7cOJGdnR123htvvFGn9/r48eNr/T0297zKIIQQDdN9QERERERERETNCRf5EREREREREUUoJv1EREREREREEYpJPxEREREREVGEYtJPREREREREFKGY9BMRERERERFFKCb9RERERERERBGKST8RERERERFRhGLST0RERERERBShmPQTERERERERRSgm/UREREREREQRikk/ERERNaolS5bAYDDgp59+avSfZTAY8Oijjzb6zyEiImopmPQTERFFkFCCXfErJSUFF154IdasWXPczzt79mx88MEHDdfQevrmm28wZMgQpKWlwWazoWPHjhg2bBiWLVumW5uIiIhaApPeDSAiIqKGN2PGDKSnp0MIgYMHD2LJkiW49NJL8dFHH+Gyyy6r9/PNnj0bI0eOxBVXXNHwjT2GVatW4ZprrsEZZ5yBO++8E/Hx8di9eze+/vprvPrqq7j22mu1c91uN0wmfrwhIiIK4V9FIiKiCDRkyBD07dtX+37ChAlo06YNli9fflxJv54effRRnHrqqfj+++9hsVjC7jt06FDY9zabrSmbRkRE1Oxxej8REVErEBcXB7vdXmUU/Omnn8a5556LxMRE2O129OnTB6tXrw47x2AwoKysDP/85z+1JQPXX3+9dn9ubi4mTJiAdu3awWq1Ij09Hbfccgt8Pl/Y83i9Xtx9991ITk5GVFQURowYgcOHDx+z7bt27UK/fv2qJPwAkJKSUqWtoTX9e/bsqbLUoeJXRZs2bUJWVhacTiccDgcGDBiAjRs3HrNtREREzR1H+omIiCJQUVER8vPzIYTAoUOHMH/+fJSWlmLcuHFh5z3//PMYPnw4xo4dC5/Ph3feeQejRo3Cxx9/jKFDhwIAli5diptuugn9+/fHxIkTAQBdunQBAOzfvx/9+/dHYWEhJk6ciO7duyM3NxerV6+Gy+UKS9Rvv/12xMfHY/r06dizZw+ee+45TJkyBStWrKj1tXTq1AlffPEF9u3bh/bt29c5BsnJyVi6dGnYMb/fj7vuuiusXV9++SWGDBmCPn36YPr06ZAkCW+88QYGDRqEDRs2oH///nX+mURERM0Nk34iIqIINHjw4LDvrVYrXn/9dVx00UVhx3fs2AG73a59P2XKFPTu3Rvz5s3Tkv5x48Zh8uTJ6Ny5c5VOgwcffBAHDhzApk2bwpYTzJgxA0KIsHMTExOxdu1abZRdURS88MILKCoqgtPprPG1PPDAA5gwYQK6dOmC8847D+effz4uvvhinHvuuZCkmictRkVFVWnvbbfdhtLSUqxbtw4AIITA5MmTtUKHobZNmjQJGRkZmDp1KtauXVvjzyAiImruOL2fiIgoAr344otYt24d1q1bh7feegsXXnghbrrpJrz33nth51VM+AsKClBUVITMzEz88ssvx/wZiqLggw8+wLBhw8IS/pDKU+gnTpwYdiwzMxOyLOOvv/6q9efceOON+PTTTzFw4EB88803mDlzJjIzM3HyySfj22+/PWY7Q958800sXLgQc+fOxYUXXggA2Lx5M7Kzs3HttdfiyJEjyM/PR35+PsrKyvC3v/0NX3/9NRRFqfPPICIiam440k9ERBSB+vfvH5aIjxkzBmeeeSamTJmCyy67TJve/vHHH+Pxxx/H5s2b4fV6tfMrJ+zVOXz4MIqLi9GzZ886taljx45h38fHxwNQOxuO5ZJLLsEll1wCl8uFn3/+GStWrMCiRYtw2WWXYfv27VXW9le2efNmTJ48GWPGjMHdd9+tHc/OzgYAjB8/vsbHFhUVaW0lIiJqaZj0ExERtQKSJOHCCy/E888/j+zsbGRkZGDDhg0YPnw4LrjgAixcuBBt27aF2WzGG2+8gWXLljV4G4xGY7XHKy8DqI3D4UBmZiYyMzORlJSExx57DGvWrKk1aS8oKMBVV12Fbt26YfHixWH3hUbxn3rqKZxxxhnVPj46OrrO7SMiImpumPQTERG1EoFAAABQWloKAHj33Xdhs9nw2WefwWq1aue98cYbVR5b3ch/cnIyYmNjsXXr1kZqce1CMxny8vJqPEdRFIwdOxaFhYX4/PPP4XA4wu4PFSSMjY2tUgeBiIgoEnBNPxERUSvg9/uxdu1aWCwW9OjRA4A68m4wGCDLsnbenj178MEHH1R5fFRUFAoLC8OOSZKEK664Ah999BF++umnKo+pzwh+bb744otqj3/yyScAgFNOOaXGxz722GP47LPPsHz5cqSnp1e5v0+fPujSpQuefvpprTOkorpsKUhERNSccaSfiIgoAq1Zswbbt28HABw6dAjLli1DdnY2/u///g+xsbEAgKFDh2LevHnIysrCtddei0OHDuHFF19E165d8dtvv4U9X58+ffD5559j3rx5aNeuHdLT03HWWWdh9uzZWLt2LQYMGICJEyeiR48eyMvLw6pVq/DNN98gLi7uhF/L5ZdfjvT0dAwbNgxdunRBWVkZPv/8c3z00Ufo168fhg0bVu3jfv/9d8ycORMXXHABDh06hLfeeivs/nHjxkGSJCxevBhDhgxBRkYGbrjhBqSlpSE3Nxf/+c9/EBsbi48++uiEXwMREZFemPQTERFFoGnTpmm3bTYbunfvjpdeegmTJk3Sjg8aNAivvfYannjiCfzjH/9Aeno6nnzySezZs6dK0j9v3jxMnDgRU6dOhdvtxvjx43HWWWchLS0NmzZtwiOPPIK3334bxcXFSEtLw5AhQ6pMpT9eixcvxocffoiVK1di//79EEKgc+fOePjhh/HAAw/AZKr+48yRI0cghMBXX32Fr776qsr9oe38Bg4ciO+++w4zZ87EggULUFpaitTUVJx11llh8SIiImqJDKKh5t4RERERERERUbPCNf1EREREREREEYpJPxEREREREVGEYtJPREREREREFKGY9BMRERERERFFKCb9RERERERERBGKST8RERERERFRhGLST0RERERERBShmPQTERERERERRSgm/UREREREREQRikk/ERERERERUYRi0k9EREREREQUoZj0ExEREREREUUoJv1EREREREREEYpJPxEREREREVGE+n9xzzyHJsDfNAAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1200x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "def plot_throughputs(perf_metrics):\n",
    "    plt.figure(figsize=(10, 6), dpi=120)\n",
    "    maximums = []\n",
    "    for i, (label, metric) in enumerate(perf_metrics.items()):\n",
    "        xs = np.arange(len(metric[\"batch_sizes\"]))\n",
    "        y_mean, y_std = np.array(metric[\"throughputs_avg\"]), np.array(metric[\"throughputs_std\"])\n",
    "        maximums.append((np.argmax(y_mean), max(y_mean)))\n",
    "        plt.plot(xs, y_mean, label=label, alpha=0.8)\n",
    "        plt.fill_between(xs, y_mean - y_std, y_mean + y_std, alpha=0.2)\n",
    "        for x, y in zip(xs, y_mean):\n",
    "            plt.text(x, y - 40 * (i % 2), f'{y:.1f}', ha='right', va='bottom', fontsize=7, color=\"black\")\n",
    "\n",
    "    plt.suptitle(\"Throughput vs Batch Size\")\n",
    "    plt.title('\\nBAAI/bge-small-en-v1.5 SetFit model trained on SetFit/sst2', fontsize=8)\n",
    "    plt.ylabel(\"Sentences / Second\")\n",
    "    plt.xlabel(\"Batch Size\")\n",
    "    plt.xticks(xs, pb.batch_sizes[:len(xs)])\n",
    "    plt.legend()\n",
    "    plt.grid(True, linestyle='--', alpha=0.3)\n",
    "    \n",
    "    # Add Speedup Arrow\n",
    "    max_opt_thoughput = max(perf_metrics['bge-small (optimum-int8)']['throughputs_avg'])\n",
    "    max_transformers_thoughput = max(perf_metrics['bge-small (transformers)']['throughputs_avg'])\n",
    "    speedup =  max_opt_thoughput / max_transformers_thoughput\n",
    "    arrow_kwargs = dict(head_width = 0.6, width=0.15, head_length=60, color=\"green\", alpha=0.6)\n",
    "    plt.arrow(6.3, 200, 0, 500, **arrow_kwargs)\n",
    "    plt.text(6.5, 340, f'Max\\nThroughput\\nSpeedup={speedup:.1f}x', color=\"green\", fontsize=14)\n",
    "    \n",
    "    plt.savefig(\"throughput_vs_batch_size.png\")\n",
    "    plt.show()\n",
    "    \n",
    "plot_throughputs(perf_metrics)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "777c6ab3-0d71-453a-9af3-67f8947fa6ce",
   "metadata": {},
   "source": [
    "#### The throughput is where our optimization shines - the optimized model achieves 7.8x speedup in max throughput!"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "gpuType": "T4",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "setfit",
   "language": "python",
   "name": "setfit"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}