soft_prompt_tuning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_name = \"bigscience/bloomz-560m\"\n",
    "#model_name=\"bigscience/bloom-1b1\"\n",
    "# NUM_VIRTUAL_TOKENS = 4\n",
    "NUM_EPOCHS = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0,1\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/xzrderek/.local/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "from transformers import AutoModelForCausalLM, AutoTokenizer\n",
    "\n",
    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
    "foundational_model = AutoModelForCausalLM.from_pretrained(\n",
    "    model_name,\n",
    "    trust_remote_code=True\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_outputs(model, inputs, max_new_tokens=100):\n",
    "    outputs = model.generate(\n",
    "        input_ids=inputs[\"input_ids\"],\n",
    "        attention_mask=inputs[\"attention_mask\"],\n",
    "        max_new_tokens=max_new_tokens,\n",
    "        repetition_penalty=1.5, #Avoid repetition.\n",
    "        early_stopping=True, #The model can stop before reach the max_length\n",
    "        eos_token_id=tokenizer.eos_token_id\n",
    "    )\n",
    "    return outputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset({\n",
       "    features: ['Lines', 'input_ids', 'attention_mask'],\n",
       "    num_rows: 25\n",
       "})"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "dataset_sentences = load_dataset('csv', data_files='data/spongebob_lines.csv')\n",
    "\n",
    "data_sentences = dataset_sentences.map(lambda samples: tokenizer(samples[\"Lines\"]), batched=True)\n",
    "data_sentences = data_sentences['train'].train_test_split(test_size=0.2)\n",
    "\n",
    "train_sample_sentences = data_sentences[\"train\"].select(range(25))\n",
    "test_sample_sentences = data_sentences[\"test\"].select(range(25))\n",
    "# train_sample_sentences = train_sample_sentences.remove_columns(['author', 'tags'])\n",
    "display(train_sample_sentences)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Change Hyperparameters Here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from peft import  get_peft_model, PromptTuningConfig, TaskType, PromptTuningInit\n",
    "\n",
    "tuning_init = [PromptTuningInit.RANDOM, PromptTuningInit.TEXT]\n",
    "virtual_tokens = [2,4,8,16,32]\n",
    "\n",
    "\n",
    "chosen_prompt_tuning_init, chosen_num_virtual_tokens = tuning_init[1], virtual_tokens[4]\n",
    "\n",
    "generation_config = PromptTuningConfig(\n",
    "    task_type=TaskType.CAUSAL_LM,\n",
    "    prompt_tuning_init=chosen_prompt_tuning_init,\n",
    "    num_virtual_tokens=chosen_num_virtual_tokens,\n",
    "    prompt_tuning_init_text=\"Imagine you are Spongebob Squarepants from the Television show Spongebob\",\n",
    "    tokenizer_name_or_path=model_name\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainable params: 32,768 || all params: 559,247,360 || trainable%: 0.005859303475299373\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "peft_model_sentences = get_peft_model(foundational_model, generation_config)\n",
    "print(peft_model_sentences.print_trainable_parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import TrainingArguments\n",
    "def create_training_arguments(learning_rate=0.0035, epochs=6):\n",
    "    training_args = TrainingArguments(\n",
    "        output_dir=\"outputs\",\n",
    "        # use_cpu=True, # This is necessary for CPU clusters.\n",
    "        auto_find_batch_size=True, # Find a suitable batch size that will fit into memory automatically\n",
    "        learning_rate= learning_rate, # Higher learning rate than full fine-tuning\n",
    "        logging_steps=1,\n",
    "        max_steps=20,\n",
    "        num_train_epochs=epochs\n",
    "    )\n",
    "    return training_args"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_args_sentences = create_training_arguments(0.003, NUM_EPOCHS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import Trainer, DataCollatorForLanguageModeling\n",
    "def create_trainer(model, training_args, train_dataset):\n",
    "    trainer = Trainer(\n",
    "        model=model,\n",
    "        args=training_args,\n",
    "        train_dataset=train_dataset, \n",
    "        data_collator=DataCollatorForLanguageModeling(tokenizer, mlm=False) \n",
    "    )\n",
    "    return trainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Detected kernel version 5.4.0, which is below the recommended minimum of 5.5.0; this can cause the process to hang. It is recommended to upgrade the kernel to the minimum version or higher.\n",
      "You're using a BloomTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.\n",
      "/home/xzrderek/.local/lib/python3.11/site-packages/torch/nn/parallel/_functions.py:68: UserWarning: Was asked to gather along dimension 0, but all input tensors were scalars; will instead unsqueeze and return a vector.\n",
      "  warnings.warn('Was asked to gather along dimension 0, but all '\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "    <div>\n",
       "      \n",
       "      <progress value='20' max='20' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
       "      [20/20 00:11, Epoch 10/10]\n",
       "    </div>\n",
       "    <table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       " <tr style=\"text-align: left;\">\n",
       "      <th>Step</th>\n",
       "      <th>Training Loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>5.397900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.679300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>4.552100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>4.244800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>4.121400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>4.124600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>3.826400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>4.137000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>3.897600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>3.714300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>3.817500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>3.548500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>3.616100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>3.876700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>3.651500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>3.524800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>3.599600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>3.601300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>3.547500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>3.395000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table><p>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Detected kernel version 5.4.0, which is below the recommended minimum of 5.5.0; this can cause the process to hang. It is recommended to upgrade the kernel to the minimum version or higher.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "\n",
       "    <div>\n",
       "      \n",
       "      <progress value='20' max='20' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
       "      [20/20 00:12, Epoch 10/10]\n",
       "    </div>\n",
       "    <table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       " <tr style=\"text-align: left;\">\n",
       "      <th>Step</th>\n",
       "      <th>Training Loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>3.359900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>4.050800</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>3.200900</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>3.657200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>5</td>\n",
       "      <td>3.299600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>6</td>\n",
       "      <td>3.014200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>7</td>\n",
       "      <td>3.264600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>8</td>\n",
       "      <td>2.749000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>9</td>\n",
       "      <td>2.797700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>10</td>\n",
       "      <td>3.174000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>11</td>\n",
       "      <td>2.717200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>12</td>\n",
       "      <td>3.347600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>13</td>\n",
       "      <td>2.587200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>14</td>\n",
       "      <td>3.118500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>15</td>\n",
       "      <td>2.665600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>16</td>\n",
       "      <td>2.569600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>17</td>\n",
       "      <td>2.507600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>18</td>\n",
       "      <td>2.891000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>19</td>\n",
       "      <td>2.888200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>20</td>\n",
       "      <td>2.107000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table><p>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "TrainOutput(global_step=20, training_loss=2.9983570337295533, metrics={'train_runtime': 12.9927, 'train_samples_per_second': 24.629, 'train_steps_per_second': 1.539, 'total_flos': 16831000240128.0, 'train_loss': 2.9983570337295533, 'epoch': 10.0})"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainer_sentences = create_trainer(peft_model_sentences, training_args_sentences, train_sample_sentences)\n",
    "trainer_sentences.train()\n",
    "\n",
    "test_sentences = create_trainer(peft_model_sentences, training_args_sentences, test_sample_sentences)\n",
    "test_sentences.train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No artists with labels found to put in legend.  Note that artists whose label start with an underscore are ignored when legend() is called with no argument.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "prompt_tuning_init=PromptTuningInit.TEXT, num_virtual_tokens=32 \n",
      "[5.3979, 4.6793, 4.5521, 4.2448, 4.1214, 4.1246, 3.8264, 4.137, 3.8976, 3.7143, 3.8175, 3.5485, 3.6161, 3.8767, 3.6515, 3.5248, 3.5996, 3.6013, 3.5475, 3.395]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No artists with labels found to put in legend.  Note that artists whose label start with an underscore are ignored when legend() is called with no argument.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "prompt_tuning_init=PromptTuningInit.TEXT, num_virtual_tokens=32 \n",
      "[3.3599, 4.0508, 3.2009, 3.6572, 3.2996, 3.0142, 3.2646, 2.749, 2.7977, 3.174, 2.7172, 3.3476, 2.5872, 3.1185, 2.6656, 2.5696, 2.5076, 2.891, 2.8882, 2.107]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAj8AAAHHCAYAAABQhTneAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAAB6N0lEQVR4nO3dd1xTV/8H8E/YG3EgIFRRwS1WHMXdiuKoSqtVW61atVpXx6+tLX3aqq0WtXs9apcdttVq66h1oRUn7oWzDtyidQCCCgjn98d5biAQIIEkNyGf9+uVV25uzr05NwHy5ZzvOUcjhBAgIiIishMOaleAiIiIyJIY/BAREZFdYfBDREREdoXBDxEREdkVBj9ERERkVxj8EBERkV1h8ENERER2hcEPERER2RUGP0RERGRXGPyQzXr//fdRt25dODo6okWLFmpXhww0YsQI1KlTR+1qWL0uXbqgS5cualeDypCYmAiNRoMlS5aoXRUyAoMfsojk5GQMGDAAtWvXhpubG2rVqoVu3brh888/L9f51q1bh8mTJ6N9+/aYP38+3nvvPVy+fBlTp07FgQMHDDrH999/D41Go725ubkhPDwcEydOxNWrV8tVL7UZ8x4UvvbSbomJiWavt5ru3LmDqVOnmv06V61ahalTp5r1NYjIME5qV4Aqv+3bt+Phhx/GAw88gGeffRYBAQG4cOECduzYgU8//RSTJk0y+px///03HBwc8O2338LFxQUAsGfPHkybNg116tQxqiXonXfeQWhoKO7du4etW7dizpw5WLVqFQ4fPgwPDw+j66amy5cvG/we/PTTTzqPf/zxRyQkJBTb36hRI5PW8euvv0Z+fr5Jz1kRd+7cwbRp0wDArC0tq1atwpdffskAiMgKMPghs5sxYwZ8fX2xe/duVKlSRee5a9euleuc165dg7u7uzbwqYiePXuiVatWAIDRo0ejWrVq+Oijj7B8+XI8+eSTeo/JysqCp6dnhV9bTUOHDtV5vGPHDiQkJBTbb2rOzs5mPT8RUVnY7UVmd/r0aTRp0qRY4AMA/v7+Oo/v37+Pd999F/Xq1YOrqyvq1KmDN954A9nZ2doyGo0G8+fPR1ZWlrZr5vvvv0fr1q0BAM8884zOfmM98sgjAICUlBQAMkfFy8sLp0+fRq9eveDt7Y0hQ4YAkEHQyy+/jJCQELi6uqJBgwb44IMPIITQOadGo8HEiROxePFiNG7cGO7u7oiKikJycjIAYN68eahfvz7c3NzQpUsXnD17Vuf4Ll26oGnTpti7dy/atWsHd3d3hIaGYu7cudoyiYmJJnsPFHXq1MGIESOK7S+aj6LkPfz222+YMWMGgoOD4ebmhq5du+LUqVM6xxbN+Tl79iw0Gg0++OADfPXVV9rPvnXr1ti9e3ex11beQzc3NzRt2hRLly7Vm0d05coVHD9+HLm5uSVe39mzZ1GjRg0AwLRp07TvWeHWmePHj2PAgAGoWrUq3Nzc0KpVK6xYsULnPLm5uZg2bRrCwsLg5uaGatWqoUOHDkhISNBe85dffglAt7vRWNeuXcOoUaNQs2ZNuLm5ISIiAj/88EOxcgsXLkRkZCS8vb3h4+ODZs2a4dNPPzW4vvrs2bMHGo1G7+utXbsWGo0GK1euBADcvn0bL774IurUqQNXV1f4+/ujW7du2Ldvn9HXDADZ2dmYMmUK6tevD1dXV4SEhGDy5Mk6fxeAgt+zn3/+GQ0aNICbmxsiIyOxefPmYufcv38/evbsCR8fH3h5eaFr167YsWNHsXJpaWl46aWXtNcSHByMYcOG4fr16zrl8vPzy/zZP3nyJPr374+AgAC4ubkhODgYgwcPRnp6erneFyo/tvyQ2dWuXRtJSUk4fPgwmjZtWmrZ0aNH44cffsCAAQPw8ssvY+fOnYiPj8exY8ewdOlSALK75quvvsKuXbvwzTffAADCwsLwzjvv4O2338aYMWPQsWNHAEC7du2Mru/p06cBANWqVdPuu3//PmJiYtChQwd88MEH8PDwgBACffv2xcaNGzFq1Ci0aNECa9euxauvvopLly7h448/1jnvli1bsGLFCkyYMAEAEB8fj0cffRSTJ0/Gf//7X4wfPx63bt3C7NmzMXLkSPz99986x9+6dQu9evXCwIED8eSTT+K3337DuHHj4OLigpEjR6JRo0Ymew/Ka+bMmXBwcMArr7yC9PR0zJ49G0OGDMHOnTvLPPaXX37B7du3MXbsWGg0GsyePRuPP/44zpw5o20t+uuvvzBo0CA0a9YM8fHxuHXrFkaNGoVatWoVO19cXBx++OEHpKSklJhgXaNGDcyZMwfjxo3DY489hscffxwA0Lx5cwDAkSNH0L59e9SqVQuvv/46PD098dtvvyE2Nha///47HnvsMQDA1KlTER8fj9GjR6NNmzbIyMjAnj17sG/fPnTr1g1jx47F5cuX9XYrGuru3bvo0qULTp06hYkTJyI0NBSLFy/GiBEjkJaWhhdeeAEAkJCQgCeffBJdu3bFrFmzAADHjh3Dtm3btGXKqq8+rVq1Qt26dfHbb79h+PDhOs8tWrQIfn5+iImJAQA899xzWLJkCSZOnIjGjRvjxo0b2Lp1K44dO4aWLVsadd35+fno27cvtm7dijFjxqBRo0ZITk7Gxx9/jH/++QfLli3TKb9p0yYsWrQIzz//PFxdXfHf//4XPXr0wK5du7R/f44cOYKOHTvCx8cHkydPhrOzM+bNm4cuXbpg06ZNaNu2LQAgMzMTHTt2xLFjxzBy5Ei0bNkS169fx4oVK3Dx4kVUr15d+7pl/ezn5OQgJiYG2dnZmDRpEgICAnDp0iWsXLkSaWlp8PX1Nep9oQoSRGa2bt064ejoKBwdHUVUVJSYPHmyWLt2rcjJydEpd+DAAQFAjB49Wmf/K6+8IgCIv//+W7tv+PDhwtPTU6fc7t27BQAxf/58g+o1f/58AUCsX79e/Pvvv+LChQti4cKFolq1asLd3V1cvHhR+1oAxOuvv65z/LJlywQAMX36dJ39AwYMEBqNRpw6dUq7D4BwdXUVKSkp2n3z5s0TAERAQIDIyMjQ7o+LixMAdMp27txZABAffvihdl92drZo0aKF8Pf3176Xxr4HhU2YMEEU/ZNQu3ZtMXz48GJlO3fuLDp37qx9vHHjRgFANGrUSGRnZ2v3f/rppwKASE5O1u4bPny4qF27tvZxSkqKACCqVasmbt68qd2/fPlyAUD8+eef2n3NmjUTwcHB4vbt29p9iYmJAoDOOZXXKfo+6vPvv/8KAGLKlCnFnuvatato1qyZuHfvnnZffn6+aNeunQgLC9Pui4iIEL179y71dfS9v6Up+h5/8sknAoBYsGCBdl9OTo6IiooSXl5e2p+hF154Qfj4+Ij79++XeG5D6qtPXFyccHZ21vmcsrOzRZUqVcTIkSO1+3x9fcWECROMPr8+P/30k3BwcBBbtmzR2T937lwBQGzbtk27D4AAIPbs2aPdd+7cOeHm5iYee+wx7b7Y2Fjh4uIiTp8+rd13+fJl4e3tLTp16qTd9/bbbwsA4o8//ihWr/z8fCGE4T/7+/fvFwDE4sWLy/tWkAmx24vMrlu3bkhKSkLfvn1x8OBBzJ49GzExMahVq5ZO98GqVasAAP/3f/+nc/zLL78MQP7Xbw7R0dGoUaMGQkJCMHjwYHh5eWHp0qXFWhPGjRun83jVqlVwdHTE888/X6y+QgisXr1aZ3/Xrl11WiCU/y779+8Pb2/vYvvPnDmjc7yTkxPGjh2rfezi4oKxY8fi2rVr2Lt3r5FXbR7PPPOMTh6W0vpU9Fr0GTRoEPz8/Eo89vLly0hOTsawYcPg5eWlLde5c2c0a9as2Pm+//57CCHKPaz+5s2b+PvvvzFw4EDcvn0b169fx/Xr13Hjxg3ExMTg5MmTuHTpEgCgSpUqOHLkCE6ePFmu1zLEqlWrEBAQoJOH5uzsjOeffx6ZmZnYtGmTti5ZWVmldmGVt76DBg1Cbm4u/vjjD+2+devWIS0tDYMGDdI5/86dO3H58mWjzq/P4sWL0ahRIzRs2FD7GVy/fl3bPb1x40ad8lFRUYiMjNQ+fuCBB9CvXz+sXbsWeXl5yMvLw7p16xAbG4u6detqywUGBuKpp57C1q1bkZGRAQD4/fffERERoW3hK6xot2VZP/tKy87atWtx586dcr8fZBoMfsgiWrdujT/++AO3bt3Crl27EBcXh9u3b2PAgAE4evQoAODcuXNwcHBA/fr1dY4NCAhAlSpVcO7cObPU7csvv0RCQgI2btyIo0eP4syZM9rme4WTkxOCg4N19p07dw5BQUE6gQtQMDqqaH0feOABncfKH8OQkBC9+2/duqWzPygoqFiSdXh4OAAUyxFSS9FrVIKZotdSnmOV97Poz0dJ+yrq1KlTEELgrbfeQo0aNXRuU6ZMAVCQsP/OO+8gLS0N4eHhaNasGV599VUcOnTIpPU5d+4cwsLC4OCg+2e76M/b+PHjER4ejp49eyI4OBgjR47EmjVrdI4pb30jIiLQsGFDLFq0SLtv0aJFqF69ujYYAYDZs2fj8OHDCAkJQZs2bTB16lSDAmB9Tp48iSNHjhT7DJSf/aKDJsLCwoqdIzw8HHfu3MG///6Lf//9F3fu3EGDBg2KlWvUqBHy8/Nx4cIFALILvKyuekVZP7+hoaH4v//7P3zzzTeoXr06YmJi8OWXXzLfRyUMfsiiXFxc0Lp1a7z33nuYM2cOcnNzsXjxYp0y5UkErYg2bdogOjoaXbp0QaNGjYp9uQCAq6ur3v3GcHR0NGq/KJI0rYaSPou8vDy9+ytyLdb2PijD8V955RUkJCTovSlBV6dOnXD69Gl89913aNq0Kb755hu0bNlSm5NmSf7+/jhw4ABWrFihzUnr2bOnTp5OReo7aNAgbNy4EdevX0d2djZWrFiB/v37w8mpIIV04MCBOHPmDD7//HMEBQXh/fffR5MmTYq1hhoiPz8fzZo1K/EzGD9+vNHnNAdDfn4//PBDHDp0CG+88Qbu3r2L559/Hk2aNMHFixctVU36HwY/pBplePmVK1cAyMTo/Pz8Yk3xV69eRVpaGmrXrl3q+SwdNNWuXRuXL1/G7du3dfYfP35c+7wpXb58GVlZWTr7/vnnHwDQdu2Y+j3w8/NDWlpasf3maoUrjfJ+Fh1BU9I+Q5X0nildIs7OzoiOjtZ7K9zqV7VqVTzzzDP49ddfceHCBTRv3lxn1FhFP5vatWvj5MmTxeZI0vfz5uLigj59+uC///0vTp8+jbFjx+LHH3/UeZ/Kqm9JBg0ahPv37+P333/H6tWrkZGRgcGDBxcrFxgYiPHjx2PZsmVISUlBtWrVMGPGDKOvu169erh58ya6du2q9zMo2oKjryvvn3/+gYeHh7bVyMPDAydOnChW7vjx43BwcNC2xtarVw+HDx82us6ladasGd58801s3rwZW7ZswaVLl3RGbZJlMPghs9u4caPe/96VHB/lj1evXr0AAJ988olOuY8++ggA0Lt371JfR+kS0vdlbQ69evVCXl4evvjiC539H3/8MTQaDXr27GnS17t//z7mzZunfZyTk4N58+ahRo0a2hwHU78H9erVw44dO5CTk6Pdt3LlSm23gCUFBQWhadOm+PHHH5GZmandv2nTJu2UAYUZMtQdgHYiy6Lvmb+/P7p06YJ58+ZpA/TC/v33X+32jRs3dJ7z8vJC/fr1dYZiV/Sz6dWrF1JTU3W6nO7fv4/PP/8cXl5e6Ny5s966ODg4aEevKfUxpL4ladSoEZo1a4ZFixZh0aJFCAwMRKdOnbTP5+XlFevK8ff3R1BQkM75r1+/juPHj5eZ/zJw4EBcunQJX3/9dbHn7t69W+wfgqSkJJ0h9RcuXMDy5cvRvXt3ODo6wtHREd27d8fy5ct1uouvXr2KX375BR06dICPjw8AmY938OBB7UjTwoxtkczIyMD9+/d19jVr1gwODg4Gve9kWhzqTmY3adIk3LlzB4899hgaNmyInJwcbN++HYsWLUKdOnXwzDPPAJD5BMOHD8dXX32FtLQ0dO7cGbt27cIPP/yA2NhYPPzww6W+Tr169VClShXMnTsX3t7e8PT0RNu2bREaGmqW6+rTpw8efvhh/Oc//8HZs2cRERGBdevWYfny5XjxxRdRr149k75eUFAQZs2ahbNnzyI8PByLFi3CgQMH8NVXX2mHgpv6PRg9ejSWLFmCHj16YODAgTh9+jQWLFhg8msz1HvvvYd+/fqhffv2eOaZZ3Dr1i188cUXaNq0qU5ABBg21B0A3N3d0bhxYyxatAjh4eGoWrUqmjZtiqZNm+LLL79Ehw4d0KxZMzz77LOoW7curl69iqSkJFy8eBEHDx4EADRu3BhdunRBZGQkqlatij179miHeiuUAPX5559HTEwMHB0d9baYlGTMmDGYN28eRowYgb1796JOnTpYsmQJtm3bhk8++UTbCjV69GjcvHkTjzzyCIKDg3Hu3Dl8/vnnaNGihTY/yJD6lmbQoEF4++234ebmhlGjRul0Cd++fRvBwcEYMGAAIiIi4OXlhfXr12P37t348MMPteW++OILTJs2DRs3bix1Zu2nn34av/32G5577jls3LgR7du3R15eHo4fP47ffvsNa9eu1bYiA0DTpk0RExOjM9QdgHYWbwCYPn06EhIS0KFDB4wfPx5OTk6YN28esrOzMXv2bG25V199FUuWLMETTzyBkSNHIjIyEjdv3sSKFSswd+5cREREGPR+AXJW+okTJ+KJJ55AeHg47t+/j59++gmOjo7o37+/wechE1FtnBnZjdWrV4uRI0eKhg0bCi8vL+Hi4iLq168vJk2aJK5evapTNjc3V0ybNk2EhoYKZ2dnERISIuLi4nSGGguhf6i7EHJ4dOPGjYWTk1OZQ76Voe67d+8utf4lvZYQQty+fVu89NJLIigoSDg7O4uwsDDx/vvva4fBKgAUG/qrDPF+//33dfYrQ2cLD4nt3LmzaNKkidizZ4+IiooSbm5uonbt2uKLL76o0HtQWElDsT/88ENRq1Yt4erqKtq3by/27NlT4lD3osN4lWssXIeShroXfR+EEHqHoC9cuFA0bNhQuLq6iqZNm4oVK1aI/v37i4YNG+qUM3SouxBCbN++XURGRgoXF5dir3n69GkxbNgwERAQIJydnUWtWrXEo48+KpYsWaItM336dNGmTRtRpUoV4e7uLho2bChmzJihM53D/fv3xaRJk0SNGjWERqMpc9h70fdYCCGuXr0qnnnmGVG9enXh4uIimjVrVuzzXbJkiejevbvw9/cXLi4u4oEHHhBjx44VV65cMaq+pTl58qR2WPnWrVt1nsvOzhavvvqqiIiIEN7e3sLT01NERESI//73vzrlpkyZIgCIjRs3lvl6OTk5YtasWaJJkybC1dVV+Pn5icjISDFt2jSRnp6uLaf8ni1YsECEhYUJV1dX8eCDD+p9jX379omYmBjh5eUlPDw8xMMPPyy2b99erNyNGzfExIkTRa1atYSLi4sIDg4Ww4cPF9evXxdCGP6zf+bMGTFy5EhRr1494ebmJqpWrSoefvhhsX79+jKvn0xPI4QVZFUSUam6dOmC69evmzz/oLJo0aIFatSoUerwbqr8NBoNJkyYUKwrmqgo5vwQkc3Izc0tljeRmJiIgwcPmnVRUiKqXJjzQ0Q249KlS4iOjsbQoUMRFBSE48ePY+7cuQgICMBzzz2ndvWIyEYw+CEim+Hn54fIyEh88803+Pfff+Hp6YnevXtj5syZOmuxERGVhjk/REREZFeY80NERER2hcEPERER2RXm/OiRn5+Py5cvw9vb2+JLJhAREVH5CCFw+/ZtBAUFlboeI4MfPS5fvlxspW0iIiKyDRcuXEBwcHCJzzP40UOZJv7ChQvaNV6IiIjIumVkZCAkJERn0WF9GPzooXR1+fj4MPghIiKyMWWlrDDhmYiIiOwKgx8iIiKyKwx+iIiIyK4w54eIiIisRl5eHnJzc/U+5+zsDEdHxwq/BoMfIiIiUp0QAqmpqUhLSyu1XJUqVRAQEFChefgY/BAREZHqlMDH398fHh4exYIbIQTu3LmDa9euAQACAwPL/VoMfoiIiEhVeXl52sCnWrVqJZZzd3cHAFy7dg3+/v7l7gJjwjMRERGpSsnx8fDwKLOsUqakvCBDMPghIiIiq2BIHo8p1txk8ENERER2hcEPERER2RUGP0RERGRXGPxUJkLIGxERkQ0SBnyHGVKmLAx+KpOnngJq1AD+NwcCERGRLXB2dgYA3Llzp8yyShnlmPLgPD+VRXY28PvvQG4usHkzMGCA2jUiIiIyiKOjI6pUqaKdwLCsSQ6rVKlSoWUuGPxUFocPy8AHAI4dU7cuRERERgoICAAAbQBUEmV5i4pg8FNZ7NtXsM3gh4iIbIxGo0FgYCD8/f25sCkZaO/egu2jR9WrBxERUQU4OjqaJMApjdUkPM+cORMajQYvvvhiqeUWL16Mhg0bws3NDc2aNcOqVat0nhdC4O2330ZgYCDc3d0RHR2NkydPmrHmVqJw8HPiBJCXp15diIiIrJhVBD+7d+/GvHnz0Lx581LLbd++HU8++SRGjRqF/fv3IzY2FrGxsTh8+LC2zOzZs/HZZ59h7ty52LlzJzw9PRETE4N79+6Z+zLUk5MDHDpU8PjePeDcOfXqQ0REZMVUD34yMzMxZMgQfP311/Dz8yu17KeffooePXrg1VdfRaNGjfDuu++iZcuW+OKLLwDIVp9PPvkEb775Jvr164fmzZvjxx9/xOXLl7Fs2TILXI1KjhyRAZCfH9CsmdzHri8iIiK9VA9+JkyYgN69eyM6OrrMsklJScXKxcTEICkpCQCQkpKC1NRUnTK+vr5o27attkylpHR5tWwJNG4st5n0TEREpJeqCc8LFy7Evn37sHv3boPKp6amombNmjr7atasidTUVO3zyr6SyuiTnZ2N7Oxs7eOMjAyD6mM1Cgc/Xl5ym8EPERGRXqq1/Fy4cAEvvPACfv75Z7i5ualVDQBAfHw8fH19tbeQkBBV62M0JfiJjAQaNZLb7PYiIiLSS7XgZ+/evbh27RpatmwJJycnODk5YdOmTfjss8/g5OSEPD2jlQICAnD16lWdfVevXtVOdqTcl1ZGn7i4OKSnp2tvFy5cqOjlWU5ubkGyc2SkbrcX1/kiIiIqRrXgp2vXrkhOTsaBAwe0t1atWmHIkCE4cOCA3jH+UVFR2LBhg86+hIQEREVFAQBCQ0MREBCgUyYjIwM7d+7UltHH1dUVPj4+OjebcfSoXNrC1xeoVw8ICwMcHYGMDODKFbVrR0REZHVUy/nx9vZG06ZNdfZ5enqiWrVq2v3Dhg1DrVq1EB8fDwB44YUX0LlzZ3z44Yfo3bs3Fi5ciD179uCrr74CAO08QdOnT0dYWBhCQ0Px1ltvISgoCLGxsRa9PospnO+j0QAuLjII+ucfGRgFBalbPyIiIiuj+miv0pw/fx5XCrVetGvXDr/88gu++uorREREYMmSJVi2bJlOEDV58mRMmjQJY8aMQevWrZGZmYk1a9aonldkNoXzfRQc8UVERFQijRBMDCkqIyMDvr6+SE9Pt/4usIceAnbuBH79FRg8WO574w0gPh547jlgzhx160dERGQhhn5/W3XLD5Xh/n3g4EG5XbjlRxnxxZYfIiKiYhj82LKjR+VSFt7eMs9HwW4vIiKiEjH4sWWFk50dCn2UDRvK+2vXgBs3LF8vIiIiK8bgx5bt2yfvC3d5AYCnJ/DAA3KbrT9EREQ6GPzYMn0jvRTs+iIiItKLwY+tun8fOHBAbusLfrjMBRERkV4MfmzV8ePA3bsy2TksrPjzHPFFRESkF4MfW6V0eT34oG6ys4LdXkRERHox+LFVhUd66aO0/Jw/D2RmWqZORERENoDBj60qLdkZAKpWBfz95fbx45apExERkQ1g8GOL8vJKT3ZWsOuLiIioGAY/tujECeDOHTmfT3h4yeU44ouIiKgYBj+2qHCys6NjyeXY8kNERFQMgx9bVFa+j4ItP0RERMUw+LFFxgY/p08D2dnmrRMREZGNYPBja/LygP375XZZwU9gIODrC+TnAydPmr9uRERENoDBj6355x8gKwvw8AAaNCi9rEbDri8iIqIiGPzYGqXLq0WL0pOdFVzmgoiISAeDH1tjaL6PgiO+iIiIdDD4sTX79sl7Q4MfdnsRERHpYPBjS/LzDU92VijBzz//yGRpIiIiO8fgx5acPAncvg24uwMNGxp2TO3asnx2NpCSYt76ERER2QAGP7akcLKzk5Nhxzg6FowKY9cXERERgx+bYmyys4IjvoiIiLQY/NgSJfhp2dK44zjii4iISIvBj63Izzd+pJeCI76IiIi0GPzYilOnZLKzm1tBS46hCnd7CWH6uhEREdkQBj+2Qmn1iYgwPNlZUb++PCYzE7h40fR1IyIisiEMfmxFeZOdAcDFRQZAAPN+iIjI7jH4sRUVCX4AJj0TERH9D4MfWyBE+ZOdFUx6JiIiAsDgxzacPg2kpwOursYnOys41w8REREABj+2Qenyat4ccHYu3znY7UVERARA5eBnzpw5aN68OXx8fODj44OoqCisXr26xPJdunSBRqMpduvdu7e2zIgRI4o936NHD0tcjvlUNN8HkEtcaDTA9evAv/+apl5EREQ2yMgx06YVHByMmTNnIiwsDEII/PDDD+jXrx/279+PJk2aFCv/xx9/ICcnR/v4xo0biIiIwBNPPKFTrkePHpg/f772saurq/kuwhJMEfx4eMhFTs+ela0/NWqYpGpERES2RtXgp0+fPjqPZ8yYgTlz5mDHjh16g5+qVavqPF64cCE8PDyKBT+urq4ICAgwfYXVYIpkZ0XjxgXBT6dOFa4aERGRLbKanJ+8vDwsXLgQWVlZiIqKMuiYb7/9FoMHD4anp6fO/sTERPj7+6NBgwYYN24cbty4Uep5srOzkZGRoXOzGikpQFqanKtHT0BoFI74IiIiUrflBwCSk5MRFRWFe/fuwcvLC0uXLkVjA0Y07dq1C4cPH8a3336rs79Hjx54/PHHERoaitOnT+ONN95Az549kZSUBEdHR73nio+Px7Rp00xyPSZXONnZxaVi5+KILyIiImiEUHexp5ycHJw/fx7p6elYsmQJvvnmG2zatKnMAGjs2LFISkrCoUOHSi135swZ1KtXD+vXr0fXrl31lsnOzkZ2drb2cUZGBkJCQpCeng4fHx/jL8qUXn8dmDULGDsWmDu3YudKSgLatQOCg4ELF0xTPyIiIiuRkZEBX1/fMr+/Ve/2cnFxQf369REZGYn4+HhERETg008/LfWYrKwsLFy4EKNGjSrz/HXr1kX16tVx6tSpEsu4urpqR5wpN6thimRnhdLyc/EiYE1de0RERBakevBTVH5+vk4rjD6LFy9GdnY2hg4dWub5Ll68iBs3biAwMNBUVbQcIQqCn5YtK36+KlUAJRH8+PGKn4+IiMgGqRr8xMXFYfPmzTh79iySk5MRFxeHxMREDBkyBAAwbNgwxMXFFTvu22+/RWxsLKpVq6azPzMzE6+++ip27NiBs2fPYsOGDejXrx/q16+PmJgYi1yTSZ09C9y6JSc2bNrUNOdUuhOZ9ExERHZK1YTna9euYdiwYbhy5Qp8fX3RvHlzrF27Ft26dQMAnD9/Hg4OuvHZiRMnsHXrVqxbt67Y+RwdHXHo0CH88MMPSEtLQ1BQELp37453333XNuf6UVp9mjWTS1uYQqNGwN9/M+mZiIjslqrBT9GRWkUlJiYW29egQQOUlKPt7u6OtWvXmqJq1sGU+T4KjvgiIiI7Z3U5P1SIqSY3LIzdXkREZOcY/FirwsnO5mj5SUkB7t0z3XmJiIhsBIMfa3X+PHDjhkx2btbMdOetWRPw8wPy84F//jHdeYmIiGwEgx9rpbT6NG1qumRnQK7szmUuiIjIjjH4sVamnN+nKCY9ExGRHWPwY63Mke+jUJKeGfwQEZEdYvBjjcyV7KxgtxcREdkxBj/W6MIF4Pp1wMlJruZuakrw888/wP37pj9/eQkhb0RERGbE4McaKfP7NGkCuLmZ/vwPPAB4eAC5ucCZM6Y/f3ns2iWv9b331K4JERFVcgx+rJE5u7wAwMEBaNhQbltL19dXXwE5OcAHHwB376pdGyIiqsQY/Fgjcwc/gHWN+MrPB1aulNtpacDSpapWh4iIKjcGP9bG3MnOCmta5mLPHuDq1YLH332nXl2IiKjSY/BjbS5dAq5dAxwdzZPsrLCmlp8//5T3Dz0kJ2HcsAE4e1bVKhERUeXF4MfaKK0+jRsD7u7mex0l+Dl+XHY7qUkJfsaNA7p2ldvz56tXHyIiqtQY/FgbS3R5AUC9enLdsKwsObReLefPAwcPyiTsXr2AUaPk/vnzgbw89epFRESVFoMfa2Op4MfZGQgLk9tqdn0pic5RUUD16kBsLFCligzI/v5bvXoREVGlxeDH2ihz/Jg7+AGsI+9H6fLq00feu7kBQ4bIbSY+ExGRGTD4sSaXLwOpqbILKCLC/K+n9oivzMyC1h0l+AEKur6WLgVu3rR8vYiIqFJj8GNNCic7e3iY//XUbvlJSJATG9atW1AXAHjwQaBFCyA7G/jlF3XqRkRElRaDH2tiqXwfReGWHzXW1Crc5aXR6D43cqS8Z9cXERGZGIMfa6IEPy1bWub1wsNl0HHrlpxbyJLy84G//pLbhbu8FE89Bbi4APv3yxsREZGJMPixJpZu+XF3B0JD5balu7527ZIBl48P0LFj8eerVQMee0xus/WHiIhMiMGPtbhyRd4cHGS+i6UoXV+WDn6ULq8ePWQLjz5K19fPPwP37lmmXkREVOkx+LEWSqtPw4aAp6flXldJNLb0iK+iQ9z16doVCAmR3XLLl1umXkREVOkx+LEWlpzfpzA1RnydOwckJ8tWrp49Sy7n6AiMGCG3v/3WIlUjIqLKj8GPtbB0vo9Cjbl+lFaf9u1lbk9pnnlG3q9fL4MmIiKiCmLwYy3UCn4aNpT3V64A6emWeU1DurwUoaHAI4/Iofg//GDeehERkV1g8GMNrl4FLl2Sw84tmewMAL6+QFCQ3LZE19ft20Biotw2JPgBChKf589XfwV6IiKyeQx+rIHS6tOgAeDlZfnXt2TX17p1clbn+vXl9Rri8cdlkHb2LLBxo1mrR0RElR+DH2ugVpeXwpJJz0qXV9++xWd1Lom7u5z0EOCcP0REVGEMfqyBvQQ/eXmlz+pcGqXr6/ff5dB3IiKicmLwYw3UDn4s1e21cydw/TpQpYoc6WWMyEigWTO52Omvv5qlekREZB9UDX7mzJmD5s2bw8fHBz4+PoiKisLq1atLLP/9999Do9Ho3Nzc3HTKCCHw9ttvIzAwEO7u7oiOjsbJkyfNfSnld+0acPGi7AJ68EF16qC0/Jw9C9y9a77XUbq8evYEnJ2NO1ajAUaNktvs+iIiogpQNfgJDg7GzJkzsXfvXuzZswePPPII+vXrhyNHjpR4jI+PD65cuaK9nSsy98vs2bPx2WefYe7cudi5cyc8PT0RExODe9a6PIIyuWF4OODtrU4datQAqlaVw8lPnDDf6xgzxF2fIUNk0LR3L3DwoOnqRUREdkXV4KdPnz7o1asXwsLCEB4ejhkzZsDLyws7duwo8RiNRoOAgADtrWbNmtrnhBD45JNP8Oabb6Jfv35o3rw5fvzxR1y+fBnLli2zwBWVg9pdXoBsVTF319eZM8CRI3LW5h49yneO6tWBfv3k9vz5pqsbERHZFavJ+cnLy8PChQuRlZWFqKioEstlZmaidu3aCAkJKdZKlJKSgtTUVERHR2v3+fr6om3btkhKSirxnNnZ2cjIyNC5WYw1BD+A+ZOelVafjh0BP7/yn0dJfP7pJ5n/Q0REZCTVg5/k5GR4eXnB1dUVzz33HJYuXYrGSitEEQ0aNMB3332H5cuXY8GCBcjPz0e7du1w8eJFAEBqaioA6LQGKY+V5/SJj4+Hr6+v9hYSEmKiqzOAEvy0bGm519TH3Ku7V7TLS9G9O1CrFnDzJrBiRcXrRUREdkf14KdBgwY4cOAAdu7ciXHjxmH48OE4WkLXS1RUFIYNG4YWLVqgc+fO+OOPP1CjRg3MmzevQnWIi4tDenq69nbhwoUKnc9g168D58/LbbWSnRXmXN09PR3YtEluVzT4KbzYKROfiYioHFQPflxcXFC/fn1ERkYiPj4eERER+PTTTw061tnZGQ8++CBOnToFAAgICAAAXL16Vafc1atXtc/p4+rqqh1xptwsQmn1CQuTMxirSQl+Tp4EcnNNe+61a4H79+WMzmFhFT+fEvysXQtYKlAlIqJKQ/Xgp6j8/HxkG5jLkZeXh+TkZAQGBgIAQkNDERAQgA0bNmjLZGRkYOfOnaXmEanGWvJ9ACAkRC6tcf8+8L9g0mRM1eWlqF8f6NyZi50SEVG5qBr8xMXFYfPmzTh79iySk5MRFxeHxMREDBkyBAAwbNgwxMXFacu/8847WLduHc6cOYN9+/Zh6NChOHfuHEaPHg1AjgR78cUXMX36dKxYsQLJyckYNmwYgoKCEBsbq8Yllk4Z5m4NwY9GU7DCuynzfu7fB1atktumCn6Agjl/uNgpEREZyUnNF7927RqGDRuGK1euwNfXF82bN8fatWvRrVs3AMD58+fh4FAQn926dQvPPvssUlNT4efnh8jISGzfvl0nQXry5MnIysrCmDFjkJaWhg4dOmDNmjXFJkO0CtbU8gPIrq89e0wb/CQlyeRkPz+gXTvTnbd/f2DCBDmEfvNmoEsX052biIgqNY0QQqhdCWuTkZEBX19fpKenmy//58YNOW8NAKSlqZ/zAwAzZwJxcXIR0Z9/Ns05J08G3n9fTlC4YIFpzqkYOxb46itg6FA59J2IiOyaod/fVpfzYzeULq/69a0j8AHMM9ePqfN9ClO6vpYskSPKiIiIDMDgRy3WMr9PYUrwc/y4afJoTp2S53JyKv+szqVp3Rpo0gS4dw9YuND05yciokqJwY9arC3fBwDq1gVcXOTipkXWTCsXpdWnUyfztG5pNAUzPnPOHyIiMhCDH7VYY/Dj5CQXWAVM0/Vlzi4vxdChst67dgGHD5vvdYiIqNJg8KOGmzeBlBS5bU3dXoDp8n7S0oAtW+S2OYMff3+gb1+5zdYfIiIyAIMfNezfL+/r1q3YIp/mYKrV3deskXP8NGoE1KtX8XqVpvBipzk55n0tIiKyeQx+1GCNXV4KU7X8WKLLSxETAwQGyrXSlNclIiIqAYMfNdhK8FPeKaDu3wdWr5bblgh+nJyA4cPlNru+iIioDAx+1GDNwU94OODgIHN2UlPLd45t24Bbt4Bq1QBLrammdH2tWQNcumSZ1yQiIpvE4MfS0tKA06fl9oMPqloVvdzcZC4SUP6uL6XrqVcvwNHRNPUqS1gY0LGjnJ/oxx8t85pERGSTGPxYmjKzc506smXEGlU06dmS+T6FFZ7zh6u2EBFRCRj8WJo1d3kpKpL0/M8/8ubsLBORLemJJwAvLzmztDLMnoiIqAgGP5ZW2YMfpdWnc2fAXIvClsTTExg8WG4z8ZmIiErA4MfSlG4vaw5+KtLtpVaXl0Lp+lq8GMjIUKcORERk1Rj8WFJ6OnDypNy25uCnYUN5f/WqHLVlqFu3gK1b5bZawc9DD8n637kDLFqkTh2IiMiqMfixJGVm59q1rTfZGQC8vYHgYLltTNfX6tVAXp5caT001Dx1K4tGA4waJbfZ9UVERHow+LEkW8j3UZSn60vtLi/F00/LIfY7dlR8mQ4iIqp0GPxYkhL8WNtipvoYm/Scm2vZWZ1LU7Mm8OijcputP0REVASDH0uaPBn4/HP1gwNDGBv8bN0qc5qqVwfatjVfvQylJD7/+KMMzIis0a1bwPvvG5dbR0QV5qR2BexKixbyZguM7fZSurwefdRyszqXplcvICBALtHx119AbKzaNSIq7t13gY8/Bq5dk0EQEVkEW35IP6Xl59w5ICur9LJCACtWyG1radVycgKGDZPb336rbl2ISqKMjtyzR916ENkZBj+kX/Xq8gYAJ06UXvb4cblemYsL0L27+etmqGeekferVgGXL6tbF6Ki7t4tGAF66BCXZCGyIAY/VDJDu76ULq+HH5bLS1iLhg2Bdu3kYqc//aR2bYh07d8P3L8vt2/eZIBOZEEMfqhkhiY9W8sQd30Kz/nD/6zJmuzcqfv40CF16kFkhxj8UMmU4Ke0lp8bN4Dt2+W2MrzcmjzxhFzz659/gG3b1K4NUYEdO3QfM/ghshgGP1QypdurtJafVatkt1Lz5nLmamvj7Q0MHCi3OecPWRMl+OnWTd4fPKheXYjsDIMfKpnS8nPqFJCTo7+MNXd5KZSur99+A27fVrcuRABw5Qpw/jzg4FDw88mWHyKLYfBDJatVS7ac5OUVLMhaWE4OsGaN3Lbm4KddOyA8XA7Z//lntWtDVJDv06QJ0L693D5+HMjOVq9ORHaEwQ+VTKMpPel582bZklKzJtC6tWXrZgyNBhg/Xm7Pnl0wwoZILUqX10MPyX8y/PzkPxnGLCRMROXG4IdKV1rwo3R59e4tm++t2bPPAjVqACkpwK+/ql0bsneFgx+NRubMAcz7IbIQK//GItWVNNePELaR76Pw8ABeeklux8fLJG0iNdy/D+zeLbcfekjeK8EP836ILILBD5WupJafo0dlK4qra8FoFWs3fjzg6yuvZelStWtD9urIEeDOHcDHR07ECQAREfKewQ+RRaga/MyZMwfNmzeHj48PfHx8EBUVhdWrV5dY/uuvv0bHjh3h5+cHPz8/REdHY9euXTplRowYAY1Go3Pr0aOHuS+l8lKCnxMnZE6CQmn1eeQROY+OLfD1BSZNktszZtjWpIdXr5a9xhrZBqXLq02bgu5itvwQWZSqwU9wcDBmzpyJvXv3Ys+ePXjkkUfQr18/HDlyRG/5xMREPPnkk9i4cSOSkpIQEhKC7t2749KlSzrlevTogStXrmhvvzLHo/xCQ2Xrzr17wNmzBfttqcursBdekF1g+/cXjFSzdgcPys+hZ0+1a0KmUDjfR9Gkicz9uXZNBrpEZFaqBj99+vRBr169EBYWhvDwcMyYMQNeXl7YUXTm0//5+eefMX78eLRo0QINGzbEN998g/z8fGzYsEGnnKurKwICArQ3Pz8/S1xO5eToCDRoILeVrq9//wWSkuS2Nc7qXJrq1YHnnpPbttL6M3myXARzyxau/1QZ6At+PDyAsDC5zaRnIrOzmpyfvLw8LFy4EFlZWYiKijLomDt37iA3NxdVq1bV2Z+YmAh/f380aNAA48aNw40bN0o9T3Z2NjIyMnRuVEjRvJ9Vq2TQ0KIFEBKiWrXK7eWXZWvWtm1yuL41W7dO3hRFAn2yMWlpcj4fAGjbVvc55v0QWUyFg5+8vDwcOHAAt27dKtfxycnJ8PLygqurK5577jksXboUjZURRmV47bXXEBQUhOjoaO2+Hj164Mcff8SGDRswa9YsbNq0CT179kRe4XyVIuLj4+Hr66u9hdjiF7o5FR3xZatdXoqgIGDkSLk9fbq6dSlNXp5s9QFkvhIArF+vXn2o4pQcxXr1ZCtkYcz7IbIcYaQXXnhBfPPNN0IIIe7fvy/at28vNBqN8PT0FBs3bjT2dCI7O1ucPHlS7NmzR7z++uuievXq4siRI2UeFx8fL/z8/MTBgwdLLXf69GkBQKxfv77EMvfu3RPp6ena24ULFwQAkZ6ebvT1VEq//SYEIETbtkLcuyeEl5d8vGuX2jUrv5QUIRwd5XXs3Kl2bfT74QdZP19fIRYtktuBgULk56tdMyqvadPk5zhkSPHnli+Xz0VEWLxaRJVFenq6Qd/fRrf8LFmyBBH/a579888/kZKSguPHj+Oll17Cf/7zH6ODLxcXF9SvXx+RkZGIj49HREQEPv3001KP+eCDDzBz5kysW7cOzZX/lkpQt25dVK9eHadOnSqxjKurq3bEmXKjQgqv7r5pE5CZCQQEAJGR6tarIurUAYYOldszZqhaFb3u3gXefFNux8UBffsCbm5yTSjOAmy79OX7KJS/ZUePlryWHhGZhNHBz/Xr1xEQEAAAWLVqFZ544gmEh4dj5MiRSE5OrnCF8vPzkV3K+jazZ8/Gu+++izVr1qBVq1Zlnu/ixYu4ceMGAgMDK1w3uxUWJhOfb98G5s2T+x591PpndS5LXJwcYbNiBWCCn12T+uwz4MIFmVP1/PMy8OnYUT7Hri/bJETBml76gp/ateXcP7m5cmoJIjIbo7+9atasiaNHjyIvLw9r1qxBt/9NcHfnzh04Ojoada64uDhs3rwZZ8+eRXJyMuLi4pCYmIghQ4YAAIYNG4a4uDht+VmzZuGtt97Cd999hzp16iA1NRWpqanIzMwEAGRmZuLVV1/Fjh07cPbsWWzYsAH9+vVD/fr1ERMTY+ylksLVVeYoAAWTA9pqvk9hDRoAAwbI7ffeU7cuhV2/XlCf6dMBd3e5reS2MfixTadOATdvyt8nfS3WhZe5YN4PkVkZHfw888wzGDhwIJo2bQqNRqNNNt65cycaKrOVGujatWsYNmwYGjRogK5du2L37t1Yu3atNqA6f/48rly5oi0/Z84c5OTkYMCAAQgMDNTePvjgAwCAo6MjDh06hL59+yI8PByjRo1CZGQktmzZAldXV2MvlQpTur6EkK0QhZLMbdobb8j7334D/vlH3boopk8HMjLk6J///SMAoOA9T0yUrQNkW5Qur8hIwMVFfxkGP0QW4WTsAVOnTkXTpk1x4cIFPPHEE9qgwtHREa+//rpR5/r2229LfT4xMVHn8dnCk+zp4e7ujrVr1xpVBzJQ48bA8uVyu2tXOS9JZdCihezCW7kSmDkT+O47detz+jTw3//K7fffl92NihYtgKpVZevB7t1Au3aqVJHKqbR8HwWDHyKLKFfSxoABA/DSSy8hODgYAJCWlobhw4ejX79+Jq0cWRGl5QeoHF1ehSmJ+j/9BJw7p35dcnOB7t2Lr5nm4CADTwBISLB83ahiSsv3UXB1dyKLMDr4mTVrFhYtWqR9PHDgQFSrVg3BwcE4xP9WKq/Ccy/Z2qzOZXnoIblG2f37srVFLbt2AYsWydyP2bP1l1ECIub92JY7dwoCmqKTGxbWtKm8v3JFzqRORGZhdPAzd+5c7SSACQkJSEhIwOrVq9GjRw+88sorJq8gWYkWLWT+yeuvA7VqqV0b01Naf775BkhNtfzrCwG8+qrcHjasYLbfopS8nx075Og7sg379sngOjCw9FnRvb0LBhdY2whEokrE6OAnNTVVG/ysXLkSAwcORPfu3TF58mTs3r3b5BUkK+HoCCxYAMTHq10T83j4YdkClJ0NfPSR5V//zz/lUhtubsC775ZcLjQUqFtXfpFa+9IcVKBwvo9GU3pZ5v0QmZ3RwY+fnx8uXLgAAFizZo12tJcQotQlJIismkZT0PozZ45MKraU+/eB116T2y++WPZ6aRzybnsMyfdRMPghMjujg5/HH38cTz31FLp164YbN26gZ8+eAID9+/ejfv36Jq8gkcX07i27mzIz5SSDlvLdd3Kxy2rVZLdiWRj86Fq9umD+KWultPyUlu+jYNIzkdkZHfx8/PHHmDhxIho3boyEhAR4eXkBAK5cuYLx48ebvIJEFlO49efTT+VcO+aWmQm8/bbcfuutggVMS/PII7Kuhw/LxFh7duWKHH3Yvz+QkqJ2bfS7eFHeHBwAA2al1+Z7HTkiWwWJyOSMDn6cnZ3xyiuv4NNPP8WDDz6o3f/SSy9h9OjRJq0ckcU9/ric+TktTXZ/mduHHwJXr8o8nnHjDDumWjWgZUu5vWGD+epmC379FcjLkwnjf/6pdm30U7q8mjcHPD3LLh8aKstlZwMnT5q3bkR2qlzz/Jw+fRqTJk1CdHQ0oqOj8fzzz+PMmTOmrhuR5Tk6yjW/AJn4fPeu+V4rNbVgaH18fMmz/urDri/pxx8Ltq09+DEk3weQLUTNmslt5v0QmYXRwc/atWvRuHFj7Nq1C82bN0fz5s2xc+dObTcYkc176im56vu1a3Lou7lMnQpkZQFt2gBPPGHcsYWDHyFMXjWbkJws82KUBXY3bbJMV6WxjMn3UTDvh8isjA5+Xn/9dbz00kvYuXMnPvroI3z00UfYuXMnXnzxRbymjFghsmXOzsDkyXJ79mwgJ8f0r3HsWEFg9f77ZQ9/Lqp9e7lA5qVL9rsC+E8/yft+/YDwcDkztrUtb5ObC+zZI7cNbfkBOOKLyMyMDn6OHTuGUaNGFds/cuRIHD161CSVIlLdM8/ICekuXiz4kjWluDiZq9K3L9Cpk/HHu7sDHTrIbXvs+srLA37+WW4//XTBkivW1vWVnCy7TqtUkQGaoZSkZwY/RGZhdPBTo0YNHDhwoNj+AwcOwN/f3xR1IlKfmxugzFgeH2/aUTdbtshFYh0d5WKq5aV0fdljd/PffwOXL8uFXnv1Kgh+Vq2SgZG1UPJ92rYt6J4zhJLzc+ECcOuW6etFZOeMDn6effZZjBkzBrNmzcKWLVuwZcsWzJw5E2PHjsWzzz5rjjoSqWPsWDmy6vRp4LffTHNOIQqCqtGjdReMNZayztfGjfY3JFpJdB40SHb/tW8P+PkBN24ASUnq1q2w8uT7AHLKg9q15TaXuSAyOaODn7feegtvv/02Pv/8c3Tu3BmdO3fGF198galTp+Ktt94yRx2J1OHpKWdcBoD33gPy8yt+zsWL5QKmnp4y4bkiWrSQLR+3bwP2tLRMZibwxx9ye9gwee/kBPxvwlWr6voqvKyFsZj0TGQ2Rgc/Go0GL730Ei5evIj09HSkp6fj4sWLePbZZ7F9+3Zz1JFIPRMnAj4+csK5FSsqdq6cnIJh9K++CgQEVOx8jo5ywkPAvvJ+li6Vq6SHhem2qFhb3s/Nm8A//8jtNm2MP97a837u35ezog8caL8jDslmlWueH4W3tze8vb0BACdPnkTHjh1NUikiq1GlCjBhgtyeMaNif+TnzgXOnAFq1gReftkk1bPL+X6ULq+hQ3VHyfXoIVuAjh2TXZVq27VL3oeHy+5TY1n7iK/du2WO1eLF7Jojm1Oh4IfILrz0khxdtWdP+ZOL09KAd96R29OmAf9bFqbClOAnKUl2B1V2ly4VzGo9dKjuc1WqAMo/YNbQ+lPefB+FEvwcPmxdSdyKv/8u2K5oqyiRhTH4ISpLjRrAmDFye/r08p1j1iyZjNuwIaBnqohyq1tXTsiYmwts3my681qrn3+WrW8dOshrL8qaur4qku8DAPXry1GHd+5YR0tWUYWDH2t4v4mMwOCHyBCvvCKXn9iyRd6MceEC8MkncnvWLNk1YyoaTcGor8re9SVEwZxLSqJzUY8+Ku83bwbS0y1TL33y841f1qIoR0egaVO5bW1dX/fuAdu2FTzetYuL7JJNMfiv8IoymjVTrHVFZSJTCA4GRowAvvpK5v6sWWP4sW+9Jb8sOnYsaJkwpeho4OuvK3/wc/Cg7AJydS15OZCwMLkw7YkT8jMaNMiydVScPCm7Ot3dC+bsKY+ICNndeugQMGCAyapXYUlJcuHVoCCgVi2Z/7NyJcDpTshGGBz8xMbGlllGY+wU/US25LXXgG+/lUso7NkDtGpV9jEHDxYk6H7wgfHLWBhCGfGVnCwXS63oKDJrpbyPffrI/J6S9Okjg5+VK9ULfpQur8hIuVxKeVlr0rOSd/XIIzLY3L1bdn0x+CEbYXC3V35+fpm3PGtMyiMylbp1gSeflNvvvWfYMa+9JrtrBg4s33BnQ1SvDjz4oNwunIdRmdy/D/zyi9wuqctLUXi2Z7Umf6xovo/CWoMf5efskUfkEi2AHAxw5456dSIyAnN+iIyhzNOzdKmc+6c0CQmylcjZ2fBgqbwq+5D3hATg6lUZ6PXoUXrZdu3kbM83b6o327Opgh+lyywlxXpWrL99u2AY/yOPyDo+8IDs2lVahIisHIMfImM0bgw8/rjcjo8vuVx+vpzIEADGjwfq1TNvvQoHP5Vxwjkl0Xnw4LK7kZyc5HpfgDqjkLKyCua9qWjwU62azKkBrGcunS1b5ND7evXkEhwaTUHrD4e8k41g8ENkrP/8R97/+itw6pT+Mj//LPN9fHyAN980f506dpSJwBcuFMwqXFlkZADLlsntsrq8FGoOed+7VwYHwcEFgUtFWNtMz4W7vBRK8PPnn6ZZBobIzBj8EBmrZUu5jlR+vhy6XtTduwUB0htvyK4ac3N3l4t7ApWv6+v33+V72qCBYUnmQMFsz8ePlxygmktFJzcsytryfvQFP507A97esmtyzx516kVkBAY/ROWhBDc//CBbWwr7/HO5LzgYeP55y9Wpsub9FJ7bx9DRcr6+QKdOctvSrT+myvdRWNMCpzduAAcOyO2HHy7Y7+JSkIvFri+yAUYHP3Xr1sWNGzeK7U9LS0NdfTOuElVG7dvL/3Zzc+UQdsWNGwXJzdOnyxYZS1GCn40b1RvlZGrnz8vrAYAhQ4w7Vo2uLyHMF/wkJ6vfpZSYKK+xSRO5Rl1h1jS7NlEZjA5+zp49q3dIe3Z2Ni5dumSSShHZBKX15+uvgWvX5Pb06XJm4YiI4mtPmVvLlnL+m/R0mXdSGfz8s7zv3Fkm1xpD+TLeskVOOGgJFy/KmY6dnOTnYQoNGsiWlcxM4OxZ05yzvPR1eSl69QIcHGT3nNr1JCpDuWZ4Xrt2LXx9fbWP8/LysGHDBtSpU8eklSOyatHRQOvWcoK3jz+WE7x9+aV8bvZsuTyBJTk6yi+lP/6QXV+myjlRiyHLWZSmXj2gUSO5yvuaNXKkmLkprT7NmwMeHqY5p5OTbGnZv18GFmq2sCvBT9euxZ+rVk2uubZ5s2z9mTTJsnUjMoLRMzxrNBoMHz5c5zlnZ2fUqVMHH374oUkrR2TVNBrZ+hMbK4Oe5GTZDdatG9C9uzp16tZNBj8JCQUtU7Zq714ZuLi5lX9phz595Dn+/NOywY+purwUzZsXBD8GzLZvFpcvywRyBwfZEqdPnz4MfsgmGD3D8wMPPIBr167pzOycnZ2NEydO4FFlUUEDzZkzB82bN4ePjw98fHwQFRWF1atXl3rM4sWL0bBhQ7i5uaFZs2ZYtWqVzvNCCLz99tsIDAyEu7s7oqOjcfLkSaPqRWSwPn3kJG+3bwN//SUDotmz1auPkvezfbucb8aWKa0+sbFyyoDyULq+Vq+2TB5URRczLYk1JD0ruVdK96o+ypD3xETrmZSRSA+jc35SUlJQvcjQ3bRy9qcHBwdj5syZ2Lt3L/bs2YNHHnkE/fr1w5ESZs7dvn07nnzySYwaNQr79+9HbGwsYmNjcfjwYW2Z2bNn47PPPsPcuXOxc+dOeHp6IiYmBvfu3StXHYlK5eAgh7Mrnn4aaNFCtepoJ57LzTV+9Xlrkpsr51EC5HtaXlFRQNWqwK1buquQm0NOTkGulbmCHzWHu5eW76MID5e33Fw5uzmRtRJGmjlzpli4cKH28YABA4RGoxFBQUHiwIEDxp6uGD8/P/HNN9/ofW7gwIGid+/eOvvatm0rxo4dK4QQIj8/XwQEBIj3339f+3xaWppwdXUVv/76q8F1SE9PFwBEenp6Oa6A7M79+0JERgpRvboQ586pXRshRo0SAhDi5ZfVrkn5/fmnvAZ/fyFycyt2rqFDLfN+7N4tX6dqVSHy80177mvX5Lk1GiFu3zbtuQ2Rny9E7dqyDmvWlF72lVdkuaFDLVI1osIM/f42uuVn7ty5CAkJAQAkJCRg/fr1WLNmDXr27IlXlen8yyEvLw8LFy5EVlYWoqKi9JZJSkpCtNKs/z8xMTFI+t/6PSkpKUhNTdUp4+vri7Zt22rLEJmco6NsVTh7Vq5xpLbKMN+P0uX11FMy4bciLDUEu/DkhobOR2SoGjWAgACZBF7WmnLmkJICnDsnP4sOHUovq3R9/fVX5ZlygSodo/+qpKamaoOflStXYuDAgejevTvq1KmDtuUYXZKcnIyoqCjcu3cPXl5eWLp0KRo3blzia9csMrdEzZo1kZqaqn1e2VdSGX2ys7ORnZ2tfZzBvmoylqurvFkDpVvi4EE5BN/fX936GCstDVi+XG5XpMtLERMjv7T/+UfewsMrfk59zJXsrGjeHEhNlZ+rpUfyKV1eDz0EeHqWXlbparx5U+aeKZNNElkRo1t+/Pz8cOF/M9quWbNG28oihNA7/09ZGjRogAMHDmDnzp0YN24chg8fjqNHjxp9noqIj4+Hr6+v9qYEd0Q2yd+/IO/IFlfZXrIEyM6Ww7sffLDi5/P1LRidZM7WH3MlOyvUzPspbYh7UU5OQO/ecpuzPZOVMjr4efzxx/HUU0+hW7duuHHjBnr27AkA2L9/P+rXr290BVxcXFC/fn1ERkYiPj4eERER+PTTT/WWDQgIwNWrV3X2Xb16FQEBAdrnlX0lldEnLi4O6enp2tuFossVENkaW+76Urq8nn7adN1H5u76un69YA2xNm3M8xpqLXAqhGHJzoVxtmeyckYHPx9//DEmTpyIxo0bIyEhAV5eXgCAK1euYPz48RWukDJ0Xp+oqChsKPKfbEJCgjZHKDQ0FAEBATplMjIysHPnzhLziADA1dVVO9xeuRHZNCX4SUiQX1624uxZOU+MRmP8chalUb6Mt26VI79MTWn1adiw5GHgFVW45ceSn+mxY3LBUnd3w7vbYmIAZ2fZzXjihHnrR1Qelsm/1u/1118XmzZtEikpKeLQoUPi9ddfFxqNRqxbt04IIcTTTz8tXn/9dW35bdu2CScnJ/HBBx+IY8eOiSlTpghnZ2eRnJysLTNz5kxRpUoVsXz5cnHo0CHRr18/ERoaKu7evWtwvTjai2xeZqYQLi5y1M0//6hdG8O9+66sc9eupj9348by3L/8Yvpzv/mmPPeIEaY/tyI7WwgnJ/k6lhxV+Pnn8jW7dTPuuO7d5XGzZ5unXkR6mG20FwD89NNP6NChA4KCgnDu3DkAwCeffILlSpKiga5du4Zhw4ahQYMG6Nq1K3bv3o21a9eiW7duAIDz58/jypUr2vLt2rXDL7/8gq+++goRERFYsmQJli1bhqZNm2rLTJ48GZMmTcKYMWPQunVrZGZmYs2aNXBzcyvPpRLZJk9PoF07uW0rXV9CAD/+KLdNkehclDm7Ysyd7wPI9b0aNZLblpzsUGlJN7TLS8GuL7JmxkZV//3vf0X16tXF9OnThbu7uzh9+rQQQoj58+eLLl26lC9UszJs+aFKYfp0+Z/344+rXRPD7Ngh6+vhIURGhunPv3WrPH+VKkLk5JjuvHl5Qvj4yHPv32+68+ozZIh8nenTzfs6ivv35fsFCLFzp3HHnj0rj3NwEOL6dfPUj6gIs7X8fP755/j666/xn//8B46FFm5s1aoVkpOTTRiWEVGF/K8FFX//DZRjJKbFKYnOjz0GeHub/vwPPQRUry6H0ptytufjx+VSDh4eQKFWaLOwdNLzgQPy/fLxMX6V+tq1ZZ5Sfj5QZBkiIrWVa3mLB/UMP3V1dUWWra8lRFSZREbKYd5paQXLLlirnBzTLGdRGkdHoFcvuW3Krhhlfp/WrSs+IWNZLD3cXRnl1aVL+a5NmfCQQ97Jyhgd/ISGhuLAgQPF9q9ZswaNlP5oIlKfo2NBnoa15/2sXi0nxQsMNGwumfIyRx6KJfJ9FErw888/wN275n89Y4e4F6UEP2vWyLmbiKyEwcHPO++8gzt37uD//u//MGHCBCxatAhCCOzatQszZsxAXFwcJk+ebM66EpGxbGW+HyXR2RTLWZSme3c5BPvkSdMNwS68rIW5BQTIrrv8fPMvc5GTU7A4bnmDn8hIWefMTGDTJtPVjaiCDA5+pk2bhszMTIwePRqzZs3Cm2++iTt37uCpp57CnDlz8Omnn2Lw4MHmrCsRGUsJfrZtA+7cUbcuJbl1C1i5Um4PG2be1/LxkV04gGlaf27fBg4fltuWCH40Gsvl/ezaBWRlyXXFmjQp3zkcHApa29j1RVbE4OBHFJpUa8iQITh58iQyMzORmpqKixcvYtSoUWapIBFVQFgYEBIi/4vfulXt2uj322+yfs2bF3TrmJMpu7727JGtMA88AAQFVfx8hrBU3o/S5fXwwzKIKa/C77ctTbhJlZpRP9GaIlPNe3h4wN/WFk0ksicaTcGor4QEdetSEnPO7aPPo4/K+23bZJ5RRVgy30dh6eCnvF1eiq5d5ezQ58+rsy4ZkR5GBT/h4eGoWrVqqTcisjLWnPdz+rRc+dvBQeb7WEJoqOzGycuTidYVYcl8H4US/Bw8aL6WlDt3gKQkuV3R4MfDoyAAZ9cXWQmjMgunTZsGX19fc9WFiMxB+fI6cAD491+Zw2EtlLl9oqMt120EyK6YI0dkV0x51xAToiD4sWTLT+PGciTfzZvA5ctArVqmf43t22VXZEgIUI4Fq4vp00cGPn/+Cbz1VsXPR1RBRgU/gwcPZjcXka2pWVO2Fhw6JLsyBg1Su0aSEMCCBXLb3InORfXpA8ycKYdg5+bKEWDGOndOLvjp7AzomfvMbNzcgAYNgKNH5WdqjuCncJdXkXSHclG6GnfvlgGbJQNdIj0M7vYqmu9DRDbEGru+kpJkt5enJxAba9nXbttWDhlPTy8Yzm0sJd+nRQuZ02JJ5s77MVW+jyIgAGjTRm7/9ZdpzklUAQa3/Ahm6RPZruho4KOPZNKzEKb5b76ilETn/v1lAGRJjo5A797ADz/IrpjyfMmr0eWlaN4cWLjQPMFPerpsoQHkSC9T6dtXDp9fsQJ49lnTndceCSFbLo8dky2PLi7yXrkVfVyeMp6eQL161vG3wgwMDn7y8/PNWQ8iMqdOneQftHPnZGuLKfI4KiI7G1i0SG5bustL0adPQfDz0UfG/5FXI9lZUTjp2dQ2b5bD95VpEkylb1/gzTdl6+OdOzIRmspn0ybgjTfM/zpvvgm8+675X0cFZl6Ihoisgqcn0K6d/KO5fr36wc/KlXLNsVq1CiYdtLTu3eV/uqdPy8VJjVmeJzsb2LdPbqvV8gPIemdnA66upju3qbu8FE2bysVOz52TP4PK0hdkPCVX7uGH5Ui6nByZu6bcKvo4M1NOPrpmDYMfIrJx0dEFwc9zz6lbF2WU19ChsgtKDd7eMvBat062/hgT/Bw8KL8wqlcH6tY1WxVLFBwM+PnJL6hjx2TekamYK/jRaGTA8/nnsuuLwU/53LsHLFkit6dMATp3Nv1rnDgBNGwoR0Tm5an3O2pGFZi2k4hsipL0/Pff8g+aWq5fB1atktuWmtiwJMrsw8ryGoYqnO+jRk6ERmOepOd//y04nynzfRRKwLNypexaI+P99ZfMywoJATp2NM9r1K8vRxXevQucOWOe11AZgx8ie9GqlVzb6tYtYP9+9eqxaJFsWn/wwfKvGWUqSvCzbRtw44bhx6mZ76MwR95PYmLBuc0xH1SnTvJn8OrVgqRqMo7S5fXUUxVbdqQ0jo5yPikASE42z2uojMEPkb1wcir4b17NIe9Kl5daic6F1a4NNGsmWyGMme1ZzZFeCnO0/Jiry0vh4gL06CG3Oduz8W7eLGg1HTrUvK/VrJm8Z/BDRDZP7XW+TpyQ8+M4OgJPPqlOHYoydqHTa9eAlBTZ9dS6tfnqVRZzrO6+YYO8N1fwA5h2YVl7s2SJzDWLiJAJ5ObE4IeIKg0l72frVjnc2NKUJvuYGDnztDVQvozXrJFfLGVRJjds3BhQc7mfJk1kAHbtmuxGqqgLF4CTJ2VXSqdOFT9fSXr1ksFvcrIMIslwyu9PeZdkMQaDHyKqNMLD5UihnByZ52JJ+fkFf7zVTnQurE0bmd+SkWHYbM/WkO8DyHlywsLktilafzZulPetWpk3qKtaFejQQW6z9cdw587Jn0+NxjKtpkrwc+qUTHyuZBj8ENkTjUa9pS62bgXOnpVDzPv1s+xrl8bBQc72DBj2ZWwN+T4KUyY9K/k+XbtW/FxlYdeX8X75Rd4//LD8B8bcAgJkoJqfL9eRq2QY/BDZG7WCHyXR+YknLL8WVlkKfxmXtpRPXl7BKCVrCn4q2vIjhPmTnQtThrwnJsph21S6wosAW6LLC5D/KFXiri8GP0T2Rgl+9u+Xc+5Ywt27wG+/yW1r6vJSKLM9nzkjJw0sybFjwO3bgJdXwVBgNZkq6fn0aZnz4+IiZwI3t7AwuTL9/fvA2rXmfz1bd/CgbH1xdZVr4VkKgx8iqjRq1pR/1Ar/t29uf/4pc2oeeMC8ybTl5eVVMA1AaV0xSpdX69bWMeut0vJz9KicO6m8lFFeUVGWW3NLaf3hkPeyKa0+fftaNsmewQ8RVSqW7voqvJyFuSZmqyhD8lCsKd8HkPMUeXvLwOf48fKfx5JdXgol+Fm1SrYAkX55eQX5Ppbq8lIw+CGiSsWSwc+1awUTCFpjl5fi0UflfVJSyd2ByjB3awl+TLHMRX5+wUgvSwY/UVFAtWpyxnFLjzy0JYmJwJUrMvm4Z0/LvrYyl1BqquW6yC2EwQ+RPerUSc74nJJi3rV7cnOBzz6T/722bi0XS7RWtWvLQCI/v2AW3cIyMuRCj4D6w9wLq2jez5Ejck0vDw857N9SHB0LRtmx66tkSpfXwIEyJ8uSvL2BOnXk9uHDln1tM2PwQ2SPvLzkf96A6Vt/hJD/yY8fDwQGAjNmyP3WsJxFWUrr+tq9W15bnTrWM0EjUPGWH6XLq1Mny3+5Ku/3ihWlj7KzV3fvAr//Lrct3eWlqKRdXwx+iOyVstSFqYKfo0eB//wHqFtXTmI3Z45cLNTfH3j9deC550zzOuakfBmvXVt8tmdry/dRmCr4sWSXlyImRgZcp07JpU9I159/ytGFdepYZhSePgx+iKhSUfJ+NmyQ3VLlcfEi8P77QIsWcrmF996TExl6ecmWnrVrgUuXgPh42c1m7Vq3lq06t28DmzbpPmdt+T4KJS/j8mXj8zLu3y9YyV2N4Mfbu2CUHbu+iis8t49aAwUY/BBRpdK6tfzyuXkTOHDA8ONu3QK++UZ+aT3wADB5spyHxNlZjuBZuFCuNfXDD3L+HFsIehQlzfYshPW2/Hh7y9Y2wPjWn337ZC5TlSoygFUDZ3vW7/r1goECanV5AQXBz+HDMh+uklA1+ImPj0fr1q3h7e0Nf39/xMbG4kQZTZ9dunSBRqMpduut/MECMGLEiGLP9+jRw9yXQ2RbnJwK/usuq+vr3j25ovRjj8lp7599VrYYCAF07AjMnStHpCxfDgwaZLm5YsxB32zPKSkyKdjFRb0goTTlTXpWury6dFFv3iLl/d6+vdKNKKqQxYtly1zLlkCjRurVIzxc/mOTmSnXF6skVA1+Nm3ahAkTJmDHjh1ISEhAbm4uunfvjqysrBKP+eOPP3DlyhXt7fDhw3B0dMQTTzyhU65Hjx465X799VdzXw6R7SltyHtenuwSGzlSdgU98QSwbJnMhWnWDJg5U3Zxbd4MjB0rhy1XBt26yZl0z54tWNNIafV58EH5nLUpb96Pmvk+igcekMFbSaPs7JXS5TV0qLr1cHYuCL4qUdeXqu3Ra9as0Xn8/fffw9/fH3v37kWnEmaBrVq1qs7jhQsXwsPDo1jw4+rqioCAANNWmKiyUYKfLVvkyBI3N9kV8vPPsvvqypWCsiEhwFNPySZ4pSm8MvL0lMHA6tWy9adJE+vN91GUJ/jJzpaLzQLqBj+A7C49eFDm/djCqEBzO3NGtoQ5OACDB6tdG/n7fuiQDH6UySltnFXl/KT/b4G7ogFOab799lsMHjwYnp6eOvsTExPh7++PBg0aYNy4cbhx44ZJ60pUKTRsCNSqJb8Ix42T/+G1agV8/LEMfPz8ZKvOpk2yJWTmzMod+CiK5qFYa76PQgl+Dh82fLbknTtlwFuzpvrrlClfqGvXyp9Fe6fM6Ny1q5wuQm2VMOnZaoKf/Px8vPjii2jfvj2aKqMXyrBr1y4cPnwYo0eP1tnfo0cP/Pjjj9iwYQNmzZqFTZs2oWfPnsgrYURLdnY2MjIydG5EdkGjKWj9+eEHOdzYzU1OqLZ8uZzZde5cOQeMtS5LYQ5KDmFSkhzRtn+/fGxNkxsWVreubLHKzgZOnjTsmMJdXhqN+epmiJYt5Zd8ZmbB6DN7VXgFd7W7vBSVMPiBsBLPPfecqF27trhw4YLBx4wZM0Y0a9aszHKnT58WAMT69ev1Pj9lyhQBoNgtPT3d4LoQ2aykJCGCgoTo3l2IH34Qgj/3UkSEEIAQzz0n7/39hcjPV7tWJXvoIVnPhQsNK9+hgyz/9dfmrZehxoyR9Rk/Xu2aqGv3bvk+uLsLkZGhdm2k8+dlnRwdhbh3T+3alCo9Pd2g72+r+Fdu4sSJWLlyJTZu3Ijg4GCDjsnKysLChQsxatSoMsvWrVsX1atXx6lTp/Q+HxcXh/T0dO3twoULRtWfyKY99JCci2ftWplv4eOjdo2sg9L19c038v6hh9RvISmNMXk/WVkFXXlq5/so9I2ys0c//yzv+/WT0xhYg+BguZp8Xl7FFtC1IqoGP0IITJw4EUuXLsXff/+N0NBQg49dvHgxsrOzMdSAZsGLFy/ixo0bCCyh79TV1RU+Pj46NyKyc8qXsZJDY635Pgol+Dl4sOyyW7fK66pdGzDi765Zde0KuLsDFy4Ydg2V0f37gDIy2Vq6vAAZ9Feyri9Vg58JEyZgwYIF+OWXX+Dt7Y3U1FSkpqbi7t272jLDhg1DXFxcsWO//fZbxMbGolqR4bWZmZl49dVXsWPHDpw9exYbNmxAv379UL9+fcTExJj9moiokmjVSs5ppLDWfB+FMS0/1pTvo3B3l5NiAvY72/OGDXKC0OrVC94La8Hgx3TmzJmD9PR0dOnSBYGBgdrbokWLtGXOnz+PK4WH2wI4ceIEtm7dqrfLy9HREYcOHULfvn0RHh6OUaNGITIyElu2bIGrNc7PQUTWqfBszxqNnBHbminBz4ULchbu0ljD/D762Ptsz0qX16BBcn4da1LJgh+NEPbcuapfRkYGfH19kZ6ezi4wInu2apUMgFq1kqu6W7s6deQsvJs2yRF6+ty6JVsW8vNlrldQkEWrWKqrV+WoLyGsr27mlpUlpx3IypKjDK2tm3XrVjmbe3CwDLCtlKHf31aR8ExEZJV69pTLeijzrlg7Q7q+Nm+WgU/DhtYXXNSsCbRpI7dXrlS3Lpa2fLkMfOrVs84uVmUKmosXy25ZtAEMfoiISqLRAP37A2FhatfEMIYkPW/YIO+trctLoUx4aG95P0qX15Ah1pOHVViVKnKWd0BOpmnjGPwQEVUWhrT8WGu+j0IJftavly0h9uDaNTnVBKDuCu5lqUR5Pwx+iIgqC2V198OH5ZwsRV29Chw5Ire7dLFYtYzSpInMXcrO1r/gbmW0aJH8vFq3lquoWysGP0REZHXq15fLk9y5IxfHLGrjRnnfogVQZJoQq6HR2F/Xl9LlZU1z++jD4IeIiKyOo2NBYqq+vB9r7/JSKEPeV66UydmV2cmTcpFZR0c5xN2aKcHP4cM2Pws3gx8iosqktLwfJfjp2tVy9SmPTp3kMivXrgG7dqldG/NSWn26dZOj3axZw4aAkxOQnm7Vw90NweCHiKgyUfJ+igY/584Bp0/LFoaOHS1fL2O4uMhpBgDbmWagPISwnS4vQH4uDRrIbRvv+mLwQ0RUmZTU8qO0+rRpYz0LZpamf395//nncgRUJZhbpphdu4BTpwBPTyA2Vu3aGKaS5P0w+CEiqkyUL6eUFCAjo2C/reT7KAYMAKZOlS1Vv/wigzrlGiqLBQvk/WOPyQDIFjD4ISIiq1OtGlCrltxWvqCEsL3gR6MBpkwBtm2To9guXpS5Si+/DNy7p3btKi43Vw5xB6x7bp+iGPwQEZFVKtr19c8/wOXLgKsrEBWlXr3Ko21b4MABYOxY+fijj2TXnSGr11uzhATg338Bf38gOlrt2hhOCX6OH5cBnI1i8ENEVNkUTXpWWn3atQPc3dWpU0V4egJz58p5f/z9ZatD69bABx/Y7lB4pcvrySflCCpbUbu2zBnLzQVOnFC7NuXG4IeIqLIp2vJjK0Pcy9Knjwx8+vQBcnKAV1+V13T+vNo1M87t28CyZXLblrq8ANkdqcwlZcNdXwx+iIgqm8LBT15ewczOtpLvUxp/f7kC+ldfyRahxER5vbY0JH7ZMuDuXbmURatWatfGeJUg74fBDxFRZRMeLudkycyUgcKNG4CXl21+0eqj0QDPPitzgdq2lZPuDRkiu5BsYUi80uU1dKh1ruBeFgY/RERkdZydgcaN5fYnn8j7Tp3k/sqkfn1g61Zg2jQ5JH7hQtkKtGGD2jUrWWpqwYKtTz2lbl3Ki8EPERFZJSXpecsWeV8Zurz0cXIC3n4b2L4dCAuTQ+Kjo4H/+z/rHBK/cKFM0o6KAurVU7s25aMEP+fO6c4lZUMY/BARVUZK3o+isgY/ijZtgP37geeek48//liOCNO3wKuaCnd52aqqVYGgILl9+LC6dSknBj9ERJVR4eCnatWClqDKzNMTmDNHrgZfs6b8Ym7TBnj/fZn4rbbjx4G9e2Vr1cCBatemYmy864vBDxFRZVQ4+Hn4YcDBjv7c9+4tv5T79ZND4idPli1f586pWy9lEdMePYDq1dWtS0Ux+CEiIqvj7w8EBMjtyt7lpU+NGsDSpcA338gWoc2bZUD4009yuQ9Ls7UV3MvC4IeIiKzSCy/I4e1PPKF2TdSh0QCjRsm8n6gomZw7bBgwaBBw86Zl65KUJBeb9fKSkzTausLBjxrBZAUx+CEiqqxefx3YvVu2gtizevVky8+778p8m8WL5Ze3MuTcEpRE5/79AQ8Py72uuTRqJKcXuHVLrhtnYxj8EBFR5efkBLz5pmyBadBAfmF36wY8+qicJdqcrRc5OQUruFeGLi8AcHOTUwsANtn1xeCHiIjsR6tWwL59wPjxslvsr79kQnjr1sCvv5pnpfK1a2U3W2CgfK3Kwobzfhj8EBGRffHwAL78Uq5KPn68XOl+714543L9+sBHH5l28r7CK7g7OpruvGpj8ENERGRjwsJkEHT+vMwH8veX2y+/DISEyCHyFy9W7DUyMoAVK+R2ZenyUjD4ISIislHVq8t8oHPngK+/Bho2lEHL++8DoaHA00/LRVTL448/5DIbjRoBLVqYstbqU4KfY8eA+/fVrYuRGPwQEREBMol39GjgyBE5S3SXLvJLfcEC4MEH5Zpha9YYlxxt6yu4lyY0VM6hlJ0NnDypdm2MwuCHiIioMAcHOUv0xo3Anj0FuTobNgA9e8oWj/nz5Zd+aS5dAv7+W27b6grupXFwAJo0kds21vXF4IeIiKgkkZHAL78AZ87IleK9vWXL0MiRQJ06wHvvlTxh4sKFspWoQwdZtjKy0bwfBj9ERERleeAB4MMPgQsXZC5QrVpAairwn//I5OhJk4DTp3WPqQwruJeFwY/x4uPj0bp1a3h7e8Pf3x+xsbE4ceJEqcd8//330Gg0Ojc3NzedMkIIvP322wgMDIS7uzuio6Nx0sb6I4mIyAr5+gKvvCKXqliwQCYx37kDfPEFEB4ODBgA7NghV5Q/cABwdq7cy4sw+DHepk2bMGHCBOzYsQMJCQnIzc1F9+7dkZWVVepxPj4+uHLlivZ2rshKvbNnz8Znn32GuXPnYufOnfD09ERMTAzu3btnzsshIiJ74ewMDBkiJ0xcv17mAuXnA7//LtcRUxaT7dULqFpV3bqakxL8nDkDZGaqWxcjOKn54mvWrNF5/P3338Pf3x979+5Fp06dSjxOo9EgQFmtuAghBD755BO8+eab6NevHwDgxx9/RM2aNbFs2TIMHjzYdBdARET2TaMBunaVt8OH5QSJP/8M/PuvfL4yd3kBct24mjWBq1dlLlTbtmrXyCBWlfOTnp4OAKhaRpScmZmJ2rVrIyQkBP369cORI0e0z6WkpCA1NRXR0dHafb6+vmjbti2SkpL0ni87OxsZGRk6NyIiIqM0bQp89x1w9izw9tvAa68BsbFq18r8bLDry2qCn/z8fLz44oto3749mjZtWmK5Bg0a4LvvvsPy5cuxYMEC5Ofno127drj4v1k4U1NTAQA1a9bUOa5mzZra54qKj4+Hr6+v9hYSEmKiqyIiIrsTGAhMmwbMnCkXVK3sGPyU34QJE3D48GEsXLiw1HJRUVEYNmwYWrRogc6dO+OPP/5AjRo1MG/evHK/dlxcHNLT07W3CxculPtcREREdsUGgx+rCEknTpyIlStXYvPmzQgODjbqWGdnZzz44IM4deoUAGhzga5evYrAwEBtuatXr6JFCVOLu7q6wtXVtXyVJyIismeFgx8hbGIma1VbfoQQmDhxIpYuXYq///4boaGhRp8jLy8PycnJ2kAnNDQUAQEB2LBhg7ZMRkYGdu7ciaioKJPVnYiIiAA0biwDnuvXZeKzDVA1+JkwYQIWLFiAX375Bd7e3khNTUVqairu3r2rLTNs2DDExcVpH7/zzjtYt24dzpw5g3379mHo0KE4d+4cRo8eDUCOBHvxxRcxffp0rFixAsnJyRg2bBiCgoIQaw+JZ0RERJbk4QHUry+3baTrS9Vurzlz5gAAunTporN//vz5GDFiBADg/PnzcHAoiNFu3bqFZ599FqmpqfDz80NkZCS2b9+Oxo0ba8tMnjwZWVlZGDNmDNLS0tChQwesWbOm2GSIREREZALNmsnFTZOTgW7d1K5NmTRCGLM8rX3IyMiAr68v0tPT4ePjo3Z1iIiIrNvUqXKE24gRctFXlRj6/W01o72IiIjIRtnYiC8GP0RERFQxSvBz5AiQl6duXQzA4IeIiIgqpl49wN0duHev+Or2VojBDxEREVWMo6Mc8g7YRNcXgx8iIiKqOBvK+2HwQ0RERBXH4IeIiIjsCoMfIiIisitK8HPqFHDnjrp1KQODHyIiIqq4mjWB6tXl4qZHj6pdm1Ix+CEiIqKK02hspuuLwQ8RERGZBoMfIiIisisMfoiIiMiuMPghIiIiu9Kkiby/ehX4919161IKBj9ERERkGl5eQN26cvvwYXXrUgoGP0RERGQ6NtD1xeCHiIiITIfBDxEREdkVBj9ERERkV5Tg5/BhID9f3bqUgMEPERERmU5YGODqCmRlAWfPql0bvRj8EBERkek4OQGNGsltK+36YvBDREREpmXleT8MfoiIiMi0GPwQERGRXWHwQ0RERHZFCX7++QfIzla3Lnow+CEiIiLTCgoC/PyAvDzg2DG1a1MMgx8iIiIyLY3Gqru+GPwQERGR6TH4ISIiIrvC4IeIiIjsCoMfIiIisitNm8r7S5eAW7fUrUsRDH6IiIjI9Hx8gNq15baVtf6oGvzEx8ejdevW8Pb2hr+/P2JjY3HixIlSj/n666/RsWNH+Pn5wc/PD9HR0di1a5dOmREjRkCj0ejcevToYc5LISIioqKstOtL1eBn06ZNmDBhAnbs2IGEhATk5uaie/fuyMrKKvGYxMREPPnkk9i4cSOSkpIQEhKC7t2749KlSzrlevTogStXrmhvv/76q7kvh4iIiAqz0uDHSc0XX7Nmjc7j77//Hv7+/ti7dy86deqk95iff/5Z5/E333yD33//HRs2bMCwYcO0+11dXREQEGD6ShMREZFhrDT4saqcn/T0dABA1apVDT7mzp07yM3NLXZMYmIi/P390aBBA4wbNw43btwo8RzZ2dnIyMjQuREREVEFKcHP4cOAEOrWpRCNENZRm/z8fPTt2xdpaWnYunWrwceNHz8ea9euxZEjR+Dm5gYAWLhwITw8PBAaGorTp0/jjTfegJeXF5KSkuDo6FjsHFOnTsW0adOK7U9PT4ePj0/5L4qIiMie5eYCnp7y/uzZggRoM8nIyICvr2+Z399WE/yMGzcOq1evxtatWxEcHGzQMTNnzsTs2bORmJiI5s2bl1juzJkzqFevHtavX4+uXbsWez47OxvZhRZey8jIQEhICIMfIiKiimreXHZ7/fkn8OijZn0pQ4Mfq+j2mjhxIlauXImNGzcaHPh88MEHmDlzJtatW1dq4AMAdevWRfXq1XHq1Cm9z7u6usLHx0fnRkRERCZghXk/qiY8CyEwadIkLF26FImJiQgNDTXouNmzZ2PGjBlYu3YtWrVqVWb5ixcv4saNGwgMDKxolYmIiMgYVhj8qNryM2HCBCxYsAC//PILvL29kZqaitTUVNy9e1dbZtiwYYiLi9M+njVrFt566y189913qFOnjvaYzMxMAEBmZiZeffVV7NixA2fPnsWGDRvQr18/1K9fHzExMRa/RiIiIrvG4EfXnDlzkJ6eji5duiAwMFB7W7RokbbM+fPnceXKFZ1jcnJyMGDAAJ1jPvjgAwCAo6MjDh06hL59+yI8PByjRo1CZGQktmzZAldXV4tfIxERkV1Tgp/jx4GcHHXr8j9Wk/BsTQxNmCIiIqIyCAH4+QHp6cChQwXBkBnYVMIzERERVVIaTcEip1bS9cXgh4iIiMzLyvJ+GPwQERGReTH4ISIiIrvC4IeIiIjsipLzc/68THxWGYMfIiIiMi8/P0BZweHwYXXrAgY/REREZAlW1PXF4IeIiIjMj8EPERER2RUGP0RERGRXCgc/Ki8uweCHiIiIzK9hQ8DREUhLAy5dUrUqDH6IiIjI/FxdgQYN5LbKXV8MfoiIiMgyrCTvh8EPERERWQaDHyIiIrIrDH6IiIjIrijBz7FjQG6uatVg8ENERESWUbs24OUF5OQAJ0+qVg0GP0RERGQZDg4Fi5yq2PXF4IeIiIgsp1kzwNMTuHlTtSpohFB5mkUrlJGRAV9fX6Snp8PHx0ft6hAREVUemZmAh4dsBTIxQ7+/nUz+ykREREQl8fJSuwbs9iIiIiL7wuCHiIiI7AqDHyIiIrIrDH6IiIjIrjD4ISIiIrvC4IeIiIjsCoMfIiIisisMfoiIiMiuMPghIiIiu8Lgh4iIiOwKgx8iIiKyKwx+iIiIyK4w+CEiIiK7wlXd9RBCAAAyMjJUrgkREREZSvneVr7HS8LgR4/bt28DAEJCQlSuCRERERnr9u3b8PX1LfF5jSgrPLJD+fn5uHz5Mry9vaHRaEx23oyMDISEhODChQvw8fEx2XmtlT1dL6+18rKn6+W1Vl72cr1CCNy+fRtBQUFwcCg5s4ctP3o4ODggODjYbOf38fGp1D98RdnT9fJaKy97ul5ea+VlD9dbWouPggnPREREZFcY/BAREZFdYfBjQa6urpgyZQpcXV3VropF2NP18lorL3u6Xl5r5WVv11sWJjwTERGRXWHLDxEREdkVBj9ERERkVxj8EBERkV1h8ENERER2hcGPiX355ZeoU6cO3Nzc0LZtW+zatavU8osXL0bDhg3h5uaGZs2aYdWqVRaqacXEx8ejdevW8Pb2hr+/P2JjY3HixIlSj/n++++h0Wh0bm5ubhaqcflNnTq1WL0bNmxY6jG2+rkCQJ06dYpdr0ajwYQJE/SWt6XPdfPmzejTpw+CgoKg0WiwbNkyneeFEHj77bcRGBgId3d3REdH4+TJk2We19jfe0so7Vpzc3Px2muvoVmzZvD09ERQUBCGDRuGy5cvl3rO8vwuWEpZn+2IESOK1b1Hjx5lntfWPlsAen9/NRoN3n///RLPac2frTkw+DGhRYsW4f/+7/8wZcoU7Nu3DxEREYiJicG1a9f0lt++fTuefPJJjBo1Cvv370dsbCxiY2Nx+PBhC9fceJs2bcKECROwY8cOJCQkIDc3F927d0dWVlapx/n4+ODKlSva27lz5yxU44pp0qSJTr23bt1aYllb/lwBYPfu3TrXmpCQAAB44oknSjzGVj7XrKwsRERE4Msvv9T7/OzZs/HZZ59h7ty52LlzJzw9PRETE4N79+6VeE5jf+8tpbRrvXPnDvbt24e33noL+/btwx9//IETJ06gb9++ZZ7XmN8FSyrrswWAHj166NT9119/LfWctvjZAtC5xitXruC7776DRqNB//79Sz2vtX62ZiHIZNq0aSMmTJigfZyXlyeCgoJEfHy83vIDBw4UvXv31tnXtm1bMXbsWLPW0xyuXbsmAIhNmzaVWGb+/PnC19fXcpUykSlTpoiIiAiDy1emz1UIIV544QVRr149kZ+fr/d5W/1cAYilS5dqH+fn54uAgADx/vvva/elpaUJV1dX8euvv5Z4HmN/79VQ9Fr12bVrlwAgzp07V2IZY38X1KLveocPHy769etn1Hkqy2fbr18/8cgjj5RaxlY+W1Nhy4+J5OTkYO/evYiOjtbuc3BwQHR0NJKSkvQek5SUpFMeAGJiYkosb83S09MBAFWrVi21XGZmJmrXro2QkBD069cPR44csUT1KuzkyZMICgpC3bp1MWTIEJw/f77EspXpc83JycGCBQswcuTIUhf5tdXPtbCUlBSkpqbqfHa+vr5o27ZtiZ9deX7vrVV6ejo0Gg2qVKlSajljfhesTWJiIvz9/dGgQQOMGzcON27cKLFsZflsr169ir/++gujRo0qs6wtf7bGYvBjItevX0deXh5q1qyps79mzZpITU3Ve0xqaqpR5a1Vfn4+XnzxRbRv3x5NmzYtsVyDBg3w3XffYfny5ViwYAHy8/PRrl07XLx40YK1NV7btm3x/fffY82aNZgzZw5SUlLQsWNH3L59W2/5yvK5AsCyZcuQlpaGESNGlFjGVj/XopTPx5jPrjy/99bo3r17eO211/Dkk0+Wuuilsb8L1qRHjx748ccfsWHDBsyaNQubNm1Cz549kZeXp7d8Zflsf/jhB3h7e+Pxxx8vtZwtf7blwVXdqcImTJiAw4cPl9k/HBUVhaioKO3jdu3aoVGjRpg3bx7effddc1ez3Hr27Kndbt68Odq2bYvatWvjt99+M+i/KVv27bffomfPnggKCiqxjK1+riTl5uZi4MCBEEJgzpw5pZa15d+FwYMHa7ebNWuG5s2bo169ekhMTETXrl1VrJl5fffddxgyZEiZgxBs+bMtD7b8mEj16tXh6OiIq1ev6uy/evUqAgIC9B4TEBBgVHlrNHHiRKxcuRIbN25EcHCwUcc6OzvjwQcfxKlTp8xUO/OoUqUKwsPDS6x3ZfhcAeDcuXNYv349Ro8ebdRxtvq5Kp+PMZ9deX7vrYkS+Jw7dw4JCQmltvroU9bvgjWrW7cuqlevXmLdbf2zBYAtW7bgxIkTRv8OA7b92RqCwY+JuLi4IDIyEhs2bNDuy8/Px4YNG3T+Ky4sKipKpzwAJCQklFjemgghMHHiRCxduhR///03QkNDjT5HXl4ekpOTERgYaIYamk9mZiZOnz5dYr1t+XMtbP78+fD390fv3r2NOs5WP9fQ0FAEBATofHYZGRnYuXNniZ9deX7vrYUS+Jw8eRLr169HtWrVjD5HWb8L1uzixYu4ceNGiXW35c9W8e233yIyMhIRERFGH2vLn61B1M64rkwWLlwoXF1dxffffy+OHj0qxowZI6pUqSJSU1OFEEI8/fTT4vXXX9eW37Ztm3BychIffPCBOHbsmJgyZYpwdnYWycnJal2CwcaNGyd8fX1FYmKiuHLlivZ2584dbZmi1ztt2jSxdu1acfr0abF3714xePBg4ebmJo4cOaLGJRjs5ZdfFomJiSIlJUVs27ZNREdHi+rVq4tr164JISrX56rIy8sTDzzwgHjttdeKPWfLn+vt27fF/v37xf79+wUA8dFHH4n9+/drRzjNnDlTVKlSRSxfvlwcOnRI9OvXT4SGhoq7d+9qz/HII4+Izz//XPu4rN97tZR2rTk5OaJv374iODhYHDhwQOd3ODs7W3uOotda1u+Cmkq73tu3b4tXXnlFJCUliZSUFLF+/XrRsmVLERYWJu7du6c9R2X4bBXp6enCw8NDzJkzR+85bOmzNQcGPyb2+eefiwceeEC4uLiINm3aiB07dmif69y5sxg+fLhO+d9++02Eh4cLFxcX0aRJE/HXX39ZuMblA0Dvbf78+doyRa/3xRdf1L43NWvWFL169RL79u2zfOWNNGjQIBEYGChcXFxErVq1xKBBg8SpU6e0z1emz1Wxdu1aAUCcOHGi2HO2/Llu3LhR78+tcj35+fnirbfeEjVr1hSurq6ia9euxd6D2rVriylTpujsK+33Xi2lXWtKSkqJv8MbN27UnqPotZb1u6Cm0q73zp07onv37qJGjRrC2dlZ1K5dWzz77LPFgpjK8Nkq5s2bJ9zd3UVaWprec9jSZ2sOGiGEMGvTEhEREZEVYc4PERER2RUGP0RERGRXGPwQERGRXWHwQ0RERHaFwQ8RERHZFQY/REREZFcY/BAREZFdYfBDRKSHRqPBsmXL1K4GEZkBgx8isjojRoyARqMpduvRo4faVSOiSsBJ7QoQEenTo0cPzJ8/X2efq6urSrUhosqELT9EZJVcXV0REBCgc/Pz8wMgu6TmzJmDnj17wt3dHXXr1sWSJUt0jk9OTsYjjzwCd3d3VKtWDWPGjEFmZqZOme+++w5NmjSBq6srAgMDMXHiRJ3nr1+/jsceewweHh4ICwvDihUrtM/dunULQ4YMQY0aNeDu7o6wsLBiwRoRWScGP0Rkk9566y30798fBw8exJAhQzB48GAcO3YMAJCVlYWYmBj4+flh9+7dWLx4MdavX68T3MyZMwcTJkzAmDFjkJycjBUrVqB+/fo6rzFt2jQMHDgQhw4dQq9evTBkyBDcvHlT+/pHjx7F6tWrcezYMcyZMwfVq1e33BtAROWn9sqqRERFDR8+XDg6OgpPT0+d24wZM4QQQgAQzz33nM4xbdu2FePGjRNCCPHVV18JPz8/kZmZqX3+r7/+Eg4ODtqVvIOCgsR//vOfEusAQLz55pvax5mZmQKAWL16tRBCiD59+ohnnnnGNBdMRBbFnB8iskoPP/ww5syZo7OvatWq2u2oqCid56KionDgwAEAwLFjxxAREQFPT0/t8+3bt0d+fj5OnDgBjUaDy5cvo2vXrqXWoXnz5tptT09P+Pj44Nq1awCAcePGoX///ti3bx+6d++O2NhYtGvXrlzXSkSWxeCHiKySp6dnsW4oU3F3dzeonLOzs85jjUaD/Px8AEDPnj1x7tw5rFq1CgkJCejatSsmTJiADz74wOT1JSLTYs4PEdmkHTt2FHvcqFEjAECjRo1w8OBBZGVlaZ/ftm0bHBwc0KBBA3h7e6NOnTrYsGFDhepQo0YNDB8+HAsWLMAnn3yCr776qkLnIyLLYMsPEVml7OxspKam6uxzcnLSJhUvXrwYrVq1QocOHfDzzz9j165d+PbbbwEAQ4YMwZQpUzB8+HBMnToV//77LyZNmoSnn34aNWvWBABMnToVzz33HPz9/dGzZ0/cvn0b27Ztw6RJkwyq39tvv43IyEg0adIE2dnZWLlypTb4IiLrxuCHiKzSmjVrEBgYqLOvQYMGOH78OAA5EmvhwoUYP348AgMD8euvv6Jx48YAAA8PD6xduxYvvPACWrduDQ8PD/Tv3x8fffSR9lzDhw/HvXv38PHHH+OVV15B9erVMWDAAIPr5+Ligri4OJw9exbu7u7o2LEjFi5caIIrJyJz0wghhNqVICIyhkajwdKlSxEbG6t2VYjIBjHnh4iIiOwKgx8iIiKyK8z5ISKbw956IqoItvwQERGRXWHwQ0RERHaFwQ8RERHZFQY/REREZFcY/BAREZFdYfBDREREdoXBDxEREdkVBj9ERERkVxj8EBERkV35fzaEu/IVBH1XAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "train_losses = []\n",
    "\n",
    "for log in trainer_sentences.state.log_history:\n",
    "    if 'loss' in log:\n",
    "        train_losses.append(log['loss'])\n",
    "\n",
    "print(f\"prompt_tuning_init={chosen_prompt_tuning_init}, num_virtual_tokens={chosen_num_virtual_tokens} \\n\" + str(train_losses))\n",
    "# print(train_losses)\n",
    "\n",
    "plt.title(\"Soft Prompt Tuning: training loss vs. epochs\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Training Loss\")\n",
    "plt.plot(range(len(train_losses)), train_losses, color =\"red\")\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "test_losses = []\n",
    "\n",
    "for log in test_sentences.state.log_history:\n",
    "    if 'loss' in log:\n",
    "        test_losses.append(log['loss'])\n",
    "\n",
    "print(f\"prompt_tuning_init={chosen_prompt_tuning_init}, num_virtual_tokens={chosen_num_virtual_tokens} \\n\" + str(test_losses))\n",
    "\n",
    "plt.title(\"Soft Prompt Tuning: test loss vs. epochs\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Test Loss\")\n",
    "plt.plot(range(len(test_losses)), test_losses, color =\"red\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "kibsi",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}