ACCDA.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```\n",
    "conda create --prefix /global/cscratch1/sd/saksman/accda_conda/ --file MAS.txt\n",
    "conda activate /global/cscratch1/sd/saksman/accda_conda\n",
    "pip install imageio umap matplotlib faiss-gpu tensorflow-addons==0.9.1 numpy==1.19.5 tensorflow==2.2\n",
    "pip install tensorflow_gpu==1.15.5\n",
    "python3 -m ipykernel install --user --name=accda\n",
    "```"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: tensorflow in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (2.2.0)\n",
      "Collecting tensorflow\n",
      "  Downloading tensorflow-2.6.2-cp36-cp36m-manylinux2010_x86_64.whl (458.3 MB)\n",
      "\u001b[K     |████████████████████▊           | 296.0 MB 170.3 MB/s eta 0:00:01"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "IOPub data rate exceeded.\n",
      "The Jupyter server will temporarily stop sending output\n",
      "to the client in order to avoid crashing it.\n",
      "To change this limit, set the config variable\n",
      "`--ServerApp.iopub_data_rate_limit`.\n",
      "\n",
      "Current values:\n",
      "ServerApp.iopub_data_rate_limit=1000000.0 (bytes/sec)\n",
      "ServerApp.rate_limit_window=3.0 (secs)\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[K     |████████████████████████████████| 458.3 MB 56 kB/s \n",
      "\u001b[?25hCollecting gast==0.4.0\n",
      "  Using cached gast-0.4.0-py3-none-any.whl (9.8 kB)\n",
      "Requirement already satisfied: grpcio<2.0,>=1.37.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.44.0)\n",
      "Requirement already satisfied: protobuf>=3.9.2 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (3.19.4)\n",
      "Requirement already satisfied: numpy~=1.19.2 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.19.5)\n",
      "Collecting clang~=5.0\n",
      "  Using cached clang-5.0.tar.gz (30 kB)\n",
      "Requirement already satisfied: six~=1.15.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.15.0)\n",
      "Collecting keras<2.7,>=2.6.0\n",
      "  Using cached keras-2.6.0-py2.py3-none-any.whl (1.3 MB)\n",
      "Requirement already satisfied: astunparse~=1.6.3 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.6.3)\n",
      "Collecting wrapt~=1.12.1\n",
      "  Downloading wrapt-1.12.1.tar.gz (27 kB)\n",
      "Requirement already satisfied: opt-einsum~=3.3.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (3.3.0)\n",
      "Collecting h5py~=3.1.0\n",
      "  Downloading h5py-3.1.0-cp36-cp36m-manylinux1_x86_64.whl (4.0 MB)\n",
      "\u001b[K     |████████████████████████████████| 4.0 MB 132.9 MB/s eta 0:00:01\n",
      "\u001b[?25hRequirement already satisfied: google-pasta~=0.2 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (0.2.0)\n",
      "Requirement already satisfied: keras-preprocessing~=1.1.2 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.1.2)\n",
      "Collecting flatbuffers~=1.12.0\n",
      "  Using cached flatbuffers-1.12-py2.py3-none-any.whl (15 kB)\n",
      "Collecting typing-extensions~=3.7.4\n",
      "  Downloading typing_extensions-3.7.4.3-py3-none-any.whl (22 kB)\n",
      "Requirement already satisfied: wheel~=0.35 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (0.37.1)\n",
      "Collecting tensorboard<2.7,>=2.6.0\n",
      "  Downloading tensorboard-2.6.0-py3-none-any.whl (5.6 MB)\n",
      "\u001b[K     |████████████████████████████████| 5.6 MB 24.5 MB/s eta 0:00:01\n",
      "\u001b[?25hCollecting absl-py~=0.10\n",
      "  Using cached absl_py-0.15.0-py3-none-any.whl (132 kB)\n",
      "Collecting tensorflow-estimator<2.7,>=2.6.0\n",
      "  Using cached tensorflow_estimator-2.6.0-py2.py3-none-any.whl (462 kB)\n",
      "Requirement already satisfied: termcolor~=1.1.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorflow) (1.1.0)\n",
      "Collecting cached-property\n",
      "  Downloading cached_property-1.5.2-py2.py3-none-any.whl (7.6 kB)\n",
      "Requirement already satisfied: requests<3,>=2.21.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (2.27.1)\n",
      "Requirement already satisfied: markdown>=2.6.8 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (3.3.6)\n",
      "Requirement already satisfied: google-auth-oauthlib<0.5,>=0.4.1 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (0.4.6)\n",
      "Requirement already satisfied: tensorboard-plugin-wit>=1.6.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (1.8.1)\n",
      "Requirement already satisfied: setuptools>=41.0.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (58.0.4)\n",
      "Requirement already satisfied: google-auth<2,>=1.6.3 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (1.35.0)\n",
      "Collecting tensorboard-data-server<0.7.0,>=0.6.0\n",
      "  Using cached tensorboard_data_server-0.6.1-py3-none-manylinux2010_x86_64.whl (4.9 MB)\n",
      "Requirement already satisfied: werkzeug>=0.11.15 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from tensorboard<2.7,>=2.6.0->tensorflow) (2.0.3)\n",
      "Requirement already satisfied: rsa<5,>=3.1.4 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from google-auth<2,>=1.6.3->tensorboard<2.7,>=2.6.0->tensorflow) (4.8)\n",
      "Requirement already satisfied: pyasn1-modules>=0.2.1 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from google-auth<2,>=1.6.3->tensorboard<2.7,>=2.6.0->tensorflow) (0.2.8)\n",
      "Requirement already satisfied: cachetools<5.0,>=2.0.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from google-auth<2,>=1.6.3->tensorboard<2.7,>=2.6.0->tensorflow) (4.2.4)\n",
      "Requirement already satisfied: requests-oauthlib>=0.7.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.7,>=2.6.0->tensorflow) (1.3.1)\n",
      "Requirement already satisfied: importlib-metadata>=4.4 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from markdown>=2.6.8->tensorboard<2.7,>=2.6.0->tensorflow) (4.8.3)\n",
      "Requirement already satisfied: zipp>=0.5 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from importlib-metadata>=4.4->markdown>=2.6.8->tensorboard<2.7,>=2.6.0->tensorflow) (3.6.0)\n",
      "Requirement already satisfied: pyasn1<0.5.0,>=0.4.6 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from pyasn1-modules>=0.2.1->google-auth<2,>=1.6.3->tensorboard<2.7,>=2.6.0->tensorflow) (0.4.8)\n",
      "Requirement already satisfied: urllib3<1.27,>=1.21.1 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from requests<3,>=2.21.0->tensorboard<2.7,>=2.6.0->tensorflow) (1.26.9)\n",
      "Requirement already satisfied: certifi>=2017.4.17 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from requests<3,>=2.21.0->tensorboard<2.7,>=2.6.0->tensorflow) (2020.6.20)\n",
      "Requirement already satisfied: idna<4,>=2.5 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from requests<3,>=2.21.0->tensorboard<2.7,>=2.6.0->tensorflow) (3.3)\n",
      "Requirement already satisfied: charset-normalizer~=2.0.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from requests<3,>=2.21.0->tensorboard<2.7,>=2.6.0->tensorflow) (2.0.12)\n",
      "Requirement already satisfied: oauthlib>=3.0.0 in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard<2.7,>=2.6.0->tensorflow) (3.2.0)\n",
      "Requirement already satisfied: dataclasses in /global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages (from werkzeug>=0.11.15->tensorboard<2.7,>=2.6.0->tensorflow) (0.8)\n",
      "Building wheels for collected packages: clang, wrapt\n",
      "  Building wheel for clang (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for clang: filename=clang-5.0-py3-none-any.whl size=30694 sha256=eabfaa576a15a2d39578aa955a4f079b3d44ab55c6e6ed160a3ce07c6e5a2d4b\n",
      "  Stored in directory: /global/u2/s/saksman/.cache/pip/wheels/22/4c/94/0583f60c9c5b6024ed64f290cb2d43b06bb4f75577dc3c93a7\n",
      "  Building wheel for wrapt (setup.py) ... \u001b[?25ldone\n",
      "\u001b[?25h  Created wheel for wrapt: filename=wrapt-1.12.1-cp36-cp36m-linux_x86_64.whl size=70495 sha256=d05b97d6c7954695b4743f99a03c7b770ef6afabcc4ed0081985ec8b6db04906\n",
      "  Stored in directory: /global/u2/s/saksman/.cache/pip/wheels/32/42/7f/23cae9ff6ef66798d00dc5d659088e57dbba01566f6c60db63\n",
      "Successfully built clang wrapt\n",
      "Installing collected packages: typing-extensions, tensorboard-data-server, cached-property, absl-py, wrapt, tensorflow-estimator, tensorboard, keras, h5py, gast, flatbuffers, clang, tensorflow\n",
      "  Attempting uninstall: typing-extensions\n",
      "    Found existing installation: typing-extensions 4.1.1\n",
      "    Uninstalling typing-extensions-4.1.1:\n",
      "      Successfully uninstalled typing-extensions-4.1.1\n",
      "  Attempting uninstall: absl-py\n",
      "    Found existing installation: absl-py 1.0.0\n",
      "    Uninstalling absl-py-1.0.0:\n",
      "      Successfully uninstalled absl-py-1.0.0\n",
      "  Attempting uninstall: wrapt\n",
      "    Found existing installation: wrapt 1.14.0\n",
      "    Uninstalling wrapt-1.14.0:\n",
      "      Successfully uninstalled wrapt-1.14.0\n",
      "  Attempting uninstall: tensorflow-estimator\n",
      "    Found existing installation: tensorflow-estimator 2.2.0\n",
      "    Uninstalling tensorflow-estimator-2.2.0:\n",
      "      Successfully uninstalled tensorflow-estimator-2.2.0\n",
      "  Attempting uninstall: tensorboard\n",
      "    Found existing installation: tensorboard 2.2.2\n",
      "    Uninstalling tensorboard-2.2.2:\n",
      "      Successfully uninstalled tensorboard-2.2.2\n",
      "  Attempting uninstall: h5py\n",
      "    Found existing installation: h5py 2.10.0\n",
      "    Uninstalling h5py-2.10.0:\n",
      "      Successfully uninstalled h5py-2.10.0\n",
      "  Attempting uninstall: gast\n",
      "    Found existing installation: gast 0.3.3\n",
      "    Uninstalling gast-0.3.3:\n",
      "      Successfully uninstalled gast-0.3.3\n",
      "  Attempting uninstall: tensorflow\n",
      "    Found existing installation: tensorflow 2.2.0\n",
      "    Uninstalling tensorflow-2.2.0:\n",
      "      Successfully uninstalled tensorflow-2.2.0\n",
      "Successfully installed absl-py-0.15.0 cached-property-1.5.2 clang-5.0 flatbuffers-1.12 gast-0.4.0 h5py-3.1.0 keras-2.6.0 tensorboard-2.6.0 tensorboard-data-server-0.6.1 tensorflow-2.6.2 tensorflow-estimator-2.6.0 typing-extensions-3.7.4.3 wrapt-1.12.1\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install -U tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# https://stackoverflow.com/questions/37893755/tensorflow-set-cuda-visible-devices-within-jupyter\n",
    "import os\n",
    "os.environ[\"CUDA_DEVICE_ORDER\"]=\"PCI_BUS_ID\"   # see issue #152\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\"\n",
    "\n",
    "# https://stackoverflow.com/questions/56008683/could-not-create-cudnn-handle-cudnn-status-internal-error\n",
    "import tensorflow as tf\n",
    "gpu_options = tf.compat.v1.GPUOptions(per_process_gpu_memory_fraction=0.95)\n",
    "config = tf.compat.v1.ConfigProto(gpu_options=gpu_options)\n",
    "config.gpu_options.allow_growth = True\n",
    "session = tf.compat.v1.Session(config=config)\n",
    "# tf.compat.v1.enable_eager_execution()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# from tensorflow.python.framework.ops import disable_eager_execution\n",
    "# disable_eager_execution()\n",
    "\n",
    "#tensorflow 2.6 kernel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "\n",
    "import wasserstein_utils\n",
    "import data_utils\n",
    "import losses\n",
    "import networks\n",
    "import deeplabv3 as dlv3\n",
    "import utils\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "# End-End training and testing of the new pytorch-based VGG16-deeplabv3 architecture\n",
    "# https://stackoverflow.com/questions/37893755/tensorflow-set-cuda-visible-devices-within-jupyter\n",
    "import os\n",
    "# os.environ[\"CUDA_DEVICE_ORDER\"]=\"PCI_BUS_ID\"   # see issue #152\n",
    "# os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\"\n",
    "\n",
    "# https://stackoverflow.com/questions/56008683/could-not-create-cudnn-handle-cudnn-status-internal-error\n",
    "import tensorflow as tf\n",
    "gpu_options = tf.compat.v1.GPUOptions(per_process_gpu_memory_fraction=0.95)\n",
    "config = tf.compat.v1.ConfigProto(gpu_options=gpu_options)\n",
    "config.gpu_options.allow_growth = True\n",
    "session = tf.compat.v1.Session(config=config)\n",
    "# tf.compat.v1.enable_eager_execution()\n",
    "import importlib\n",
    "import matplotlib.pyplot as plt\n",
    "# import wasserstein_utils\n",
    "import data_utils\n",
    "import losses\n",
    "import networks\n",
    "import deeplabv3 as dlv3\n",
    "import utils\n",
    "import pickle\n",
    "import time\n",
    "import numpy as np\n",
    "import importlib\n",
    "\n",
    "import faiss\n",
    "\n",
    "import data_utils\n",
    "import losses\n",
    "\n",
    "from IPython import display\n",
    "import umap\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "import numpy as np\n",
    "\n",
    "import data_utils\n",
    "import data_utils_new\n",
    "import losses\n",
    "\n",
    "from IPython import display"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "backbone = 'vgg16'\n",
    "dataset = \"SYNTHIA_VIPER_DENSE_13_1\"\n",
    "\n",
    "# H x W x C\n",
    "img_shape = (512,1024,3)\n",
    "\n",
    "# 13 classes + void\n",
    "num_classes = 14\n",
    "\n",
    "batch_size=4\n",
    "\n",
    "do_training = True\n",
    "\n",
    "epochs=100000\n",
    "epoch_step=250\n",
    "\n",
    "num_projections=100\n",
    "\n",
    "source_data_dir = '/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/SYNTHIA/train/'\n",
    "cityscapes_data_dir = '/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/'\n",
    "\n",
    "fn_w_dlv3 = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + \"/\" + backbone + \"_deeplabv3.h5\"\n",
    "fn_w_cls = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + \"/\" + backbone + \"_deeplabv3_classifier.h5\"\n",
    "\n",
    "fn_w_adapted_dlv3 = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + \"/\" + backbone +\"_deeplabv3_adapted.h5\"\n",
    "fn_w_adapted_cls = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + \"/\" + backbone + \"_deeplabv3_classifier_adapted.h5\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3.h5\n",
      "/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_adapted.h5\n"
     ]
    }
   ],
   "source": [
    "print(fn_w_dlv3)\n",
    "print(fn_w_adapted_dlv3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "importlib.reload(losses)\n",
    "\n",
    "deeplabv3 = dlv3.deeplabv3(activation=None, \\\n",
    "                           backbone=backbone, \\\n",
    "                           num_classes=num_classes)\n",
    "\n",
    "X = deeplabv3.input\n",
    "Y = tf.keras.layers.Input(shape=(img_shape[0], img_shape[1], num_classes,), dtype='float32', name='label_input')\n",
    "\n",
    "C_in = tf.keras.layers.Input(shape=deeplabv3.layers[-1].output_shape[1:], dtype='float32', name='classifier_input')\n",
    "classifier = tf.keras.Model(C_in, networks.classifier_layers(C_in, num_classes = num_classes, activation='softmax'))\n",
    "\n",
    "# A combined model, giving the output of classifier(deeplabv3(X))\n",
    "combined = tf.keras.Model(X, classifier(deeplabv3(X)))\n",
    "combined.compile(loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False))\n",
    "\n",
    "# A model outputting hxwx1 labels for each image. Also useful to verify the\n",
    "# mIoU with Keras' built-in function. Will however also consider the 'ignore' class. \n",
    "combined_ = tf.keras.Model(X, tf.cast(tf.keras.backend.argmax(combined(X), axis=-1), 'float32'))\n",
    "combined_.compile(metrics=[tf.keras.metrics.MeanIoU(num_classes=num_classes)], loss=tf.keras.losses.CategoricalCrossentropy(from_logits=False))\n",
    "\n",
    "# Set up training\n",
    "opt = tf.keras.optimizers.Adam(lr=1e-4, epsilon=1e-8, decay=1e-6)\n",
    "# opt = tf.keras.optimizers.SGD(lr=1e-4, momentum=.9)\n",
    "\n",
    "loss_function = losses.weighted_ce_loss(num_classes, 0)\n",
    "wce_loss = loss_function(Y, classifier(deeplabv3(X)), from_logits=False)\n",
    "\n",
    "# https://stackoverflow.com/questions/55434653/batch-normalization-doesnt-have-gradient-in-tensorflow-2-0\n",
    "params = deeplabv3.trainable_weights + classifier.trainable_weights\n",
    "\n",
    "updates = opt.get_updates(wce_loss, params)\n",
    "\n",
    "train = tf.keras.backend.function(inputs=[X,Y], outputs=[wce_loss], updates=updates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# deeplabv3.load_weights(fn_w_dlv3)\n",
    "# classifier.load_weights(fn_w_cls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt \n",
    "# importlib.reload(data_utils)\n",
    "\n",
    "# do_training = True\n",
    "\n",
    "# # Training on source domain\n",
    "# if do_training == True:\n",
    "#     try:\n",
    "#         deeplabv3.load_weights(fn_w_dlv3)\n",
    "#         classifier.load_weights(fn_w_cls)\n",
    "\n",
    "#         print(\"Successfully loaded model. Continuing training.\")\n",
    "#     except:\n",
    "#         print(\"Could not load previous model weights. Is a new model present?\")\n",
    "        \n",
    "#     start_time = time.time()\n",
    "\n",
    "#     fig,ax = plt.subplots(1,figsize=(10,7))\n",
    "\n",
    "#     loss_history = []\n",
    "\n",
    "#     for itr in range(epochs):\n",
    "#     # for itr in range(1):\n",
    "#         source_train_data, source_train_labels = utils.sample_batch(source_data_dir, batch_size=batch_size, seed=itr)\n",
    "#         source_train_labels = tf.keras.utils.to_categorical(source_train_labels, num_classes=num_classes)\n",
    "\n",
    "#         loss_history.append(train(inputs=[source_train_data, source_train_labels ]))\n",
    "\n",
    "#         if np.isnan(np.sum(loss_history[-1])):\n",
    "#             print('ERROR. ENCOUNTERED NAN')\n",
    "#             break\n",
    "\n",
    "#         if itr%epoch_step == 0:\n",
    "#             if itr != 0:\n",
    "#                 ax.clear()\n",
    "#                 ax.plot(np.asarray(loss_history))\n",
    "\n",
    "#             ax.set_title(\"Training loss on source domain\")\n",
    "#             ax.set_xlabel(\"Epoch\")\n",
    "#             ax.set_ylabel(\"Loss\")\n",
    "\n",
    "#             display.clear_output(wait=True)\n",
    "#             display.display(plt.gcf())\n",
    "#             time.sleep(1e-3)\n",
    "            \n",
    "#         if itr % (epochs // 10) == 0 or itr == epochs - 1:\n",
    "#             deeplabv3.save_weights(fn_w_dlv3)\n",
    "#             classifier.save_weights(fn_w_cls)\n",
    "\n",
    "#     training_time = time.time() - start_time\n",
    "# else:\n",
    "#     deeplabv3.load_weights(fn_w_dlv3)\n",
    "#     classifier.load_weights(fn_w_cls)\n",
    "#     print(\"Loaded model weights\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model weights\n"
     ]
    }
   ],
   "source": [
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "print(\"Loaded model weights\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Performance before adaptation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "import data_utils_new\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "source_cat_iou,source_miou = utils.compute_miou(source_data_dir, combined_, data_utils_new.label_ids_13, [])\n",
    "\n",
    "for k in source_cat_iou:\n",
    "    print(k, source_cat_iou[k])\n",
    "print(source_miou)\n",
    "\n",
    "print('Computed ' + dataset + ' mIoU in', time.time() - start_time)\n",
    "\n",
    "\n",
    "# road 0.8615078545158885\n",
    "# sidewalk 0.9000774695735323\n",
    "# building 0.9430844244367906\n",
    "# traffic light 0.5906543878603067\n",
    "# traffic sign 0.47512503948229323\n",
    "# vegetation 0.8432760653705371\n",
    "# sky 0.9643722032928421\n",
    "# person 0.7238042430944608\n",
    "# rider 0.6191497998404211\n",
    "# car 0.882677391642491\n",
    "# bus 0.911323854999263\n",
    "# motorcycle 0.6885880663029611\n",
    "# bicycle 0.34777010181426876\n",
    "# 0.7501085309404659\n",
    "# Computed SYNTHIA_VIPER_DENSE_13_1 mIoU in 3522.347924232483"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ignore 0.0\n",
      "road 0.46600550969361687\n",
      "sidewalk 0.08673386139638939\n",
      "building 0.47567657106939204\n",
      "traffic light 0.096038243034673\n",
      "traffic sign 0.014404219487528921\n",
      "vegetation 0.13340190671890303\n",
      "sky 0.7299059721207719\n",
      "person 0.1971218314328171\n",
      "rider 0.0\n",
      "car 0.8628852727742756\n",
      "bus 0.01935999478079332\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.22010952732208294\n",
      "Computed VIPER mIoU in 394.8119444847107\n"
     ]
    }
   ],
   "source": [
    "importlib.reload(utils)\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "cit_cat_iou,cit_miou = utils.compute_miou(cityscapes_data_dir, combined_, data_utils_new.label_ids_13, [])\n",
    "\n",
    "for k in cit_cat_iou:\n",
    "    print(k, cit_cat_iou[k])\n",
    "print(cit_miou)\n",
    "\n",
    "print('Computed VIPER mIoU in', time.time() - start_time)\n",
    "\n",
    "# road 0.46610276620252705\n",
    "# sidewalk 0.08673575303732681\n",
    "# building 0.47584591747660204\n",
    "# traffic light 0.09605701550329426\n",
    "# traffic sign 0.014402255800938481\n",
    "# vegetation 0.13339936275003222\n",
    "# sky 0.7302335987742893\n",
    "# person 0.19711547556738282\n",
    "# rider 0.0\n",
    "# car 0.8627417161587038\n",
    "# bus 0.01936740214450999\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.237077020262739\n",
    "# Computed VIPER mIoU in 166.7798900604248"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load pretrained weights for SYNTHIA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# deeplabv3.load_weights(\"/pscratch/sd/c/chubakov/code/mas3-continual/weights/SYNTHIA_VIPER_DENSE_13_3/vgg16_deeplabv3_adapted.h5\")\n",
    "# classifier.load_weights(\"/pscratch/sd/c/chubakov/code/mas3-continual/weights/SYNTHIA_VIPER_DENSE_13_3/vgg16_deeplabv3_classifier_adapted.h5\")\n",
    "# print(\"Loaded model weights\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import data_utils_new\n",
    "importlib.reload(utils)\n",
    "\n",
    "start_time = time.time()\n",
    "means, _, ct = utils.learn_gaussians(source_data_dir, deeplabv3, combined, batch_size, data_utils_new.label_ids_13)\n",
    "print(\"computed means in\", time.time() - start_time)\n",
    "\n",
    "start_time = time.time()\n",
    "means, covs, ct = utils.learn_gaussians(source_data_dir, deeplabv3, combined, batch_size, data_utils_new.label_ids_13, \\\n",
    "                                  initial_means=means)\n",
    "print(\"finished training gaussians in\", time.time() - start_time)\n",
    "\n",
    "np.save(\"./extras/means_1024x512_\" + backbone + \"deeplabv3_\" + dataset + \".npy\", means)\n",
    "np.save(\"./extras/covs_1024x512_\" + backbone + \"deeplabv3_\" + dataset + \".npy\", covs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(means.shape[0]):\n",
    "    print('=====================================================')\n",
    "    print(means[i])\n",
    "    print(covs[i])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# means = np.load(\"./extras/means_1024x512_\" + backbone + \"deeplabv3_\" + dataset + \".npy\")\n",
    "# covs = np.load(\"./extras/covs_1024x512_\" + backbone + \"deeplabv3_\" + dataset + \".npy\")\n",
    "\n",
    "# means = np.load(\"./extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_COMBINED_13.npy\")\n",
    "# covs = np.load(\"./extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_COMBINED_13.npy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "# SYNTHIA -> VIPER1\n",
    "means = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "covs = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "\n",
    "\n",
    "\n",
    "means = np.nan_to_num(means)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "means = np.load(\"./extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "covs = np.load(\"./extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# means = np.load(\"./extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_COMBINED_13.npy\")\n",
    "# covs = np.load(\"./extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_COMBINED_13.npy\")\n",
    "means = np.nan_to_num(means)\n",
    "# means"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6cAAAMoCAYAAAAz+pZ9AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAACSI0lEQVR4nOzdeXxU9b3/8ddJgmBYXACtWiDWqxgFghjcWLRVcalL3Wq9AUVro9DFblb7i1epNr9rr3bB/kTFiloZW+tSra1tKVoVRKvBC9cFrNKbxF3ALRC2JOf3xyQhhIRsMzmzvJ6PRx4zc86Zcz4zgSTv+W5BGIZIkiRJkhSlnKgLkCRJkiTJcCpJkiRJipzhVJIkSZIUOcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUlpJwiCu4Ig+HGCzjU9CILFO9j/ZBAEFzfeLwmCYEEirpsNOnpvJUlqyXAqSeoVQRBUBkGwIQiCdS2+/l/UdXVFGIaxMAynRF1HZwRBMCsIgvmttjUH7VSTyrVJknpHXtQFSJKyyqlhGC6Mugh1TxAEuWEY1kddhyQpM9lyKkmKXGP3z2eCIPh5EAQfB0HwryAIjmrc/mYQBB8EQXBBq6cNCYLgb0EQ1ARB8FQQBCNanO/Axn0fBkHwWhAEX26xb3AQBH8IguDTIAieB/ZrVcvxQRCsDILgk8aW3aBVnYtbPA6DILg0CILXG+u+OQiCoHFfbhAEPw2CYE0QBP8bBME3Go9v84PhIAiuCILg7cbX81oQBMe2OM//CYJgVeO+pUEQDGvcN7vx/fm0cfukxu0nAv8HOLexhXp5EATlwCTg/7Vste7gvborCIJbgiB4LAiC9cDn26g7DILgW43fszVBENwQBEGbf180fk9faHxvXwiC4KjG7W3WJknKLoZTSVKqOBz4H2AwcC/wW2A88G/AVOLBZUCL40uA64AhwDIgBhAEQX/gb43n2AP4CjAnCIKDGp93M7AR2Au4qPGLxucOAR4Crmo87ypgQgd1n9JY5xjgy8AJjdu/BpwEjAXGAV9q7wRBEIwEvgGMD8NwYOM5Kht3fxc4DzgZGNRYb23jvhcaz7974+u9PwiCfmEY/gX4v8B9YRgOCMOwKAzDMmAR8I3Gbd/oxHsF8O9AOTAQaG/86BlAcePrPJ0W72mL17g78CfgJuLf458BfwqCYHBbtbX3XkmSMpfhVJLUmx5ubGFs+vpai33/G4bhnY3dRu8DhgHXhmG4KQzDBcBm4kG1yZ/CMHw6DMNNQBlwZGOL4ilAZeO56sIw/G/gQeCcIAhygbOAq8MwXB+G4cvA3S3OeTLwShiGD4RhuAX4BfBeB6/p+jAMPw7DsBr4O/GwCPGgOjsMw7fCMPwIuH4H56gH+gIHBUHQJwzDyjAMVzXuuxi4KgzD18K45WEYrgUIw3B+GIZrG1/nTxvPMbKDeltq971qccwjYRg+E4ZhQxiGG9s5z0/CMPyw8T34BfEw3doXgdfDMLyn8Vq/AVYCp3ahXklSBjOcSpJ605fCMNy1xdftLfa93+L+BoAwDFtva9ly+mbTnTAM1wEfAnsDI4DDW4Zg4q2snwGGEp9v4c0W56lqcX/vVucNWx3blpbhtbZFjduca0fnCcPwDeDbwCzggyAIfhsEwd6Nu4cRb8HdThAE3w+CYEVjN9mPgV2It/h21o7eq+3qDoLglWDrZFaT2nltVcRfe2t7s+173XTsPl2oV5KUwQynkqR0NazpTmN3392Bd4gHpadaheABYRjOAFYDdS2fCwxvcf/dVucNWh3bFe8Cn22r3raEYXhvGIYTiQfGEPhJ4643aTUutrG2ScAPiLfQ7haG4a7AJ2wdIxu2dZlWj3f0Xm33nDAMD27cPyAMw0XtvLbhxL8Prb3T+NpodezbO6hXkpRFDKeSpHR1chAEE4Mg2In42NPnwjB8E/gjcEAQBNOCIOjT+DU+CILCxi7DDwGzgiDIbxxb2XKipT8BBwdBcGbjxEXfYttWxK74HXBZEAT7BEGwK3BFewcGQTAyCIIvBEHQl/h42A1AQ+PuXwHXBUGwfxA3JgiCwcTHgNYRD9x5QRBcTXxMapP3gYJWkxO9D3yuxeN236suvtbLgyDYrbFb9WXEu2W39ljjtf49CIK8IAjOBQ5qrKGt2iRJWcZwKknqTY8G265z+vsenOte4Bri3XkPJT5pEmEY1gBTiE/u8w7xbrc/IT4eE+ITDw1o3H4XcGfTCcMwXEN8vOX1wFpgf+CZbtZ3O7CA+CRP/008nNURH1/aWt/Ga65prGsP4IeN+35GPOguAD4F7gB2Bv4K/AX4J/HusRvZtnvt/Y23a4MgeLHx/mzg7CAIPgqC4KZOvFed9QiwlPjEVH9qrHEbjeNkTwG+R/y9/QFwSuN7vl1tXby+JCkDBPHhNJIkKZmCIDgJuDUMw9ZdW9NaEAQhsH/juFlJkrrNllNJkpIgCIKdgyA4ubEL6z7EW3l70lIsSVJGM5xKkpQcAfAj4CPi3XpXAFdHWpEkSSnMbr2SJEmSpMjZcipJkiRJipzhVJIkSZIUubyoC2hpyJAhYUFBQdRlSJIkSZKSYOnSpWvCMBza1r6UCqcFBQVUVFREXYYkSZIkKQmCIKhqb5/deiVJkiRJkTOcSpIkSZIiZziVJEmSJEUupcactmXLli289dZbbNy4MepSMlq/fv347Gc/S58+faIuRZIkSVIWSvlw+tZbbzFw4EAKCgoIgiDqcjJSGIasXbuWt956i3333TfqciRJkiRloZTv1rtx40YGDx5sME2iIAgYPHiwrdOSJEmSIpPy4RQwmPYC32NJkiRJUUqLcBq1o446KuoSJEmSJCmjGU47YcmSJUk9f11dXVLPL0mSJEmpLvPCaeyPUHAs5Bwcv439scenHDBgAAANDQ3MnDmTAw88kOOPP56TTz6ZBx54AICCggKuueYaxo0bx+jRo1m5ciUAH374IV/60pcYM2YMRxxxBP/zP/8DwKxZs5g2bRoTJkxg2rRprF69mrPOOovx48czfvx4nnnmmR7XLUmSJEnpIuVn6+2S2B+h9GqobZzYp+rd+GOAklN6fPqHHnqIyspKXn31VT744AMKCwu56KKLmvcPGTKEF198kTlz5nDjjTfyq1/9imuuuYZDDjmEhx9+mCeeeILzzz+fZcuWAfDqq6+yePFidt55Z/793/+d73znO0ycOJHq6mpOOOEEVqxY0eOaJUmSJCkdZFY4Lfv51mDapHZjfHsCwunixYs555xzyMnJ4TOf+Qyf//znt9l/5plnAnDooYfy0EMPNT/nwQcfBOALX/gCa9eu5dNPPwXgtNNOY+eddwZg4cKFvPrqq83n+vTTT1m3bl1zq60kSZIkZbLMCqfV73Vte4L17dsXgNzc3E6NI+3fv3/z/YaGBp577jn69euXtPokSZIkKVVl1pjT4Z/p2vYumjBhAg8++CANDQ28//77PPnkkx0+Z9KkScRiMQCefPJJhgwZwqBBg7Y7bsqUKfzyl79sftzU9VeSJEmSskFmhdPy70B+q5bH/H7x7Qlw1lln8dnPfpaDDjqIqVOnMm7cOHbZZZcdPmfWrFksXbqUMWPGcOWVV3L33Xe3edxNN91ERUUFY8aM4aCDDuLWW29NSM2SJEmSlA6CMAyjrqFZcXFxWFFRsc22FStWUFhY2PmTxP4YH2Na/V68xbT8OwkZb9qkaRzo2rVrOeyww3jmmWf4zGcS0zIbtS6/15IkSZLUBUEQLA3DsLitfZk15hTiQTSBYbS1U045hY8//pjNmzfzH//xHxkTTCVJkiQpSpkXTpOsM+NMJUmSJEldk1ljTiVJkiRJaclwKkmSJEmKnOFUkiRJkhQ5w6kkSZIkKXKG0142a9YsbrzxxqjLkCRJkqSUYjjtojAMaWhoiLoMSZIkScoomRdO/xqDMwrgqJz47V9jPT5lZWUlI0eO5Pzzz2fUqFF89atfZdSoUYwePZr77rsPgHXr1nHssccybtw4Ro8ezSOPPNL8/PLycg444AAmTpzIa6+91uN6JEmSJCnTZNY6p3+NwfWlsLE2/vi9qvhjgBNKenTq119/nbvvvpu3336bW2+9leXLl7NmzRrGjx/P5MmTGTp0KL///e8ZNGgQa9as4YgjjuC0007jxRdf5Le//S3Lli2jrq6OcePGceihh/bwhUqSJElSZsmsltNby7YG0yYba+Pbe2jEiBEcccQRLF68mPPOO4/c3Fz23HNPjj76aF544QXCMOT//J//w5gxYzjuuON4++23ef/991m0aBFnnHEG+fn5DBo0iNNOO63HtUiSJElSpsmsltP3q7u2vQv69++/w/2xWIzVq1ezdOlS+vTpQ0FBARs3buzxdSVJkiQpG2RWy+mew7u2vRsmTZrEfffdR319PatXr+bpp5/msMMO45NPPmGPPfagT58+/P3vf6eqqgqAyZMn8/DDD7NhwwZqamp49NFHE1aLJEmSJGWKzGo5vbR82zGnAP3y49sT5IwzzuDZZ5+lqKiIIAj4r//6Lz7zmc9QUlLCqaeeyujRoykuLubAAw8EYNy4cZx77rkUFRWxxx57MH78+ITVIkmS0keMFZSxiGpqGM5AyplECYVdPkaSMlUQhmHUNTQrLi4OKyoqttm2YsUKCgu78EP5r7H4GNP3q+MtppeW93gypGzR5fdakiQ1B8oqasgloJ6tf1vlADuTx3rq2nzuDIqYwD5cwoJ2jwEY0RhUn+Ft5vI/zdcIgBDIJaCUMczhuAS+MklKvCAIloZhWNzmvowLp+o232tJkjqnZSBNRf3J4zam2OoqKeXsKJxmVrdeSZKkJJvJQm5hedRl7NB66pjGYwAGVElpI7MmRJIkSUqiGCtSPpg2CYHzeYwYK6IuRZI6xXAqSZLUSZfxeNQldEkDMI3HCPgpBcw1qEpKaXbrlSRJ6oQYK1jLpqjL6LKm2UWqqGEajzGVxxjBQE7mczzGv5wZWFLKMJxKkiR1Qrq1mralZVBt2T25ihpKWQA4RlVSdOzW200XX3wxr7766nbb77rrLr7xjW8k5ZqVlZWMGjUq6deRJElxMVYwhJsJ+Glatpp2RS11TG3sAtz0NYDZdgWW1GtsOe2mX/3qV1GXIEmSkijGCi7iL2ymIepSIrOeOi5w1l9JvSTzWk5fj8G9BTA3J377eqzHp1y/fj1f/OIXKSoqYtSoUdx3330cc8wxNK3Jeuedd3LAAQdw2GGH8cwzzzQ/b/Xq1Zx11lmMHz+e8ePHN+8bPXo0H3/8MWEYMnjwYH79618DcP755/O3v/2NyspKJk2axLhx4xg3bhxLlizZYX1/+tOfOPLII1mzZk2PX6skSYoH0wv4c1YH0yb1ZEaXZkmpL7PC6esxWFQK66qAMH67qLTHAfUvf/kLe++9N8uXL+fll1/mxBNPbN737rvvcs011/DMM8+wePHibbr6XnbZZXznO9/hhRde4MEHH+Tiiy8GYMKECTzzzDO88sorfO5zn2PRokUAPPvssxx11FHsscce/O1vf+PFF1/kvvvu41vf+la7tf3+97/n+uuv57HHHmPIkCE9ep2SJGW7pm68U3mM+uYRmlrLJmayMOoyJGW4zOrW+0IZ1NVuu62uNr59/5Jun3b06NF873vf44orruCUU05h0qRJzfv+8Y9/cMwxxzB06FAAzj33XP75z38CsHDhwm3C6qeffsq6deuYNGkSTz/9NCNGjGDGjBnMnTuXt99+m912243+/fvzySef8I1vfINly5aRm5vbfL7WnnjiCSoqKliwYAGDBg3q9uuTJEnxYFrKAmqpi7qUlHQry5nAPnbvlZQ0mdVyuq66a9s76YADDuDFF19k9OjRXHXVVVx77bWdel5DQwPPPfccy5YtY9myZbz99tsMGDCAyZMns2jRIhYtWtQcbB944IHm0Pvzn/+cPffck+XLl1NRUcHmzZvbPP9+++1HTU1Nu+FVkiR1LMYKCpjLVB4zmO5ACJSxKOoyJGWwzAqnA4Z3bXsnvfPOO+Tn5zN16lQuv/xyXnzxxeZ9hx9+OE899RRr165ly5Yt3H///c37pkyZwi9/+cvmx8uWLQNg2LBhrFmzhtdff53Pfe5zTJw4kRtvvJHJkycD8Mknn7DXXnuRk5PDPffcQ319fZt1jRgxggcffJDzzz+fV155pUevUZKkbBRjBRfyZ6qoibqUtFBFjbP3SkqazAqn48shL3/bbXn58e098NJLL3HYYYcxduxYfvSjH3HVVVc179trr72YNWsWRx55JBMmTKCwcGtXl5tuuomKigrGjBnDQQcdxK233tq87/DDD+eAAw4AYNKkSbz99ttMnDgRgJkzZ3L33XdTVFTEypUr6d+/f7u1HXjggcRiMc455xxWrVrVo9cpSVK2mcZjbHFsaZdcwGMM4f+Rw08pYK5hVVLCBGGYOj+Qi4uLw6YZcJusWLFim8DXoddj8TGm66rjLabjy3s03jSbdPm9liQpTcVYwTQeM5YmQD55zGWKY1EldUoQBEvDMCxua19mTYgE8SBqGJUkSe1oWr/UYJoYtdRRxiLDqaQey6xuvZIkSR0oY5HrlyZYtWN2JSWA4VSSJGUVg1Ti7U7fqEuQlAEMp5IkKasMZ2DUJWScGrY4MZKkHjOcSpKkrFLOpKhLyDibaeAC/uwMvpJ6xHAqSZKySgmFHMTuUZeRceoJCYmvhVrKAgOqpC4znHZCZWUlo0aN6vTxf/jDH7j++usBmDVrFjfeeOMOz1lRUcG3vvWtxBQrSZI6tJ4tUZeQ0WqpYyqP2YoqqUsybymZFHDaaadx2mmndfr44uJiiovbXOpHkiQlgZMi9Y4qariQPwO41IykDmVey2l9DOoKoC4nflsfS8hp6+rqKCkpobCwkLPPPpva2loKCgpYs2YNEG/9POaYYwC46667+MY3vrHdOZYuXUpRURFFRUXcfPPNzduffPJJTjnlFCDe0nrRRRdxzDHH8LnPfY6bbrqp+bjrrruOkSNHMnHiRM4777w2W2QlSVLHnBSp92wh5DIej7oMSWkgs8JpfQzCUqAKGkc9EJYmJKC+9tprzJw5kxUrVjBo0CDmzJnT5XNceOGF/PKXv2T58uU7PG7lypX89a9/5fnnn+dHP/oRW7Zs4YUXXuDBBx9k+fLl/PnPf6aioqK7L0WSpKxXziTy7UDWa9ayKeoSJKWBzAqnYRlQ22pjbeP2nhk2bBgTJkwAYOrUqSxevLhLz//444/5+OOPmTx5MgDTpk1r99gvfvGL9O3blyFDhrDHHnvw/vvv88wzz3D66afTr18/Bg4cyKmnntr9FyNJUpYroZC5TMmwP4QkKb1l2M/k6i5u77wgCLZ7nJeXR0NDAwAbN27s8TWa9O27dSHr3Nxc6urqEnZuSZIUV0IhYdRFZJGZLIy6BEkpLsPC6fAubu+86upqnn32WQDuvfdeJk6cSEFBAUuXLgXgwQcf3OHzd911V3bdddfmFtdYrGtdjSdMmMCjjz7Kxo0bWbduHX/84x+78SokSVJLjj3tPbex42FNkpRZ4TQoB/Jbbcxv3N4zI0eO5Oabb6awsJCPPvqIGTNmcM0113DZZZdRXFxMbm5uh+e48847+frXv87YsWMJw659Vjt+/HhOO+00xowZw0knncTo0aPZZZdduvtyJEkSjj3tTQ1RFyAp5QVdDUnJVFxcHLae6GfFihUUFnZh6vH6WOMY02pgeDyY5pYktM6orFu3jgEDBlBbW8vkyZOZO3cu48aNS9j5u/xeS5KUAWKs4AL+TL2dfJMu5HtRlyApYkEQLA3DsM11NDPvo8LcEiAzwmhrpaWlvPrqq2zcuJELLrggocFUkqRs1bT+5jQeM54mUdDxIZKyXOaF0wx27733Rl2CJEkZ6RneNpgmWUh8UqQ5HBd1KZJSVGaNOZUkSeqGufxP1CVkBd9nSTtiOJUkSVnP8aa9w/dZ0o4YTiVJUtbLdUSkJEXOcCpJkrJeKWOiLkGSsp7htAMff/wxc+bM6dZz77//fgoLC/n85z8PwHnnnceYMWP4+c9/ztVXX83ChQs7dZ7KykpGjRoFQEVFBd/61rc6fXxrd911F++8804XXoUkSZlvDsfR8Yrl6qkRDIy6BEkpzNl6O9AUTmfOnLndvrq6OvLy2n8L77jjDm6//XYmTpzIe++9xwsvvMAbb7zRo3qKi4spLm5zWaBOueuuuxg1ahR77713j+qQJCnTOBoy+aqoccZeSe3KuJbTVcS4jwLmkcN9FLCKWI/Od+WVV7Jq1SrGjh3L5ZdfzpNPPsmkSZM47bTTOOiggwD40pe+xKGHHsrBBx/M3LlzAbj22mtZvHgxX/3qV7n88suZMmUKb7/9NmPHjmXRokVMnz6dBx54AIAXXniBo446iqKiIg477DBqamrarefJJ5/klFNOAWD16tUcf/zxHHzwwVx88cWMGDGCNWvWAFBfX8/XvvY1Dj74YKZMmcKGDRt44IEHqKiooKSkhLFjx7Jhw4YevTeSJGWShqgLyBK3sJyZdK73mKTsklHhdBUxFlPKeqqAkPVUsZjSHgXU66+/nv32249ly5Zxww03APDiiy8ye/Zs/vnPfwIwb948li5dSkVFBTfddBNr167l6quvpri4mFgsxg033MAf/vCH5vNMmjSp+fybN2/m3HPPZfbs2SxfvpyFCxey8847d6q2H/3oR3zhC1/glVde4eyzz6a6urp53+uvv87Xv/51XnnlFXbddVcefPBBzj777Oaali1b1unrSJKUDZwUqfe4pIyktmRUOK2gjHpqt9lWTy0VlCX0Oocddhj77rtv8+ObbrqJoqIijjjiCN58801ef/31Tp/rtddeY6+99mL8+PEADBo0aIddhVtavHgxX/nKVwA48cQT2W233Zr37bvvvowdOxaAQw89lMrKyk7XJElSNnJSpN7jkjKS2pJR4XQ91V3a3l39+/dvvv/kk0+ycOFCnn32WZYvX84hhxzCxo0bE3q97ujbt2/z/dzcXOrq6iKsRpKk1DeH4ziWYVGXkRVspZbUlowKp/0Z3qXtnTFw4MAdjgH95JNP2G233cjPz2flypU899xzXTr/yJEjeffdd3nhhRcAqKmp6XSQnDBhAr/73e8AWLBgAR999FGHz+no9UiSlM0W8uWoS8gKtlJLaktGhdNiysklf5ttueRTTHm3zzl48GAmTJjAqFGjuPzyy7fbf+KJJ1JXV0dhYSFXXnklRxxxRJfOv9NOO3HffffxzW9+k6KiIo4//vhOt7xec801LFiwgFGjRnH//ffzmc98hoEDdzxF+/Tp07n00kudEEmSpHYMpl/UJWQ8Z+uV1JYgDFOnz39xcXFYUVGxzbYVK1ZQWFjY6XOsIkYFZaynmv4Mp5hy9qMk0aWmhE2bNpGbm0teXh7PPvssM2bMYNmyZd0+X1ffa0mSMlGMFVzEX9js/L1JE/K9qEuQFJEgCJaGYdjm2pgZt87pfpRkbBhtrbq6mi9/+cs0NDSw0047cfvtt0ddkiRJaa+E+Ae1U3ks4koyk6NNJbUn48JpNtl///357//+76jLkCQp45RQSBmLqMJ5GhItdfrsSUo1GTXmVJIkKVHKmUS+n+Mn3Ah2PD+GpOzlT1xJkqQ2NHXvvYA/uy5nApUzKeoSJKUoW04lSZLaUUIhd3OSLagJcizDmkO/JLVmOJUkSdqBEgqZy5TG7qghjprsulwCZlDkOrKSdshw2oGPP/6YOXPmdOu5999/P4WFhXz+858H4LzzzmPMmDH8/Oc/5+qrr2bhwoU9qu3kk0/m448/7tE5JElSx0oopJJSJvID9udeDKidl0tAHd91bVNJHbKPSgeawunMmTO321dXV0deXvtv4R133MHtt9/OxIkTee+993jhhRd44403ElbbY485xb0kSb1pD4YT8t9UcxKb2D3qctKC43UldVbGtZwuIMZZFDCJHM6igAXEenS+K6+8klWrVjF27Fguv/xynnzySSZNmsRpp53GQQcdBMCXvvQlDj30UA4++GDmzp0LwLXXXsvixYv56le/yuWXX86UKVN4++23GTt2LIsWLWL69Ok88MADALzwwgscddRRFBUVcdhhh1FTs+209e+++y6TJ09m7NixjBo1ikWLFgFQUFDAmjVrALjuuusYOXIkEydO5LzzzuPGG28E4JhjjuGKK67gsMMO44ADDmh+riRJ6rpLKKcv+Qznz9h62jm5rmwqqZMyquV0ATF+QimbqAXgfar4CaUATKGkW+e8/vrrefnll1m2bBkATz75JC+++CIvv/wy++67LwDz5s1j9913Z8OGDYwfP56zzjqLq6++mieeeIIbb7yR4uJivv71r3PKKac0n+eOO+4AYPPmzZx77rncd999jB8/nk8//ZSdd955mxruvfdeTjjhBMrKyqivr6e2tnab/S+88AIPPvggy5cvZ8uWLYwbN45DDz20eX9dXR3PP/88jz32GD/60Y963J1YkqRs1fT3xG2U8b/UUkf/iCtKfaWMiboESWkio1pOb6OsOZg22UQtt1GW0OscdthhzcEU4KabbqKoqIgjjjiCN998k9dff73T53rttdfYa6+9GD9+PACDBg3arqvw+PHjufPOO5k1axYvvfQSAwduuz7YM888w+mnn06/fv0YOHAgp5566jb7zzzzTAAOPfRQKisru/JSJUlSK1Mo4UEquYtznMW3A33JcayppE5LSDgNgmBeEAQfBEHwcotts4IgeDsIgmWNXycn4lo78gHVXdreXf37b/2U9Mknn2ThwoU8++yzLF++nEMOOYSNGzcm9HqTJ0/m6aefZp999mH69On8+te/7tLz+/btC0Bubi51dXUJrU2SpGzVchbfABhMv6hLSikBcAcnRl2GpDSSqJbTu6DNnz4/D8NwbONX0mfv2YPhXdreGQMHDtxuDGhLn3zyCbvtthv5+fmsXLmS5557rkvnHzlyJO+++y4vvPACADU1NdsFyKqqKvbcc0++9rWvcfHFF/Piiy9us3/ChAk8+uijbNy4kXXr1vHHP/6xSzVIkqTuaZrFt4HvsYavM5+TyY26qBTQlxzu4WTXNJXUJQnpixKG4dNBEBQk4lw9cQnl24w5BehLPpdQ3u1zDh48mAkTJjBq1ChOOukkvvjFL26z/8QTT+TWW2+lsLCQkSNHcsQRR3Tp/DvttBP33Xcf3/zmN9mwYQM777wzCxcuZMCAAc3HPPnkk9xwww306dOHAQMGbNdyOn78eE477TTGjBnDnnvuyejRo9lll126/ZolSVL3lFDYHMhirKCUBdSSfb2W7uBEg6mkLgvCMDEzzTWG0z+GYTiq8fEsYDrwKVABfC8Mw492dI7i4uKwoqJim20rVqygsLDzP9wWEOM2yviAavZgOJdQ3u3JkNLJunXrGDBgALW1tUyePJm5c+cybty4Lp2jq++1JEnasRgrKGMRVbTfCysTjWAglY2TUkpSS0EQLA3DsLitfckcxX8LcB3xedavA34KXNRGcaUQ/+k1fHj3u982mUJJVoTR1kpLS3n11VfZuHEjF1xwQZeDqSRJSryWLakzWcgtLI+4ot5RnWVhXFJiJC2chmH4ftP9IAhuB9ocCBmG4VxgLsRbTpNVT6a79957oy5BkiTtwByOYw7HEWMFF/FnNmfwOqm78xH3UUAx5eyXhY0GkronaUvJBEGwV4uHZwAvt3esJElStiihkE18l/mczGD6RlbHYPqRS5Dw8+ZSxxk8xnqqWEwpq4gl/BqSMlNCWk6DIPgNcAwwJAiCt4BrgGOCIBhLvFtvJXBJIq4lSZKUCVp2+YWt41OrqSGfPNYncSKlfPKYzRcAEjBp09YW4AGs5ys8wpH8NwD11FJBma2nkjolUbP1ntfG5jsScW5JkqRs0DqsHsfveJw3e3zeEQzkZD7HY/yLamoYzkDKmbTNtZpC8XAG8m/syhO82WGn4wH04VaObz7PPHKgjWetT/B685IyVzInRJIkSVI3LeTL27Sm7k5fIGAtGwnYNgbuRA51NNDQYlsfAu7kpA6XdGkdioHtrruJBtaxBYDB9GU2x273nP4MZz1V252/fw/Wm5eUXQynEXv44Yc54IADOOigg7p03NVXX83kyZM57rjjeqNMSZIUgbaCI2wbHptaQoHttnV3rdH2rrsjxZSzmFLqW6w3n0s+xT1Yb15SdjGcRuzhhx/mlFNO6VQ4bXnctdde2xvlSZKkFNReeOxuGE2EpnGlFZSxnmr6M9zZeiV1SdJm641KjEcp4PPkcCAFfJ4Yj/bofFdeeSU333xz8+NZs2Zx4403csMNNzB+/HjGjBnDNddc07z/uuuuY+TIkUycOJHzzjuPG2+8EYBVq1Zx4okncuihhzJp0iRWrlzJkiVL+MMf/sDll1/O2LFjWbVqFbfffjvjx4+nqKiIs846i9ra2jaPmz59Og888AAAjz/+OIcccgijR4/moosuYtOmTQAUFBRwzTXXMG7cOEaPHs3KlSt79F5IkiTtyH6UcC6VXEQD51JpMJXUJRkVTmM8SilXUcU7hIRU8Q6lXNWjgHruuefyu9/9rvnx7373O4YOHcrrr7/O888/z7Jly1i6dClPP/00L7zwAg8++CDLly/nz3/+MxUVFc3PKy0t5Ze//CVLly7lxhtvZObMmRx11FGcdtpp3HDDDSxbtoz99tuPM888kxdeeIHly5dTWFjIHXfc0eZxTTZu3Mj06dO57777eOmll6irq+OWW25p3j9kyBBefPFFZsyY0RyUJUmSJCnVZFS33jJ+Ri0bt9lWy0bK+BklnNqtcx5yyCF88MEHvPPOO6xevZrddtuNl156iQULFnDIIYcAsG7dOl5//XVqamo4/fTT6devH/369ePUU09t3r9kyRLOOeec5vM2tW629vLLL3PVVVfx8ccfs27dOk444YQd1vfaa6+x7777csABBwBwwQUXcPPNN/Ptb38bgDPPPBOAQw89lIceeqhb74EkSZIkJVtGhdNq3u3S9s4655xzeOCBB3jvvfc499xzqaqq4oc//CGXXLLt0q2/+MUv2nx+Q0MDu+66K8uWLevwWtOnT+fhhx+mqKiIu+66iyeffLJHtfftG1/cOzc3l7q65K2XJkmSJEk9kVHdeoezV5e2d9a5557Lb3/7Wx544AHOOeccTjjhBObNm8e6desAePvtt/nggw+YMGECjz76KBs3bmTdunX88Y9/BGDQoEHsu+++3H///QCEYcjy5csBGDhwIDU1Nc3XqqmpYa+99mLLli3EYrHm7a2PazJy5EgqKyt54403ALjnnns4+uije/R6JUmSJKm3ZVQ4Lee75NNvm2359KOc7/bovAcffDA1NTXss88+7LXXXkyZMoV///d/58gjj2T06NGcffbZ1NTUMH78eE477TTGjBnDSSedxOjRo9lll10AiMVi3HHHHRQVFXHwwQfzyCOPAPCVr3yFG264gUMOOYRVq1Zx3XXXcfjhhzNhwgQOPPDA5hpaH9ekX79+3HnnnZxzzjmMHj2anJwcLr300h69XkmSJEnqbUEYhh0f1UuKi4vDlpMIAaxYsYLCws5Pix7jUcr4GdW8y3D2opzvdnu8aXesW7eOAQMGUFtby+TJk5k7dy7jxo3rtev3RFffa0mSJEnqiiAIloZhWNzWvowacwpQwqm9GkZbKy0t5dVXX2Xjxo1ccMEFaRNMJUmSJClKGRdOo3bvvfdGXYIkSZIkpZ2MGnMqSZIkSUpPhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5x2U0FBAWvWrIm6DEmSJEnKCIZTSZIkSVLkMi6cxlhBAXPJ4acUMJcYK3p8zvXr1/PFL36RoqIiRo0axX333de8b8OGDZx00kncdttt7L///qxevRqAhoYG/u3f/q35sSRJkiSpfRkVTmOsoJQFVFFDCFRRQykLehxQ//KXv7D33nuzfPlyXn75ZU488UQA1q1bx6mnnsp5553HJZdcwtSpU4nFYgAsXLiQoqIihg4d2tOXJUmSJEkZL6PCaRmLqKVum2211FHGoh6dd/To0fztb3/jiiuuYNGiReyyyy4AnH766Vx44YWcf/75AFx00UX8+te/BmDevHlceOGFPbquJEmSJGWLjAqn1dR0aXtnHXDAAbz44ouMHj2aq666imuvvRaACRMm8Je//IUwDAEYNmwYe+65J0888QTPP/88J510Uo+uK0mSJEnZIqPC6XAGdml7Z73zzjvk5+czdepULr/8cl588UUArr32WnbbbTe+/vWvNx978cUXM3XqVM455xxyc3N7dF1JkiRJyhYZFU7LmUQ+edtsyyePcib16LwvvfQShx12GGPHjuVHP/oRV111VfO+2bNns2HDBn7wgx8AcNppp7Fu3Tq79EqSJElSF+R1fEj6KKEQiI89raaG4QyknEnN27vrhBNO4IQTTthmW2VlZfP9O++8s/n+8uXLKSoq4sADD+zRNSVJkiQpm2RUOIV4QO1pGO2u66+/nltuuaV5xl5JkiRJUudkVLfeqF155ZVUVVUxceLEqEuRJEmSpLRiOJUkSZIkRc5wKkmSJEmKnOFUkiRJkhQ5w6kkSZIkKXKGU0mSJElS5AynSVZXVxd1CZIkSZKU8jIunMbYRAEfk8OHFPAxMTb1+JyVlZUceOCBlJSUUFhYyNlnn01tbS1Lly7l6KOP5tBDD+WEE07g3XffBeCYY47h29/+NsXFxcyePZv777+fUaNGUVRUxOTJkwHYuHEjF154IaNHj+aQQw7h73//OwB33XUXZ555JieeeCL7778/P/jBD3pcvyRJkiSluryoC0ikGJsoZT21jY+raKCU9QCU0LdH537ttde44447mDBhAhdddBE333wzv//973nkkUcYOnQo9913H2VlZcybNw+AzZs3U1FRAcDo0aP561//yj777MPHH38MwM0330wQBLz00kusXLmSKVOm8M9//hOAZcuW8d///d/07duXkSNH8s1vfpNhw4b1qH5JkiRJSmUZ1XJaxobmYNqktnF7Tw0bNowJEyYAMHXqVP7617/y8ssvc/zxxzN27Fh+/OMf89ZbbzUff+655zbfnzBhAtOnT+f222+nvr4egMWLFzN16lQADjzwQEaMGNEcTo899lh22WUX+vXrx0EHHURVVVWP65ckSZKkVJZRLafVNHRpe1cEQbDN44EDB3LwwQfz7LPPtnl8//79m+/feuut/OMf/+BPf/oThx56KEuXLt3htfr23drKm5ub67hVSZIkSRkvo1pOh7fzctrb3hXV1dXNQfTee+/liCOOYPXq1c3btmzZwiuvvNLmc1etWsXhhx/Otddey9ChQ3nzzTeZNGkSsVgMgH/+859UV1czcuTIHtcpSZIkSekoo8JpOTuT32pbfuP2nho5ciQ333wzhYWFfPTRR3zzm9/kgQce4IorrqCoqIixY8eyZMmSNp97+eWXM3r0aEaNGsVRRx1FUVERM2fOpKGhgdGjR3Puuedy1113bdNiKkmSJEnZJAjDMOoamhUXF4dNkwg1WbFiBYWFhZ0+R4xNlLGBahoYTg7l7NzjyZAqKys55ZRTePnll3t0nlTX1fdakiRJkroiCIKlYRgWt7Uvo8acQnxW3p6GUUmSJElS78qobr3JUlBQkPGtppIkSZIUJcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5x2w8knn8zHH3+83fZZs2Zx44039n5BkiRJkpTmMm4pmWQLw5A//vGP5OT0PNfX1dWRl+e3QJIkSZIyruU09iYULICcR+K3sTd7fs7KykpGjhzJ+eefz6hRo8jNzWXNmjUAlJeXc8ABBzBx4kRee+215uesWrWKE088kUMPPZRJkyaxcuVKAKZPn86ll17K4Ycfzg9+8IOeFydJkiRJGSCjmu1ib0Lpcqitjz+u2hB/DFAyrGfnfv3117n77rs54ogjKCgoAGDp0qX89re/ZdmyZdTV1TFu3DgOPfRQAEpLS7n11lvZf//9+cc//sHMmTN54oknAHjrrbdYsmQJubm5PStKkiRJkjJERoXTshVbg2mT2vr49p6G0xEjRnDEEUdss23RokWcccYZ5OfnA3DaaacBsG7dOpYsWcI555zTfOymTZua759zzjkGU0mSJElqIaPCafWGrm3viv79+3f62IaGBnbddVeWLVvW43NJkiRJUjbIqDGnw3fu2vaemjx5Mg8//DAbNmygpqaGRx99FIBBgwax7777cv/99wPxSZSWL1+enCIkSZIkKQNkVDgtL4T8Vr1l83Pj25Nh3LhxnHvuuRQVFXHSSScxfvz45n2xWIw77riDoqIiDj74YB555JHkFCFJkiRJGSAIwzDqGpoVFxeHFRUV22xbsWIFhYWdT5exN+NjTKs3xFtMywt7Pt40W3T1vZYkSZKkrgiCYGkYhsVt7cuoMacQD6KGUUmSJElKLxnVrVeSJEmSlJ4Mp5IkSZKkyBlOJUmSJEmRM5xKkiRJkiJnOJUkSZIkRc5wKkmSJEmKnOE0ierq6qIuQZIkSZLSQsaF09hTUHAx5Hwpfht7KjHn/fWvf82YMWMoKipi2rRpPProoxx++OEccsghHHfccbz//vsAzJo1i2nTpjFhwgSmTZuWmItLkiRJUobLi7qARIo9BaU3Q+2m+OOq1fHHACVHd/+8r7zyCj/+8Y9ZsmQJQ4YM4cMPPyQIAp577jmCIOBXv/oV//Vf/8VPf/pTAF599VUWL17Mzjvv3MNXJEmSJEnZIaPCadk9W4Npk9pN8e09CadPPPEE55xzDkOGDAFg991356WXXuLcc8/l3XffZfPmzey7777Nx5922mkGU0mSJEnqgozq1lu9pmvbe+Kb3/wm3/jGN3jppZe47bbb2LhxY/O+/v37J/6CkiRJkpTBMiqcDh/Ste2d9YUvfIH777+ftWvXAvDhhx/yySefsM8++wBw99139+wCkiRJkpTlMiqclk+D/L7bbsvvG9/eEwcffDBlZWUcffTRFBUV8d3vfpdZs2ZxzjnncOihhzZ395UkSZIkdU8QhmHUNTQrLi4OKyoqttm2YsUKCgsLO32O2FPxMabVa+ItpuXTejbeNJt09b2WJEmSpK4IgmBpGIbFbe3LqAmRIB5EDaOSJEmSlF4yqluvJEmSJCk9GU4lSZIkSZEznEqSJEmSImc4lSRJkiRFznAqSZIkSYqc4bQTKisrGTVqVNRlSJIkSVLGMpxKkiRJkiKXceH0oRiML4B9cuK3D8USc966ujpKSkooLCzk7LPPpra2loKCAtasWQNARUUFxxxzDABPPfUUY8eOZezYsRxyyCHU1NQkpghJkiRJylAZFU4fisHlpfB2FYRh/Pby0sQE1Ndee42ZM2eyYsUKBg0axJw5c9o99sYbb+Tmm29m2bJlLFq0iJ133rnnBUiSJElSBsuocPqfZbChdtttG2rj23tq2LBhTJgwAYCpU6eyePHido+dMGEC3/3ud7npppv4+OOPycvL63kBkiRJkpTBMiqcvlPdte1dEQTBdo/z8vJoaGgAYOPGjc37rrzySn71q1+xYcMGJkyYwMqVK3tegCRJkiRlsIwKp3sP79r2rqiurubZZ58F4N5772XixIkUFBSwdOlSAB588MHmY1etWsXo0aO54oorGD9+vOFUkiRJkjqQUeH0h+Wwc/6223bOj2/vqZEjR3LzzTdTWFjIRx99xIwZM7jmmmu47LLLKC4uJjc3t/nYX/ziF4waNYoxY8bQp08fTjrppJ4XIEmSJAnqZkJdHtQFLb4KoD5BM6EqMkEYhlHX0Ky4uDisqKjYZtuKFSsoLCzs9DkeisXHmL5THW8x/WE5nFmS6EozU1ffa0mSJClp6mMQXgKsb7ExD6jbwZMGQzAbcg0AqSoIgqVhGBa3tS/jZuo5s8QwKkmSJKW1upnALW3t6OCJayGcCnUXQTDPkJpmMqpbryRJkqQ001Y33TaDaVdsbgypMxNRoXqJ4VSSJElS76iPxceHbhdE65N0wVtaXGuI41JTXMZ165UkSZKUYupjEF4KrIuwiLUQXhTPwXb3TUm2nEqSJElKnvoYhBcQbTBtsjleS13QoitxgS2qKcJwKkmSJCmxWnbfDaeSvG673VHf6rYKwlIDagownCbZk08+yZIlSyK59jHHHEPrpXkkSZKkpKibuXV8ZzgVqIq6oi6ohbAs6iKynuE0yboTTuvqOpoiW5ISbwExzqKASeRwFgUswE+QlZ1WEeM+CphHDvdRwCr/L0gdq9uHns+wG7UqqMvZvptvcytwG/uUUBkXTpfF4IYCuConfrssAf92KisrOfDAA5k+fToHHHAAJSUlLFy4kAkTJrD//vvz/PPP8+GHH/KlL32JMWPGcMQRR/A///M/VFZWcuutt/Lzn/+csWPHsmjRIiorK/nCF77AmDFjOPbYY6murgZg+vTpXHrppRx++OH84Ac/4I033uC4446jqKiIcePGsWrVKs4//3wefvjh5rpKSkp45JFHqK+v5/vf/z6jRo1izJgx/PKXv9zuNSxYsIAjjzyScePGcc4557BuXSr0+ZeUChYQ44sM4Vqm8j5VhIS8TxXXMpWJBEwk4GSGGFaVUVYRYz5DmEfAPALuZgCxxsdPMY31VAEh66niKaY2H9f0dSd5LMElKqR4cOsHvBN1JQkSEu/mOw3qjovP8NvcCty0r40uwPWx+LHODNwjGTVb77IYPFIKW2rjjz+pij8GGNvDCbneeOMN7r//fubNm8f48eO59957Wbx4MX/4wx/4v//3/zJs2DAOOeQQHn74YZ544gnOP/98li1bxqWXXsqAAQP4/ve/D8Cpp57KBRdcwAUXXMC8efP41re+1Rw433rrLZYsWUJubi6HH344V155JWeccQYbN26koaGBr371q/z85z/nS1/6Ep988glLlizh7rvvZu7cuVRWVrJs2TLy8vL48MMPt6l9zZo1/PjHP2bhwoX079+fn/zkJ/zsZz/j6quv7tmbIint/ZSZ/L4Tn3R/ylquZSqPcSe/YGEvVCb1zCpiVFDGeqoIyCWknv6MoJhy3ucZVrb6d1/PeupZ3/go7PD8IfWs5BZWckvzeffD2T+VZepj8dlv2Rx1JUkQAo+3s6+2cQ3VC4iPW80BGlodsxbC6c4M3EUZFU7/VrY1mDbZUhvf3tNwuu+++zJ69GgADj74YI499liCIGD06NFUVlZSVVXFgw8+CMAXvvAF1q5dy6effrrdeZ599lkeeughAKZNm8YPfvCD5n3nnHMOubm51NTU8Pbbb3PGGWcA0K9fPwCOPvpoZs6cyerVq3nwwQc566yzyMvLY+HChVx66aXk5cW/nbvvvvs213zuued49dVXmTBhAgCbN2/myCOP7NkbIiltLCDGL7iMT1nb43NV8Difpy91bGEPhnMJ5UzxD3L1opahs7WdGEwAbGr1bz1snPSkqRU00ZrO+xRTDarKDvWxxvGZ6TSmNBmaJlRqHUyb1EF4vgG1CzIqnH5S3bXtXdG3b9/m+zk5Oc2Pc3JyqKuro0+fPj2+Rv/+/Ts85vzzz2f+/Pn89re/5c477+zUecMw5Pjjj+c3v/lNT0uUlEZ+ykwe5jbCdn9pds+Wxk/Im7r+2pqq3rKKGIsppZ7aNvdvTsAHMD3VFFQXcwkTuc2QqsyT0a2lydDQ2Mp6GQSzDakdyKgxp7sM79r2RJo0aRKxWLxf+ZNPPsmQIUMYNGgQAwcOpKampvm4o446it/+9rcAxGIxJk2atN25Bg4cyGc/+9nm7r6bNm2itjb+i3j69On84he/AOCggw4C4Pjjj+e2225rnkipdbfeI444gmeeeYY33ngDgPXr1/PPf/4zQa9cUipq6q6b6GDalgoeZzJ5jklVUi1hJk8xtd1gmmrqWc9TXOBkSsocdTOhLrdx/KXBtOsau/nWDXFipR3IqHB6fDn0yd92W5/8+PZkmzVrFkuXLmXMmDFceeWV3H333UB8jOnvf//75gmRfvnLX3LnnXcyZswY7rnnHmbPnt3m+e655x5uuukmxowZw1FHHcV7770HwJ577klhYSEXXnhh87EXX3wxw4cPZ8yYMRQVFXHvvfduc66hQ4dy1113cd555zFmzBiOPPJIVq5cmaR3QlIqeIS5vXq9Buq5lqkGVCXFEmZuN0Y0PdTzLJdFXYTUPdvMUDuA+Ey8yf/AM7PVAWvZ4cRKWS4Iw44H/feW4uLisPW6nCtWrKCwsLDT51gWi48x/aQ63mJ6fHnPx5umktraWkaPHs2LL77ILrvsktBzd/W9lpS6JhJEct0ccnkal8NSYuxofGl6CejPcMeiKn3Ux+LBKU16KqS3EZBXGXURvSoIgqVhGBa3tS+jxpxCPIhmUhhtaeHChXz1q1/lO9/5TsKDqSQlQgP1LCDmJEnqkT9zHO+2O0tmOoovSbOY+BICBlSlvLAMg2lvScDkOBkko7r1ZrrjjjuOqqoqvv3tb0ddiiS1q2l9VNdGVXc8yMEZFky3qqeWCsqiLkPqBANT7+mFyXHSiOFUkpQUTWuj/pSZUZeiNPFnjuMTXo26jKRa7x/9Sgu7d3yIEiPohclx0ojhVJKUVL/nFr5oK6o6sISZGdtiuq3QGXyV2upjwEdRV6EsZTiVpAzUh52iLmEbn7CWa5nGRALOosCgqu281sszTEfpKaaxxB4FSlXhZTgrby8K7erfkuFUkjLQD5kXdQltiM8O/z5VjkvNQquIcR8FzCOH+QwhxhDmkcN9FPBnjiOkPuoSe1HISm61BVWppz5GfKkT9R67+rdkOO2EyspKRo0atd32iy++mFdfTdzYmAEDBiTsXJKy2xRKOIMZUZfRoU9ZSzkXGlAz3CpiLKa0cUmYkM2sZVPjWn/rqcqS7rythTzFVFtQlVpsxYtAGF9P1vVOAcNpj/zqV7/ioIMOiroMSWrT95jDLgyOuowO1bOF25zBNKNVUEa9y1K0aSW38GeOi7oMqZGteNGognAq1AWNX0OyNqxmXDitiUF1AfwrJ35bk6Dva11dHSUlJRQWFnL22WdTW1vLMcccQ0VFBQB/+ctfGDduHEVFRRx77LE0NDSw//77s3r1agAaGhr4t3/7N1avXs3777/PGWecQVFREUVFRSxZsmS7691www2MHz+eMWPGcM011yTmRUjKOpcxO+oSOuV9qphIwPEMZBI5jkvNMPEWU7XnXR63i69ShMuapIa1LcJqQVYF1YwKpzUxWFMKdfFeQ9RVxR8nIqC+9tprzJw5kxUrVjBo0CDmzJnTvG/16tV87Wtf48EHH2T58uXcf//95OTkMHXqVGKx+MUXLlxIUVERQ4cO5Vvf+hZHH300y5cv58UXX+Tggw/e5loLFizg9ddf5/nnn2fZsmUsXbqUp59+uucvQlLWmUIJg9Kg9bTJBtYREjaPS3VMamYIyI26hJTn+qdKCUE5kB91FdpGFYSlWRNQMyqcflQGYateQ2FtfHtPDRs2jAkTJgAwdepUFi9e3LzvueeeY/Lkyey7774A7L57fG2oiy66iF//+tcAzJs3jwsvvBCAJ554ghkz4mPBcnNz2WWXXba51oIFC1iwYAGHHHII48aNY+XKlbz++us9fxGSstK3mU3fNP1j49PGWX5dKzW9ZddkR91j67JSQm4JBHMhjT7UzA61WTMeOKPCaV073eTb294VQRDs8HFbhg0bxp577skTTzzB888/z0knndSpa4VhyA9/+EOWLVvGsmXLeOONN/jqV7/arbolaQolXMFcoOOfW6kp5PfcwmTybEVNQ/Huqun6b6932bVXKSG3BPLWYEBNNdkxHjijwmleO93k29veFdXV1Tz77LMA3HvvvUycOLF53xFHHMHTTz/N//7v/wLw4YcfNu+7+OKLmTp1Kueccw65ufFuTcceeyy33HILAPX19XzyySfbXOuEE05g3rx5rFu3DoC3336bDz74oOcvQlLWmkIJV3NP1GX0SAP1jUvQ5NiSmiaaZultWkZIO/YUF0RdgrRVMBu7+KaS7BgPnFHhdLdyCFr9Hwry49t7auTIkdx8880UFhby0UcfNXfLBRg6dChz587lzDPPpKioiHPPPbd532mnnca6deuau/QCzJ49m7///e+MHj2aQw89dLvlaKZMmcK///u/c+SRRzJ69GjOPvtsampqev4iJGW1KZREXUKCxFtSnTwpta0ixtNc4Cy9XVLv0jJKHc1dfEdEXYkA6B91Ab0iCMPU+TSzuLg4bJr9tsmKFSsoLCzs9DlqYvExpnXV8RbT3cphYIR/j1VUVPCd73yHRYsWRVdEJ3X1vZaUfiaTR0OGj//bkxFcQnkGhfH009RiajDtnqOZz37++1WqqTsOsnJN4lQyON6inZvePx+CIFgahmFxW/syquUU4kF0eCV8riF+G2Uwvf766znrrLP4z//8z+iKkKQWTqc06hKS7n2q+AmltqZGyHVNe8aZe5WS8hZCMD/qKrLcWgjPz+iZe/OiLiCTXXnllVx55ZVRlyFJzb5HfBms33NLxJUk1yZq+XHj+D1bUHvf+iyZuCNZnLlXKSu3BOqJr8GpXrUqgIocWE8D/ZlKMWRkD4uEtJwGQTAvCIIPgiB4ucW23YMg+FsQBK833u6WiGtJknrme8zhDGZ0fGCaa6Ce/+QiFhBjATHOosAxqr2kf5ZM3CFlpdwSyILfIalkVQCLc2B9AATx28WUZuQM34lqOb0L+H/Ar1tsuxJ4PAzD64MguLLx8RXdOXkYhp1aukXdl0pjjyUl1wJiPMbdUZfRK7awmWvZ9hP+pm6/YKtqIq0ixiIupoGNUZciKdny5kD9BAgvA9ZGXU3G2dpKGp8GaQtQ3yoK1VNLRXgZ+wWZ9XssIS2nYRg+DXzYavPp0PzXz93Al7pz7n79+rF27VrDUxKFYcjatWvp169f1KVI6gW3UcamLB8PuIlaZnNZ1GVkjCXM5CmmGkylbNK0HmowH9dETZy2Wkk3t3PsetZCXWbN8J3MMad7hmH4buP994A9u3OSz372s7z11lusXr06cZVpO/369eOzn/1s1GVI6gUfOB4QgE9YywJitp720BJmsjLDxzD3thz8sFhpJLcEKIlP0hOeDzREXVFaq8jZvpWUdjqQxheXuTXeip3mM/g26ZUJkcIwDIMgaLPpMwiCUoj3rxo+fPsxKn369GHfffdNboGSlEX2YDjvO+EKEG9FNpx23ypiBtMkmMSvoi5B6rrmyZJKIct75/TE+vZ2hGwbUkMobmi8E5ZBhvwuS+ZSMu8HQbAXQOPtB20dFIbh3DAMi8MwLB46dGgSy5EkAVxCOX3Jj7qMlGBI75nn7BqdFJk4A6eyRG4JBHOxm2/39e/Cse83h9XM6RGVzHD6B2icxz9++0gSryVJ6qQplHAFcxnkHw8AztzbA5ucCCXhdvL/pdJd01hUjo26krRU3AC5rfubtm41Jf74teZtmfOBc6KWkvkN8CwwMgiCt4Ig+CpwPXB8EASvA8c1PpYkpYAplPAYayj2jwduoyzqEtJSJi5hkApcm0AZI28hLjnTdfuFMLEB+odA2HjbjpD4BEqwPj7mNwMEqTQLbnFxcVhRURF1GZKUVSb65zCLSZ3fhekgvmzMRTS0O4ekui/gIieUUaapmwmOT++2O3MhbOdXdW5jmN0vHAF5lb1aV3cFQbA0DMPitvYls1uvJCnF2aU1zveha57jMoNpkvRn+8khpbSXNyfqCtLayMZW1LbUN66JminjTg2nkpTF7NIa9wsn9umUVcSIMcSxpkkTUEx51EVISZIbdQFp66gGOHAHATU+w29mfLBlOJWkLOaap3GfGrY6tIoYiyk1mCbRgVzqTL3KYKVRF5DWjmpofybf/gQQZMYHW4ZTScpie2TIJ62JMIlcJhIwmTx+ysyoy0k5FZRR79qFSXUUdn1UBsubQ3yCJFtQu6utmXxzQygOL43PkpwBDKeSlMUuoZxc+kRdRkoIGyehaaCe33ML3+a4iCtKLetdEzapXEJGWSFvDuTVQV4Y/2IGxpHOa2sm34kNsF9O5nyw5b8GScpiUyihjDujLiMlVfC4LaiN4svGOKtzsgT04UhmR12G1Pvy5kBefTyoBvPBD2k6tF8I59bDRfXx2/3CzFqux3AqSVluCiUM8g+CNtmCGldBGe3OxKEem8ydjjWVcksgb02LFlV17NiMmwnZcCpJ4tvMtntvOyp4nIsJuI+CxhbE7GOX3mTKMZhKreXN6WZraj5wLJnf06N//L3JWxh1IQlnOJUkNXfv3cUW1DatBH5JFRcwlcnkZl1338AJTJKoIeoCpNTVsjU1mA/s1MZBOxEPoyMgmBsPbME9ZG4X4RmQty5jJkBqzXAqSQLiAfVPrOFq5rMnI6IuJ2U10MDvuYXP048FWdKSGlIfdQmSsl1uCQTzgBFsDaPzIW8T5DVAXuXWwNYUajNqHGt+4+vNrG68rRlOJUnbmEIJD1JpQO3AFjbxn1yUFQG1v/8WksZZeqUuyC2Jh9DWYXSHx6/ZOjtwyy9mkJzuvwMaz930c7MnPU/6s02rcIa2lrZkOJUktekSyulLftRlpLQtbOZapvJ9Do66lKQqppzcVv8WAscoJ4Sz9EoRyZvTGHLbCK7N3YibWmkHA3ntnKgpfDa15NY0nruy8Vx17Zx3R3Jo7r7b2SCeIQynkqQ2TaGEK5hrC2onPMerfIE+TCKHsyjIuNbU/ShhInMbW1AD+jOCydxpq18PHcgMJ0OSUtU2rbRrILiL7bsUtwyflZ0LkM3nbScU54WNy+tkdvfd9hhOJUntsotv522mjpCQ96niJ5RmVEBdRYwKylhPNf0ZTjHl7EcJRzJ7uxZVdc6BzOAosvOPTyktdbVLsbrFcCpJ6tAllJP5U/MnziZquY2yqMtIiFXEWMRFjcvJhKynikVcxBJm8iyXUU9t1CWmHYOpJLXNcCpJ6tAUSoAw6jLSygdUR11CQjzHZTSweZttDWxmJbewmbURVZXe3uSxqEuQpJRkOJUkdYpde7smn5D7KGBVmnfv3ZTGAXQVcB8wr/F2VbTlNFufIR9cSFKiGU4lSZ0S79qrzhoGrKeKxZSmfUBNR6uAxcD6xsfrGx+nQkDtz/CoS5CklGQ4lSR1yhRKGOTsrJ22EngQqKeWp7mAeeSkZUtqus7IWwHUt9pW37g9SrnkU+wHPZLUJsOpJKnTvs1s+rBT1GWkjU+APwMh9TRNJpRuLanxdTh7soh8NNZ3cXuvCGEic10+RpLaYTiVJHXaFEr4IfOiLiOtvNvqcT21VFDGAmKcRUHKr426HyXsxK5Rl9Fl/bu4vVcEGEwlaQcMp5KkLplCCbukaVfPqCxp9fh/GtdCfZ+qlF0bdRUxYgxhHkFazspbzPbtvbmN2yVJqclwKknqssvs3tslK9l2Ip4XyWVTq/VBo1obdRUx7qNgmzGxf+Y4nmJqWs/Uux8wka0tpf0bH+8XWUXx9U0lSe3Li7oASVL6mdLYNfFapkZcSfp4jngwyiWfmlbBtMn7VDGPHPoznGLKk94FdBUxFlNKfWM966niKc4HGpJ63d6yH8kNo/0ZQS1vNY4p3rG9OJajmJPEaiQp/dlyKknqlimUuPZpF2wiHmYmMrfd9y3eypeciZPaaiGtoKw5mG6VGcE02Zpm3R1J6Q6P688IjmY+J7GwlyqTpPRlOJUkddsllNOX/KjLSBu/Yg2raPt9az0esmnipERoaiFdTxUtw2/8sbqq6UOG/SjhKOY0dtfd+idVLv05mvlcRMi5VDoJkiR1UhCGYdQ1NCsuLg4rKqJegUyS1BULiHEbZbxv0OmUvuRzBXMBuI0yPqCafEKKaasLasBFCWjJvI8Cg2jCJOZ7IknZKgiCpWEYtjk/nS2nkqQemUIJD1LJ1cwHgqjLSXlNEx81vW+LaOAiRrQ5NrI/wxNyzfVUJ+Q8Stz3RJK0PcOpJCkhplDCGVyKAbVjH7QKi8WUk7tdN9/4mMZEaC9QBdsttqKAPuS0MxN1Ir8nkqTtGU4lSQnzPeZwNfdEXUbK26NVWNyPEiYyl/6MAIJtxjQmQnvhdySlbW4/kBltbg/om5B6UlP8fZ/MnUxiXuP3YmuAT/T3RJK0PZeSkSQllMvMdGwoa1hFbJugsx8lSQs+TeetoIz1VG+zVM2eTOj0dmCbpWcgHlrz2Ln31kQNSXDjfC5Hc/d2770hVJJ6nxMiSZKSYgExfsFlfNpboSWN5AKT6cN07ky7ENS0BE1HoTUpQtgrhHcD2gyoufQnj36dCMoBENKfEb2ynqwkaasdTYhkOJUkJd0CYpRzIfVsibqUlNEfuIgRnEtl1KUkROvQOoyTeY25hNRvd2xALpO5G4CnOmphb/wzZSfgyAbYL4RVAVTkwHqgf9B2wFzCzBbXzyGXnamndpvWYUlS79tROLVbryQp6bZ29Z1Gc9rIcuuJz6LbVktkOgantrol78mENrsBtx67+QyXUse6Fs9s0bIZlLNfPRCWQuN59gthv/p8COZCbtvv1VHM4SjmJOjVSZJ6g+FUktQrmgLqTyhlU7K7f6aJKvK3CW/rqWIxpUBmjHnc0VjXlsd0+FpzgXogLAOqgeEQlLcbTCVJ6cluvZKkXrWAGLdRxgdUswfDOYqT+T23RF1WpPoDTf2bKmjsrgoUE/BFLrUFUJKUMezWK0lKGVMoaW5Fbak3AmoOuTS0MQYyauuBpxvvhy22LSaExvfFgCpJynSucypJitz3eil4pWIwbRKy/WjceuItqa8xt/cLkiSplxlOJUkpoQ87RV1CSloPbc54K0lSpjGcSpJSQh11UZeQkvoTX3pFkqRMZziVJKWEkIaoS0g5ucQnShrZOIOvJEmZzHAqSUoJObYOciDxllIabycS8EVmOBmSJCkrOFuvJCklnE5pVi8pEwBHAbnkM5G5GbHOqSRJXWHLqSQpJXyPOeRl8aRIk4H+jDCYSpKyluFUkpQy/g/zoi4hEnsBYxjBuVQaTCVJWctwKklKGVMo4QxmRF1Gr/sA+B+qoi5DkqRIGU4lSSnle8zhauazMwO227cLg8mlTwRVJVc98KITQkmSspwTIkmSUs4USpjSTvfWBcS4jTI+oJo9GM4w/o0KHu/lChOvhvqoS5AkKVKGU0lSWmkvuC4gxnWcn7brpe7JiKhLkCQpUnbrlSRlhCmU8B/8mr7kR11Kl/Uln0soj7oMSZIiZTiVJGWMKZRwBXPZhcFRl9JpOeRyBXPb7cYsSVK2MJxKkjLKFEr4E2voR/+oS+mUBuq5jTIWEIu6FEmSImU4lSRlpE3URl1Cp71PFeVcaECVJGU1w6kkKSPtwfCoS+iSerbwCy6LugxJkiJjOJUkZaRLKE+7yZE+ZW3UJUiSFBnDqSQpIzVNjpQuY0+b2LVXkpStDKeSpIw1hRIWso5ijo26lE6za68kKVsZTiVJGe8XLORq5qfFEjN27ZUkZSvDqSQpKzQtMXM18xmUBiFVkqRsYziVJGWVKZTwWGNI3ZMRUZeznYAg6hIkSYqE4VSSlJWmUMKDVHIGM6IuZRshYdQlSJIUCcOpJCmrfY85nMEMghT6leiMvZKkbJQ6v4klSYrI95jDIupTZtKk2yiLugRJknqd4VSSpEYtJ03akxEEBOSQ2+axyWxp/YDqpJ1bkqRUZTiVJKmVpvGoi2jgKu6mL/nb7O9LPl/ikqRdfw+GJ+3ckiSlKsOpJEk7MIUSrmBuc0vqnozgCubyPeaQx04Jv15f8rmE8oSfV5KkVJcXdQGSJKW6KZQwhZLtttezJSHn70d/NlHLHgznEsrbvJYkSZnOcCpJUjftwXDep6pH57ia+YZRSZKwW68kSd12CeXk9vBzXoOpJElxhlNJkrppCiWUcRcQRF2KJElpz3AqSVIPTKGExTQwgF27/NxUWFNVkqRUYTiVJCkB/sJHFHNsl55zGbOTVI0kSenHcCpJUoL8goUsJmz+OoMZ7R57BjMcbypJUguGU0mSkuR7zOFq5jOoRffdXRjM1czne8yJsDJJklKPS8lIkpRE7a2RKkmStmXLqSRJkiQpcoZTSZIkSVLkDKeSJEmSpMgZTiVJkiRJkTOcSpIkSZIiZziVJEmSJEXOcCpJkiRJipzhVJIkSZIUOcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5xKkiRJkiJnOJUkSZIkRc5wKkmSJEmKnOFUkiRJkhQ5w6kkSZIkKXKGU0mSJElS5AynkiRJkqTIGU4lSZIkSZEznEqSJEmSImc4lSRJkiRFznAqSZIkSYqc4VSSJEmSFDnDqSRJkiQpcoZTSZIkSVLkDKeSJEmSpMgZTiVJkiRJkTOcSpIkSZIiZziVJEmSJEXOcCpJkiRJipzhVJIkSZIUOcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5xKkiRJkiJnOJUkSZIkRc5wKkmSJEmKnOFUkiRJkhQ5w6kkSZIkKXKGU0mSJElS5AynkiRJkjJKjEfpyygCRjZ/Hcf0qMtSB5IeToMgqAyC4KUgCJYFQVCR7Oupc2piUF0A/8qJ39bEOvec1/NgVRD/eiOAe0ZA7M1kVytJkiR1ToxHmcr32cyWbbY/zrMEjGQmsyjg8+RwIAV8nhiPRlSpWgvCMEzuBYKgEigOw3BNR8cWFxeHFRXm12SricGaUghrt25rAAJg0z4w7CcwsGT757w/dftPM0Jg0e7QZxmUDEtq2ZIkSdJ2YjxKGT+jinfIIaCB7uebXHIp5cvMYVbiCtQ2giBYGoZhcVv77NabIW6fCd/JgzsCWNbYsvlcHjwxc/tj1162bTCF+D+EAOj3Nrw1Hf5y07b7qy9t+x9LAEz8EF4YBcs60foqSZIkJUqMR7mQK6niHYAeBVOAeuq5hd/YBTgivRFOQ2BBEARLgyAobb0zCILSIAgqgiCoWL16dS+Uk76WxeCGArgqJ37bFAZ/fhzk3AJfq4djgEHEQ+Me9fCZW7YNqDUxqF+74+v0rYNdr9raXfehGPRb1/7xATDgU3ik1IAqSZKk3nMZP2YLdQk/7+M8a3ffCPRGOJ0YhuE44CTg60EQTG65MwzDuWEYFodhWDx06NBeKCc9LYvFw98nVUAYv71/KpRPhEGPw0SgH/Gg2FI/oP/crY/XXrb9MW0ZWgMvjoCJAVx2AewgmzZ/PrWlFv5W1vnXBPHgO74A9smJ3z5kuJUkSVInreXjpJ17Kt83oPaypIfTMAzfbrz9APg9cFiyr5mJ/lYWD39HANOBC4GLgLOegSOBPjt47uD6rfc7ajVt6Wsh3AY8WQ95xMelthYCK1s8/qRq21bd1prC6G8aux4XTYXfVsGNIbxdBZddaECVJElSx3ojOE7l+wSMZAiHG1R7QVInRAqCoD+QE4ZhTeP9vwHXhmH4l7aOd0Kk9l2VA1NC2IftWz7DNra19B5wVBjvpnvE8M61nLalqcNEbovrrgSeAz4HFAP9W9SzDvgHUA3k7QNPHAL//Bv8dhMMa1VzCLwMzADyd4M3PuxmkZIkScoKBXy+eaxpb+hDHndyPSWc2mvXzEQ7mhApL8nX3hP4fRAETde6t71gqh07MB/2Wd92sNxR2NwC3AqMjsHB03tWQ1Pr6VPAv4B9gfFAYas6mm4HAkcDzwD/ehuOfBuOZ/tg2vScUcRD7r8+6lmdkiRJynzVvNur19tCHRdwBYABNUmSGk7DMPwXUJTMa2SLog07DqGtW0+b2sPzgGuAD6bGJ0rqqRzigfPoxscdtcL2aTx2MvFW1raCaUvFwKrG+0/MjI+XHVwPa3NhfSl8YU73a5ckSVLmGM5evdpyCvHZfEu5CjCgJoNLyaSJ/LYGfLYQEG/VDIHNLba1/EqUrp4zIP4PrRAY0MGx/YFgIPxqH9jnFhhaH3/u0HoYcQs8enC3y5YkSVIGObm5uaR31bKRqXyfAj7vONQEM5ymsJoYVBfEJw7qzMjgAPgr8dbKRIbRRNlRoA2Bt4mPUz0hH8a/s/0kTwFw0Kttr90qSZKk7PI7/hzp9at4h1KuMqAmkOE0BT0Ug7v6xrvi1lVtbXnsKKCuAyaQmsG0Mz4C3tof9nq//RbWgG2XxpEkSVL2ifFoUpeR6axaNnIBV5DDgbakJoDhNGItW0dfyoN5AeROhUmb2540qD0h8Codd5tNVU0TIo1+p+NwPbg+PvOwJEmSsk+MR5vHfaaCeuoJCW1JTQDDaYRqYvDe17a2jvavj08eNIqutX6GwIvEx3Sma6spxGv/UjszEre2cTj84PBkVyRJkqRUU8bPqGVj1GW0qZaNnM/lBtRuSvZSMmrhrsPhqOfj64TWA5uITwDUUlfDZUh8qZYrgUU9LzFhQrbOINyV1zQQ+HIHx+QAxwBHPA93fwY+6AefVMMuw+H4chhb0q2SJUmSlAZ6ewmZrmog5Hx+ADijb1fZctpL7jocJj0f/zQgIH6bn4DzBsChwHFATQLOlyit1zztigGdeF4A7Awc8T4MrgLC+G3OtHgX6eqCeMu0JEmSMsvu7BJ1CR1qoIFptqB2mS2nvWTC810bQ9oVOwOXAlV0vUtwMnWnFbit57S3Hbauo3oE8X/MeY2zRtVVwVsXwGeBgbakSpIkqZeFhK6J2kW2nPaSZL/Re5BawRQ6X0tTF+CeXKcf23/S0rceVk7vwYklSZKUclJhlt7OqmUjZfws6jLShuG0l/QkfHX2/KkUTLtqR2NTe/K6dq+D210XVZIkKWMEafZXbxXvRF1C2jCc9pKXc5IXUEPS+xuZzB8vW4BXb0viBSRJktRrYjxKmPRmn8SbyayoS0gL6ZxpUlpNDFYOgTcCWBLApw2JD6fdnRE3m/QBRjVEXYUkSZISIV27yN7Cbyjg806Q1AHDaRLUxOD9i2CntfE3+DPABBL/ZhtKO9Y0m7EkSZLSX6ovI7MjVbzDNL5vK+oOGE6T4KMyyNm87TZDZHQGRl2AJEmSEmI4e0VdQo+ExFtRDahtM5wmQV111BWopY0Doq5AkiRJiVDOd8mnX9Rl9Ngt/MYuvm0wnCZB3vCoK8geHY3j3QL8d/qNmZckSVIbSjiVufyYXHKjLqXHLuE/oi4h5RhOk+CNk2Fj1EVkifa6S4fAOuAZYOX63qtHkiRJyVXCqdzNT9K+BXU9G2w9bcVwmmDfuwmmzoM/AXVsnVFXvWsT8DvgX1EXIkmSpITLlBbU87mcHA50Jt9GhtMEir0Jz14FN2+CM4E8nFE3Ck0fBnyu8TZ/cMfPOfc4OD+AFwNY1eJrWd/47MuSJElKLSWcSilfjrqMHmkgJCSkinco5aqsD6iG0wT65U0wvQbGYCCNUgD0I758z77Aug1wVQ7cUADL2gia5x4HfR+HMmBXtn6gEACDNsM75xtQJUmSUk2MR/kV90ddRsLUsjFt13FNFMNpAhXfDUfQ82BqN+DE6AOMB3JqgRA+qYL7p28fUBc9Dt9oPL7N8zTElweSJElS6ijjZ2yhLuoyEiqd13FNBMNpAu2yBvr38BwG08Rq/f0I6uD+S7c/rqOevy4PJEmSlFqqeCfqEhIu3ddx7SnDaQKtGwI9nRjWMaqJdyHwZbaOQWXd9se0sWkbLg8kSZKUWtJ9MqS2vMV7WT3u1HCaQCOvgudybf1MJU1hfwBwNPGgegTbdu2ddGx8yZn6ds6xJQd2K09unZIkSeqa+nb/ektf9TQwle8TMDIrZ/A1nCbQUYPhzYHwKgbUVNQUVA8iPslRU0C9byH8cU/4O/H1acMWXzU5sPevYWBJNDVLkiSpbSPYO+oSkiobZ/A1nPbQQzE4dQjcGkAwFaZ/DCOAFRhQU1UAHNAAf2sxydFj78Eb02BuDswDfpkDt02DonqDqSRJUioq57tRl5B02TaDb17UBaSzh2Lw/y6Er2yBz0Nzr/cBQGGEdaljAfBJq0mO/uvXwK+jqEaSpOSIsYkyNlBNA8PJoZydKaFv1GVJCVHCqdzJgzzOs1GXklSZOPFTe2w57YInZsI/8uCNIH77t0tgzBaYCNsNx3ZSo9Q3NYx/L1cOcR1TSVL6i7GJAj4mhw8p4GOO4xOmsZ4qGgiBKhq4iPUM4cPmY2JsirpsqUcWchczOC/qMpIuj8Ks6N4bhGHqdD4tLi4OKyoqoi6jTX/ZDfb/eNvQGRIfX3oQhtF017AT7DnPLrySpNTRstVzdwAC1iZh0NBgAmaTb4uq0tpMZnELv4m6jKSbz42UcGrUZfRIEARLwzAsbmufLaed8OjB2wdT2Dq5TjpJnY8iUkvOZviorOPjJElKthibGMCHTG3R6rkWkhJMaTzvVNYT8CGBLapKU4/xVNQl9IpL+I+oS0gqw2knjHy1/ZbRdGsxTbd6e1NddcfHSJKUTDNZx1TW93jd9J6oooELWU+MTdt1FTa0KlVV827UJfSK9WzgOKZHXUbSGE47IfOW91Vb8oZHXYEkKZvF2MStbI66DAC2ABexntJWY1ZbtrAGfMgAPjSwKiUMZ6+oS+g1j/Nsxo4/NZx2QkfL+9pVNv1tAB6bHHUVkqRsVsaGlPqbYjNQ28Ex66E5sNqyqiiV813ysqhJaSrfz8iAajjthEeGtx9A64GVxD9hbCnE0JoK2vsehEADW79PucAb8+F7N/VWZZIkxR3HJwR8SBUNUZfSI1U0UNrYHVjqbSWcyl38hLwsWinzIn6YcQHVcNoJey+G+4dtG2ZC4q1tTwNP58DTuVDH9qG0dTjq6LES66O+8O7QbcPohsb7OcTH4AbATsCMEM6/zGVlJEm95zg+4fEO+2ilj1riLcBSFEo4lX3YI+oyes1mtnAZP466jITKno8WeqBkGMSegeNWQPUGGL4zlBfGtx/ceMyiIN761taEQ2GL29afBjhBUfKEwIBN8Id+cOtQyFsNxwH/QdvjiANgIPDe9Phjl5WRJCVbJgXTJtVp3gKs9JYtEyM1WcvHUZeQULacdlLJMKicAg2nx29Lhm3dVxODvdnxjL7hDvYrOZpaRC94E+5qDKY/pOMJrnLr4P3Lkl6eJEkZy669iko2TYzUpIDPZ0z3XsNpAnxU1nHw9I2OTgCMBK6CTi8vHqxNXj2SJGWyEBx7qsiU892saxCq4h1KuSojAqqZKQE6uz5mtv1HSSUBXVsSKMr15SRJSneOPVVUSjiVSzkv6jJ6XS0bKeNnUZfRY4bTbnooBj/Mh+UBhM5qlBY6++FACLwyOJmVSJKU+apoII8Pmcm6qEtRlpnDLHaiT9Rl9Loq3om6hB4znHbDD2fCs1Phwg3xCXRsEc0cIfAWUDg76kokSZkuxqaM/xuiHriFzQZU9bqvcnbUJfS6ANK+a6/htIu+dxPcfQtcgFMdZ4qm5X8agFeBX/WHsc7UK0lKsjI2ZM2ScnPZHHUJyjKP8VTUJfS6ENK+a6/5qgtib0Lsx/FEv3PUxajL2psxOQBqgDuBCuB/a3uzKklSNomxiTI2UE1D1gRTIAMXzFGqy7YlZZqk++u25bQLylZAsCbqKtSRppbQ1oJ2tgMMAH4H/AvYe3iSCpMkZbUYmyhlPVVZFkylKGTjkjKQ/q/bcNoF1Rugfkj8fndGTviLqHd0Z/xO0/dz53z4YXkiq5EkKa6MDdg5R+od5XyXfPpFXUavK+e7UZfQI4bTLhi+M9RMhYa+8FNgSwfHt27By/RJD3rbjsJ+e+91yPbfty3As8A+I+CGuXCm400lSUlQTUPUJUQmAHL4kAI+dv1T9YoSTmUuP2YEe0ddSq8JMiBtGE67oLwQcr4An8yEvwyFcuItbu2FpAADabK0N360o+dAfKB1Q+PjdcDiAI6YDy9UGkwlSckzPIv/7Gr6wL6KBkpZb0BVryjhVCr5O/O5MStaUUNCvsoPoy6jR7L3p2Q3lAyDuUWw54mw5nZ47E/dC0nque685/FPbbfe1gMv5sOh9zg7ryQp+crZmfyoi0gBtcAFBlT1omxqRd3ElrReTsZw2kUlw6ByCjScDgPy4hPpKD3lAeM3G0wlSb2jhL7MpT8j/POLerAFVb2qZStqpkvn5WT86dgD1Ru6NzGSUke/uqgrkCRlixibuKxxtl7FW1Avc4oo9bISTqUvfaIuI6mqeCfqErrNcNoDw3eGR+jcLLzO1CtJUvaayTqmsp61UReSYtYSMoSPbEFVr7qD/4y6BLXDcNoD5YUwZyA8Q8fhc1MnjultqVZPFDZGXYAkKePF2MStbI66jJS1ltAuvupVJZzKDM7L6HljZjIr6hK6xXDaAyXDoOTHcGUu/Ah4j3jga91ZZwvwHLCC1AqEmfwfsqUQqGP7974emJ/b+/VIkrJLGRtS6vd/Kqol/j5JvWUOs7iHGzN2kqRb+E3UJXSL4bSHfvot+H93w4oRcE4AXxkB1TMgbwQQwLrceMsqwL9FWWiWaOuXf9C4fQVbl/5ZBzwOjL+792qTJGWnbF7ftCuqaLD1VL2qaZKkwewadSlJkY6z9hpOE+DMkvgamW83wE9Oht1vgy1VEIawUx4MASYAO5E9rZVRae/97QMMB34H3AncC+TOcF1TSVLyZfP6pl1l915FYTZX0Ye8qMtIuHSctdeflgn0xEwYdgsMaoiHpADYaRMcDEmfE8zuQh3rT/x96jMYvjIfvjYn6ookSdnA9U07z+69ikIJp3In12dcC2o170ZdQpcZThNo6C3Q1hDGZLeWhr1wjUywDvh1LlyzxrVNJUm9p2l9U3WO3aAVhRJOZQ3/YAbnRV1Kwgxnr6hL6DLDaYIsixHZp6IG046FxMf+1tdHXYkkKRuV0DfqEtKG3aAVpTnMYj43pn0rag45lPPdqMvoMv/3J8jfyqKuIHMkuotyCLxC4zhTZ+eVJCmllbNz1CUoyzW1ooa8RshrUZfTLQ1p2gPBcJogn1Tj8P0ECIG3iS+/k4hzbSA+K+9/Af8CppYm4MSSJElSinNCpCy2y/D4Wqate406UVHXBMA+QB49e+9C4C/AJcAsoCoXLpgB/+kkSJIkpbSLWB91CVJGSMcJkTJvzuSIHF8Oj5TColooJj4z7DqgCtgfHGnSBS3H0HZnsqcQ+MvucMNvoHJK4uqSJEnJtznqAqQMkY4TIhlOE6Rp9te/lcHvqqHP7vDUp7ByCxwHXArsybZhy4mMOhbQ+YAaEu9a/SxwXS78tDCZlUmSpGQZwkfMJt+JpJQSBpDPOmqjLqNLAoK0nBDJcJpAY0u2XaLkoRj8Zxk8Xg0Vu8cD1kdr4/u+A5yJATVRQmAF8a7V64CcNVAyLNqaJElS96wlpLSxe68BVVG7lWuZyvejLqNLLuUrlHBq1GV0WRCGqTMqsri4OKyoqIi6jKSqicGb06BvaDDtrM60nIbAncQnUnoG2DQCXqhMcmGSJHVBLh+m6fyZ0RlBDpVpvqSHMsMQDmctH0ddRqf0Z2fWsSzqMtoVBMHSMAyL29rnhEi9qCYG718A/RIQTFPnI4XUEAJ7Ew+m7+bDD8sjLkiSpFYuYaeoS0g71cZ5pYjZXBV1CZ3WL417GxhOe9FHZZDTejpf7VBnQ3gO8bG9o/vDDXPhzJKOniFJUu+aw4CoS0g7w/1TVSmihFMZnCat+B/ySdQldJv/43tRXdWO94d0Poxt6sKx3ZEqLbMBnW9lzgO+tx4+vDOJBUmS1AMj/NOrS8rZOeoSpGazuSothuWl4yy9TfwJ2ZtyOz6kphOnCYGd2H5N1URKh/94bckBjnwcnpgZdSWSJG3PsNVZIceS62RISiklnMqlnBd1GR36N0ZEXUK3GU57UwdpsiaAX5zS9mEtW1UD4t+4XHAkRht2BvrPjboKSZK2V0JfBkddREoLgQZKqGIhu0RdjLSdOcxiPjdGXcYOPcnzUZfQbYbTXpS3gw8xQuB/p8Lxc+A/xsCnbA2ktbQ9Y21Tl9fNpE433FQx2LG9kqQUNZv+9Im6iBQ1hHd5g78yn3FRlyK1q4RTmc+N9EnRVTnrk9q/MrkMp71ot/L2Q+Qm4Ixfx9fmPPWPMOX7cORQmBjA8UPb72YbEG89fYp4i2tvhNSujI1t7/nJrmFtJ7pQS5IUhRL6MihtB9Ak1y/Yn/1wVkOlvhJO5U6uT8lJknI7M5YwRRlOe9HAEth87Pahqg5omLH1cckw+OAG+OADeLcB/nb2js/bFygm/s3saXDsjC09fH5Az+oMgVX7w5PAxjbOswFYX9rt8iRJSroP7fPUBt8TpZcSTmUN/yDkteavVOjyW8qXoy6h2wynvaxwIWyYAetz4z+C1+fC5hkwak7bx788E3a6peMJigawdSxqMtUBSxJwnqaA2h2DZsCUf8Knx8J84iG1pvF87wHPHgtfaOf9lCQpFbhESlsCLqM26iKkHinhVGZwnn0jusmfjBEYNQdG18F+Yfy2vWAKEM6lw97sbY1FbUsD3Z9AKQTWAYuBfxHvhtxTATtuhd0uvAYwcAYMbXy/vrMQiufDrSPg5AC+MgLemw8XL0xAcZIkJZGz9rZtra2nygBzmMU93MgI9iYgYDC79ur41Fv4DTEe7bXrJVIQhqnzQ6C4uDisqKiIuoyUsipIzLIuLb/L3TnfJ8ADwH7AEdA8sXtPalsHVLQ4X8tz1QH1x0L/N6CuGvKGx8fsDnQYiiQpQ/Tjw4R82JtpQnaPugQp4WI8Shk/o4p3euV6I9ibSv7eK9fqqiAIloZhWNzmPsNpanspD/pHPOFW2Oq2dXN7WzMJd2QL8AzxVliAYXlweAMMbIDaXAhKd9yiLElSuouxiamsj7qMlGM4VaYbwuGs5eOkXiMgoIGVSb1Gd+0onNqtN8UFpV0fm5nojxualqzJoe1/MF0ZP9rUPbhlMO2TD4ffBWPrO9fVWZKkTFBCX2awU9RlpJTBjtRTFpjNVeTTL6nXGM5eST1/shhOU9yoOfDWQW2Hv9bbQuKz174K1PaLjy9NnXbxeC1PAfcBK/OAAHYZAafPhbF215UkZaE5DGA+/RlBDgGk8QIQCRCGzCY/6iqkpCvhVOby4+YxqYle+iWHgHK+m9Bz9hbDaRo4+hV45dj4EilNS7BsBFYA6/vHH9cQn7V2XgDvzYBRG+DoXvzwcRMdT260Alg7As6ZD/+1BX7cAJdXGkwlSdmthL5UsisN7E4duzODnbKv/TAMmVFdRUnzrBZSZivhVCr5Ow2s5G5+krCW1J3ow6+5gRJOTcj5eptjTtPIshj8rQw+qYZdhsPx5TsOduML4LdViZlQaUfqgKeB5WPgy+/CXqu3zubbj3g33tcOgi+/kuRCJEnKIDE2UcYGqmkgoPsz7qe8MOTYNWtYOHRk1JVIkYnxKJdyNet6sJzSfG5Mi1DqhEhZ6qEYjJnadvP4jiYxamDrONO2nlfP1m5HG4m32P7qRFh2aXxbfi7MLYKSYd2vXZIkbZXDhyk1VCfRnARJigvo3oc0x3IkC7krscUkiRMiZakzS+DZ/dsem7qFtsejBoNh44z2u+huAX4N3AYsnw8Hh7BrNXz0nXiYHbGzwVSSpEQb7p9sUlYYzK5dfs5B/FvaBNOO+JMuw037J7w0Id71NiR++7cAbhkUn5xofW58e94IGDof9l0Tn4TpnRnxVtGWtgBLiLes7j8jHn4hHkQrp0DD6fFbg6kkSYlVzs7bTRWUD8ynf3ZPoiRlmNlcxU702W57wNbg2jSB0gj2Zj438gp/6sUKkysv6gKUfF9avO3jAzrxnC/MgSeA/nNhSH18wqUXiU9o9OUOxrpKkqTEapooqGkM6nByKGfn5u3TWJ/R3X6lbNE0ZrSMn1HNuwxnL8r5blqMJU0Ex5xKkiSluZms4xY2R11Gt4wgh8pudGWUlJ4ccypJkpTB5jCA/lEX0Q0B8S7LkgSGU0mSpIzQ/QUoohEAl7KTa5tKamY4lSRJygCpNKPvTsQnawrZnfn0Z0RjbU2TN40gh3vozxwGRFajpNSTOj/FJEmS1G1tzejbJACOJZfBvVDHCHKYR//mFtES+lLJroTsTh27E7I7lexqi6mk7RhOJUmSMkAJfZm7g1bKhezCmsZw2LpFM+jhtXPZ2lJq8JTUXYZTSZKkDNGVVsqWx97TGFQD4mF2Bju12wrbWj5wd4uWUknqLtc5lSRJynIl9N0uXE5gU/O6qrsTX/O89WI1g4HZBlNJCWLLqSRJkrbT1LLawO6sYXfmtWpdnU9/1rC7wVRSwthyKkmSpA611boqSYlky6kkSZIkKXKGU0mSJElS5AynkiRJkqTIGU4lSZIkSZEznEqSJEmSImc4lSRJkiRFznAqSZIkSYqc4VSSJEmSFDnDqSRJkiQpcoZTSZIkSVLkDKeSJEmSpMgZTiVJkiRJkTOcSpIkSZIiZziVJEmSJEXOcCpJkiRJipzhVJIkSZIUOcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5xKkiRJkiJnOJUkSZIkRS7p4TQIghODIHgtCII3giC4MtnXkyRJkiSln6SG0yAIcoGbgZOAg4DzgiA4KJnXlCRJkiSln2S3nB4GvBGG4b/CMNwM/BY4PcnXlCRJkiSlmWSH032AN1s8fqtxmyRJkiRJzSKfECkIgtIgCCqCIKhYvXp11OVIkiRJkiKQ7HD6NjCsxePPNm5rFobh3DAMi8MwLB46dGiSy5EkSZIkpaJkh9MXgP2DINg3CIKdgK8Af0jyNSVJkiRJaSYvmScPw7AuCIJvAH8FcoF5YRi+ksxrSpIkSZLST1LDKUAYho8BjyX7OpIkSZKk9BX5hEiSJEmSJBlOJUmSJEmRM5xKkiRJkiJnOJUkSZIkRc5wKkmSJEmKnOFUkiRJkhQ5w6kkSZIkKXKGU0mSJElS5AynkiRJkqTIGU4lSZIkSZEznEqSJEmSImc4lSRJkiRFznAqSZIkSYpcXtQFSJKk7os9BWX3QPUa2H0AbNoC6zbG9w0eCLO/BiVHR1ujJEmdYTiVJClNxZ6C0puhdlP88dqabfevrYELZ8fvG1AlSanOcCpJUgpq2SI6fAiUT4tvv+z27UPojmyph6k/i3/l5kDpCTDn0q37Z94Kc/8K9Q1t75ckqbcEYRhGXUOz4uLisKKiIuoyJEnqNbGnuh44e2rGSfEAOvNWuOXP2+/fe3d4+87eq0eSlD2CIFgahmFxm/sMp5IkRSP2FFx0E2yu6/1r5+bEW0t3xDGrkqRE21E4dbZeSZKSLPYUFFwMOV+K38aeim8vuyeaYAodB1OIt+ae//Ot9UqSlEy2nEqSlERRto4myuCBsGZ+1FVIkjLBjlpOnRBJkqQEar20S2+OJU2WtTXx12X3XklSMhlOJUlKkNYTDGVCMG3ikjSSpGRzzKkkSQkQe6rtmW8zxZb6+KzCkiQli+FUkqQEKLsn6gqSL5NagiVJqcdwKklSN8WegiFTITgdqlZHXU3vaDnbsCRJiWQ4lSSpi5pC6dSfZV9rYtXq+OseMtWQKklKLMOpJEldEHsKSm/OvlDa2tqaeEgNTrc1VZKUGIZTSZK6oOweqN0UdRWppWp1PLAbUCVJPWE4lSSpC7JlbGlX1W7KjkmhJEnJYziVJKkTYk/BgC9HXUVqq1rdAPWxqMuQJKUpw6kkSR1oGme63u68O5QThBCWGlAlSd1iOJUkqQOOM+2chjCH2NOnQ1gWdSmSpDRkOJUkqQOOM+2sgLLY/wWqoi5EkpSGDKeSJO1A7CkIoi4ijVStGQ4Edu2VJHWZ4VSSpB0ouwfCqItII7k5DUBo115JUpcZTiVJ2gG79HZNfUNu473qSOuQJKUfw6kkSUqYgIbGe8MjrUOSlH4Mp5IktRJ7CgouhuD0qCtJPyE5QD4E5VGXIklKM3lRFyBJUippWtPUpWN6IJgLuSVRVyFJSjO2nEqS1MJltxtMeyInCAymkqRuMZxKktQo9hSsrYm6ivTW4NTGkqRuMpxKktTostujriD9jRgadQWSpHRlOJUkqZGtpj1XPi3qCiRJ6cpwKkmSEqbk6KgrkCSlK8OpJEnEx5uqZ+zSK0nqCcOpJElA2T1RV5De8vvapVeS1DOGU0mSgOo1UVeQnoIg3mI69+t26ZUk9Uxe1AVIkpQKhg+BqtVRV5Fe+veFdb+LugpJUqaw5VSSJOyS2h3rN8HMW6OuQpKUKQynkiQR75I646Soq0g/t/zZyaQkSYlhOJUkqdGcS2HwwKirSD+XzV0LdTlQVwD1sajLkSSlKcOpJEktzP4aBFEXkWbWrtsdCIEqCEsNqJKkbjGcSpLUQsnRcM93IdffkN1UC2FZ1EVIktKQv3olSWql5Gio+z0cOybqStLD4IGt1+GpjqQOSVJ6M5xKktSOhdfBTrlRV5HadsrbxOyLLmu1dXgktUiS0pvhVJKkHZjXOnep2eCBG5n39UspmfybFlvzISiPrCZJUvoynEqStAMlR9u9t7WdcmH+d2HN/H6UHH0cMIL4NFIjIJgLuSURVyhJSkd5URcgSVKqW3gdHPcf8Pj/RF1JNPr3hXW/a2dnbglgGJUk9Zwtp5IkdcLC67KzBbVPLtz29airkCRlA8OpJEmdtPC6eCtiNggCGDEU7rws3rVZkqRks1uvJEldcNvX4cLZsKU+6kqSZ8RQqPxV1FVIkrKN4VSSpC5oakWc9jMIoy0lacqnRV2BJCkb2a1XkqQuKjka7vku5GdgF9/BA+3GK0mKhuFUkqRuKDka5n49HuYyRX5fmP21qKuQJGUrw6kkSd1UcjSsmR9f8zPdJ0oaPDAetm01lSRFxXAqSVIPlRwdXwc0nZea2bA56gokSdnOcCpJUoIsvC7eipqbhr9dazdB2T1RVyFJymZp+OtTkqTUVXI03P3t9JwsqXpN1BVIkrKZ4VSSpARrmixpxNDOP2en3OTVA/ExpUEQr6m9SZyGD0luDZIk7YjhVJKkJCg5Gip/BeEjMOOk9o/LzYl3Bd70UPy2KdA2dQ0ePBB26sGq5IMHxs+7Zj40PByvafbXtm/Zze/r+qaSpGgFYZg6S4gXFxeHFRUVUZchSVLCxZ6C0pvjYzub5Pft3Ay5safgstthbU3nrxcE8TC6o3OW3RPvyjt8SDyYOlOvJCnZgiBYGoZhcVv7bDmVJKkXtOzq29S9trNLt7Rcsqbl81u2tLbWURfdppbdptZUg6kkKWo96CgkSZK6ouTonoXA9p7fVousXXQlSenGllNJktJYT1pkJUlKJbacSpKU5nraIitJUiqw5VSSJEmSFDnDqSRJkiQpcoZTSZIkSVLkDKeSJEmSpMgZTiVJkiRJkTOcSpIkSZIiZziVJEmSJEXOcCpJkiRJipzhVJIkSZIUOcOpJEmSJClyhlNJkiRJUuQMp5IkSZKkyBlOJUmSJEmRM5xKkiQpvb0eg3sLYG5O/Pb1WNQVSeqGvKgLkCRJkrrt9RgsKoW62vjjdVXxxwD7l0RXl6Qus+VUkiRJ6euFsq3BtEldbXy7pLRiOJUkSVL6WlfdzvYqu/dKacZwKkmSpJQXYxMFW94mJ1xDQc1yYo+fA3MDIGz/SYtKdxxQezBWNcYmCviYHD6kgI+JsanTz5XUtiAMd/AfupcVFxeHFRUVUZchSZKkqL0ei3fNXVdN7OBSSo+YRW3uTs2787fUMvfpb1Oy6sHOnS/Ihb2OgU/fiLe29t0dNn8K4Zatx+Tlw6S5HY5VjbGJUtbTsjNxPjCX/pTQt9MvUcpGQRAsDcOwuK19tpxKkiQpdbweg7uHwN+nxrvmElI2ZsY2wRSgtk8+ZYf9R+fPG9bDO483n5NNa7cNptA8VrWjVtEyNtBqlCu1jdsldZ/hVJIkSamhaebdTWu32Vw9YJ82D29ve0/E9iymlPVU0UAIVNHANNYTtAiq1TS0XU9Y51I2Ug+4lIwkSZKi06L7LkFOvIWzleHr3qZq4LA2tyda2eHXbNcq2jQIrooGSresZvf6Taztt1vb9ayrirf6Pn0JTL4tvmPJZVsDd9/BcNRsl7mR2mDLqSRJkqLR1FLa1NW2jWAKUP78deRv2TYy5m+ppfz56xJeUnX/vXa4v7ZPPoQNHddTvz4eUv8+dduW4E1r4amLbF2V2mA4lSRJUjTaWqO0DSWrHmTu099mRM2bBGEDI2re7NpkSF3QmdbYD/vt1rN6Gja7DqvUBmfrlSRJUjTm5rDDpWAiMPOon3DLwV+FIGj3mBE1b1L5m7E9vFIApW2PXZUymbP1SpIkKfUMGB51Bdt5bMQJOwymCetOnIKvXYqa4VSSJEnRGF8edQXbaXcG4DBMXHfinJ1S8rVLUXO2XkmSJKlRezMDj1j3VgK68gIEcPQ8Z+uV2mDLqSRJknrfopnxmWxTTPJnBg5h0aVw95D4mFvXRZWaGU4lSZLUe16PxYPZiluirqRNvTIzcN26xuVlwsZ1UafFw7qU5ZytV5IkSb3jj8fBO49HXUWKCuDz99jdVxnP2XolSZIUrUUzDaY7FMKTF9jFV1nNcCpJkqTkS9FuvCklrIe/G1CVvQynkiRJUsqoj49BdbIkZSHDqSRJkpRSQrZOljQ1PlZXygKGU0mSJCmVvfM4zA3isxzbkqoMZjiVJEmS0sGmtfGW1LnB/2/vjmP0vuv7gL+/d06cnE1QcaIAAd9ZyGPQMaHpQGJNMrJES4YqhYCoMJcuqMy3QKmASUg0XtVs3SFU1LWoakhtLQ0jN1dF0BVRVG9JN0gHGlwkBKGIAfOdIZCQGIlwdjDx3Xd/PHe273xObN89932eu9dLevQ8z/d57vf72Hrke97+fr+fX+d2YMvpS9B8Z7KzDNhyYPrYltYFAACwCbzq3ZoirbU61/k7/en/TX785eTk8c747Ezy8HjnsUvT0EfMnAIA0H3X3dMJqCmtK9l4fvjQ6WC66OTx5Kv72tQDF0k4BQBgfVx3TzI+n9zwQLJlqHU1G9/skdYVwAURTgEAWF+7x5Lr9ifbh1tXsrFtfVHrCuCCCKcAAKy/3WPJO6ZjmW8XPfszjZHoK10Lp6WUu0spj5VSvrZwe1O3zgUAQJ/avrN1BRvX/C/sO6WvdHvm9A9rra9duH2+y+cCAKDfvG7C/tNump1xaRn6hmW9AAC0Y/9p983OdK6P+rmbWlcCz6nb4fS9pZSvl1LuK6X8UpfPBQBAP1rcfzpeFy43Q1f88KHk4fe0rgLOaVXhtJTyYCnl0RVutyb5eJJXJHltkh8l+YNzHGO8lDJVSpl68sknV1MOAAD97rp7ksFtravYuL718dYVwDmtKpzWWm+qtf6jFW5/VWt9otY6V2udT3IgyevPcYz9tdbRWuvoVVddtZpyAADYCK7/09YVbGz2n9Kjutmt9yVnPL0tyaPdOhcAABvI7rGF5b0uM9MVOvjSo7q55/T3SynfKKV8PckNST7QxXMBALCRXHdPcsMnNUrqhtkZs6f0pC3dOnCt9de7dWwAADaB3WOdW9IJU//rnUk92bSkDePh8c794t8v9ACXkgEAoPftHkv2Ppvc8EBSBltX0/9OHre8l57TtZlTAABYc4szff/z9rZ1bASzR1pXAEuYOQUAoL+capjEqmzf2boCWEI4BQCg/1x3T2eJ79YdrSvpX8/OaoxETxFOAQDoT7vHkjueSsZrJ6huGWpdUROTr3hrRvZ8LQN7n8zInq9l8hVvPb8fPHG00xhJQKVHCKcAAPS/3WPJdfuTwW2tK1lXk694a8av/6PMvODlqWUgMy94ecav/6PzD6gaI9FDhFMAADaG3WPJu2Y31X7Ufa//nRy/ZOmM8fFLhrLv9b9z/gfRGIkeIZwCALCxLO5HHbi0dSVdd2T7Nec1PvnYnoz87eEM/PVcRv72cCYf23P6xTJgaS89QTgFAGDj2T2W/LP7ku3DrSvpqp2zjz3v+ORjezL+jQOZeWYkNQOZeWYk4984cDqg1jl7T+kJwikAABvT7rHkHdMbOqBOfOX3MvTs8SVjQ88ez8RXfu/U833f/nCOzy3di3t8blv2ffvDpwfsPaUHCKcAAGxsr5vYsEt8x7736ez/4vsz/LPvp9T5DP/s+9n/xfdn7HufPvWeI8+sfD3Ts8btPaWxLa0LAACArto91rl/+M7k5GzbWrpg7HufXhJGl9t5+ZHMPDOy4vgS21cOsbBezJwCALDx7R5LfuNnnUZJW3e0rmZdTbzyrgwNHlsyNjR4LBOvvOv0wJahzgwzNCScAgCweeweS+54qhNStw8nKZ2wukGX/SbJ2DUHs/81ezN8+XRK5jN8+XT2v2Zvxq452HnD1h2da8QuzjBDI6XW2rqGU0ZHR+vU1FTrMgAA2Gy+M9lpCDQ707qS9TfeO3mAja+U8kitdXSl18ycAgDAYmffGx5IyiWtq1k/ZbB1BXCKcAoAAIt2jyVv/LPTS363DycvvbHzeCP6h+OtK4BTdOsFAIAz7R47e//lRlz2+9Ibk+vuaV0FnGLmFAAAns/ist/xekYzpRUMbsupGdcbHliYdV1jg9tWf4xXvTv51QdXfxxYQ2ZOAQDgQqw0s/pc7/3cTckPH1qbc28ZSgYvS+aOPf97VzK4Lbn+T3XmpSeZOQUAgG761QdXd33VxaZF24c7l3w58ZML+/lXvbsz4ztek3fNCqb0LDOnAADQbYuzrd+ZTL70vuTE0ef/me3DnaXEy53v3tftw8nrJoRR+oaZUwAAWC+7x5I7nlph7+qybsBbhjrBciWvm+i8fi4vvbFz/HdMC6b0FTOnAADQwpl7V091Az6SbN/53DOei+Pn+37oE6XW2rqGU0ZHR+vU1FTrMgAAAOiCUsojtdbRlV6zrBcAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoDnhFAAAgOaEUwAAAJoTTgEAAGhOOAUAAKA54RQAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoDnhFAAAgOaEUwAAAJoTTgEAAGhOOAUAAKA54RQAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoDnhFAAAgOaEUwAAAJoTTgEAAGhOOAUAAKA54RQAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAYA1NfmFZORfJwNv7txPfqF1RQD0A+EUAFgzk19Ixv8kmXkyqbVzP/6xY5n87duTsZLcNpIcmmxdJgA9SDgFANbMvk8mx08sHTs+ty37Dv/H5FeSXD6T3H178tH3NKkPgN4lnAIAa+bIU+cYf2ZnsiXJ6MLAZz6evKEkt1xpJhWAJJ1fEwDAZnZoMrl3X/LEkeQFL0qe/XnyzLHTr2+5NJk7mdT5pT/3wh3JBz6W3Dx2amjnpUczc2LHWafYefmRzoNty1746dHOTOq//1ed4w8MJvNzyYuHkzsnlhwbgI3NzCkAbGaHJpOPjCePz3Q2iT59dGkwTZKTv1gaTHcleVuS244m37w9eUfpzIK+YSATP/6tDNWlPz80eCwTr7yr82TZoU9ZPP78XOf+8YXlv2+wTxVgsxBOAWAzu3df8vPj5//+XUmuTbI9SVm4v3ZhPDVjJw5m/9N7M1ymUzKf4cuns/81ezN2zcHkZJKpi6jx8ZlOgBZQATa0UmttXcMpo6OjdWrqYn5rAQAXZHEp7+MzF/Zzb0snkC43m+RTy8Z2pbPHdFs6M6ZTSQ5fcKWnlYEkNbl6pyW/AH2qlPJIrXV0pdfsOQWAzeaj70k+c2+Si/gP6uV7Rp9r/HBWF0aXq/MLgXcmmb49+c/vS/7BryXf/3wyeyTZvjN53USyW2gF6EfCKQBsJh99T6dT7sU6lpVnTs+1l3QtLS4pXvz2Mnc0+dYZf5bZmeTh8c5jARWg79hzCgCbxaHJ1QXTpLM09+SysYvdS3o+FpsvvTPJ9Xn+/1Y/eTz56r4uFQNAN5k5BYDNYuJdqz/G4jLdtdxLei7LZ0rLef7c7JEuFANAtwmnALAZvOOXk2dPrM2x1nov6bmM5uK+qWzfudaVALAOLOsFgI3ut25KDv996you3LmaL51peU+nLUOdpkgA9B3hFAA2skOTydRDrau4OOdqsjSfTiidTfKthfsk2T6cXLdfMySAPmVZLwBsZPf2cXOgqSzdc5p0mi/9XZYuK/4/SS4bSj7kMjIA/czMKQBsRIcmk9tGksdnWldy8Q6nE0Rnc3qmdHkwXfTz48kfvm8diwNgrQmnALCRHJpMbrkyufv2rgTTya17MrLjcAaumsvIjsOZ3Lpnzc+xxOEkn0py/8L9czVi+unR5J+/IHlD6dxuvrLz9wFAX7CsFwA2ikOTyUfGO7OIXTC5dU/GrziQ46XTqWhmcCTjVxxInk7GThzsyjkv2DOzpx8/fTT5D3d0Ht9suS9Aryu1Lm9z187o6GidmurWVbwBYIO7+cpOIOuSkR2HMzM4ctb48Nx0po/u6tp5V+2FO5K/eap1FQAkKaU8UmsdXek1y3oBYCP46Hu6GkyT5MjAytcPPdd4z/hpd/9eAFgbwikA9LtDk8lnPt710+ycP3JB4wBwIYRTAOh3/2l9utROzN6Vobr04qND9VgmZu9al/NftMu2ta4AgPMgnAJAv+vyct5FYycOZv/TezM8N51S5zM8N539T+/tnWZI51JaFwDA+RBOAYDzNnbiYKaP7sr8k4OZPrqr94NpkjxzrLMnF4Ce5lIyANDvykBS51tX0c6uJKNJtiU5lmQqZ18P9TP3Jv/4V1xSBqCHmTkFgH63iYLp5NY9GdlxOANXzWVkx+FMvmxPcm2S7eks392ezvOzrmxTk3v3rXO1AFwI4RQA+t2Lh1tXsC4mt+7J+BUHMjM4kloGMjM4kvGTBzL5xJ6lb9ySzkzqck/oKgzQy4RTAOh3d060rmBd7Nv+4RwvSzvvHp/fln3f/vDZb16pQe/VPX49VoBNTjgFgH63SfZRHhlYOVweeWaF8WNnD+WfvmltCwJgTQmnAEBf2Dm/8rLcnZctGz+ZTlOk5T7/ieTQ5JrXBcDaEE4BYCO4YkfrCrpuYvauDNWlU6JD9Vgm5u9KZpPUdO7/Lmd3602Snx/XFAmghwmnALAR/NuPta6g68ZOHMz+p/dmeG46pc5neG46+5/em7EfHEw+leT+dO5XCqaLNEUC6FmucwoA9I2xEwczduLgxR9AUySAnmXmFAD63aHJ5CPjravoD5ukszFAPxJOAaDf3buvs5+S57dJOhsD9CPLegGg3z0+07qCNnYlGU3nmqbH0unQ+1z7TV88vB5VAXCRhFMA6HelJLW2rmJ97UpybU5/k9m+8DxZOaCWAUt6AXqcZb0A0O82WzBNOjOmy/+LfcvC+Ep+979Y0gvQ44RTAKD/bDvf8ZLc/YBgCtAHhFMA6HuldQHr79h5jr/lTsEUoE8IpwDQ795yZ+sK1t9UkpPLxk4ujJ/pg/esTz0ArJqGSADQ7xYD2H/bn8zPdZr/DG5JTv6ibV3dtNj06Lm69erOC9BXzJwCwEbwwXuS/30y+XJNvjSXPHwiecu7s95Lfie37snIjsMZuGouIzsOZ3Lrnu6d7HCSTyW5f+H+zGC65RLdeQH6jHAKABvVB+9J7v7kus0gTm7dk/ErDmRmcCS1DGRmcCTjVxzobkA9l3/3Z/aaAvSZUnuo/fzo6Gidmlq+WQQAWBOHJpO7fz1Jd373j+w4nJnBkbPGh+emM310V1fOeU5f7p3vNwCcVkp5pNa64oW/zJwCwGZx81hXV/keGdh5QeMAcCYNkQBgM7l6Z/L4TFcOvXP+yIozpzvnj1z8QXfluZsereSFOy7+fAA0Y+YUADaTOyeSy4aWjl02lIzemAwMrurQE7N3ZaguvdDoUD2Widm7Lu6Au5Jcm2R7OjO+2xeeP9cK4UsuTT7wsYs7HwBNCacAsJncPJZ8aH+nSVIpnfsP7U/++MHT3X4v0tiJg9n/9N4Mz02n1PkMz01n/9N7M3bi4MUdcDRnr/HasjC+khcPJ/vu0wgJoE9piAQALHXLlclPj7auInlnVt4jW9O5fEzSuVzO4nVeAeh5XWuIVEp5Wynlm6WU+VLK6LLXfruU8t1SyrdLKTev5jwAwDrqlf+3PvYc41suSe5+QDAF2EBWu6z30SRvSfLFMwdLKa9O8vYkv5zkliT3lFJWt5EFAFgfP/tJ6wo6ppKcXDZ2MsnXL3UdU4ANaFXdemut30qSUs5ac3Nrkj+vtZ5IcriU8t0kr0/y5dWcDwBYB13s6HtBFrvyntmtd+DG5P4H29UEQNd0qyHSNUm+f8bzHyyMnaWUMl5KmSqlTD355JNdKgcAOG8rdfS95NLkioVLtJR17Kd4OMmn0tljevjG5P2CKcBG9by/XUopD5ZSHl3hdutaFFBr3V9rHa21jl511VVrcUgAYDVW6ui7777k0FOdbr5fmuvc3/3A+l1T9C3v7nQUBmDDet5lvbXWmy7iuI8lefkZz1+2MAYA9IObx55/T+fy99w2snbLgQcGkzePa3gEsIl0a13OZ5O8vZSytZSyK8nuJF/p0rkAgF6w0nLg87G4TPjFw53Z2C/XzjVXBVOATWVVDZFKKbcl+eMkVyX561LK12qtN9dav1lK+Yskf59OX73frLXOrb5cAKBnLc6i3rvv/GZQX7gj+cDHdN0FIElSau2Vi5klo6OjdWpqqnUZAMBaODSZTPxG8uwvTo+N3mjvKMAmVkp5pNY6utJrq5o5BQA4p/PZtwoAC9axFzwAAACsTDgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoDnhFAAAgOaEUwAAAJoTTgEAAGhOOAUAAKA54RQAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoDnhFAAAgOaEUwAAAJoTTgEAAGhOOAUAAKA54RQAAIDmhFMAAACaE04BAABoTjgFAACgOeEUAACA5oRTAAAAmhNOAQAAaE44BQAAoLlSa21dwymllCeTzLSug55zZZKnWhcBC3we6SU+j/QSn0d6ic9j7xqutV610gs9FU5hJaWUqVrraOs6IPF5pLf4PNJLfB7pJT6P/cmyXgAAAJoTTgEAAGhOOKUf7G9dAJzB55Fe4vNIL/F5pJf4PPYhe04BAABozswpAAAAzQmn9LxSyt2llMdKKV9buL2pdU1sPqWUW0op3y6lfLeU8qHW9bC5lVKmSynfWPg3cap1PWw+pZT7Sik/LqU8esbYi0op/6OU8p2F+19qWSObxzk+j74/9iHhlH7xh7XW1y7cPt+6GDaXUspgkj9J8i+TvDrJnlLKq9tWBblh4d9El0qghfuT3LJs7ENJHqq17k7y0MJzWA/35+zPY+L7Y98RTgGe3+uTfLfW+v9qrb9I8udJbm1cE0AztdYvJvnJsuFbk3xi4fEnkrx5PWti8zrH55E+JJzSL95bSvn6wrINy4RYb9ck+f4Zz3+wMAat1CT/vZTySCllvHUxsODqWuuPFh4/nuTqlsVAfH/sO8IpPaGU8mAp5dEVbrcm+XiSVyR5bZIfJfmDlrUC9IBra63/JJ2l5r9ZSrm+dUFwptq5HIRLQtCS7499aEvrAiBJaq03nc/7SikHknyuy+XAco8lefkZz1+2MAZN1FofW7j/cSnlL9NZev7FtlVBniilvKTW+qNSykuS/Lh1QWxetdYnFh/7/tg/zJzS8xZ+wS26Lcmj53ovdMlXk+wupewqpVya5O1JPtu4JjapUsq2UsoLFh8n+Rfx7yK94bNJ7lh4fEeSv2pYC5uc74/9ycwp/eD3SymvTWd50HSSf9O0GjadWuvJUsp7kxxKMpjkvlrrNxuXxeZ1dZK/LKUknd/j/7XW+jdtS2KzKaUcTPLGJFeWUn6Q5HeTfCTJX5RS3pVkJsmvtauQzeQcn8c3+v7Yf0pnSwAAAAC0Y1kvAAAAzQmnAAAANCecAgAA0JxwCgAAQHPCKQAAAM0JpwAAADQnnAIAANCccAoAAEBz/x97r1FnKUv5/QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1152x1008 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "18.559155464172363\n"
     ]
    }
   ],
   "source": [
    "import data_utils_new\n",
    "import matplotlib.pyplot as plt \n",
    "import umap.umap_ as umap\n",
    "# Generate data from the gmm model and plot it\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "n_samples = np.ones(num_classes, dtype=int)\n",
    "n_samples[0] = 0\n",
    "n_samples *= 2000\n",
    "\n",
    "xx, yy = utils.sample_from_gaussians(means, covs, n_samples=n_samples)\n",
    "\n",
    "NUM_COLORS = num_classes\n",
    "\n",
    "reducer = umap.UMAP()\n",
    "\n",
    "umap_embedding = reducer.fit_transform(xx)\n",
    "\n",
    "plt.figure(figsize=(16,14))\n",
    "cm = plt.get_cmap('gist_rainbow')\n",
    "\n",
    "shift = 1 / len(data_utils_new.label_ids_13.keys())\n",
    "idx = 0\n",
    "for label in data_utils_new.label_ids_13:\n",
    "    ind = yy == data_utils_new.label_ids_13[label]\n",
    "    \n",
    "    plt.scatter(umap_embedding[:,0][ind], umap_embedding[:,1][ind], label=label, \\\n",
    "                color=cm(1.*idx/NUM_COLORS))\n",
    "    idx += 1\n",
    "\n",
    "plt.title(\"Embedding scatter-plot\")\n",
    "plt.legend()\n",
    "    \n",
    "plt.show()\n",
    "\n",
    "print(time.time() - start_time)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initalize loss function "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importlib.reload(wasserstein_utils)\n",
    "import tensorflow_addons as tfa\n",
    "\n",
    "Z_s = tf.keras.layers.Input(shape=(img_shape[0], img_shape[1], num_classes,) )\n",
    "X_labeled = tf.keras.layers.Input(shape=(img_shape[0], img_shape[1], 3,) )\n",
    "Y_s = tf.keras.backend.placeholder(shape=(None, img_shape[0], img_shape[1], num_classes), dtype='float32') # labels of input images oneHot\n",
    "Y_labeled = tf.keras.backend.placeholder(shape=(None, img_shape[0], img_shape[1], num_classes), dtype='float32') # labels of input images oneHot\n",
    "lambda1 = 1\n",
    "lambda2 = 5e-1\n",
    "lambda3 = 5e-1\n",
    "    \n",
    "loss_function = losses.weighted_ce_loss(num_classes, 0)\n",
    "wce_loss_labeled = loss_function(Y_labeled, classifier(deeplabv3(X_labeled)), from_logits=False)\n",
    "wce_loss_unlabeled = loss_function(Y_s, classifier(Z_s), from_logits=False)\n",
    "\n",
    "# Wasserstein matcing loss\n",
    "# theta = tf.keras.backend.placeholder(shape = (num_projections, num_classes), dtype='float32')\n",
    "# matching_loss = wasserstein_utils.sWasserstein_hd(deeplabv3(X), deeplabv3(Z_s), theta, nclass=num_classes, Cp=None, Cq=None,)\n",
    "\n",
    "# Overall loss is a weighted combination of the losses\n",
    "# total_loss = lambda1 * wce_loss_unlabeled + lambda2 * wce_loss_labeled + lambda3 * matching_loss\n",
    "total_loss = lambda1 * wce_loss_unlabeled + lambda2 * wce_loss_labeled\n",
    "# total_loss = lambda2 * wce_loss_labeled\n",
    "\n",
    "params = deeplabv3.trainable_weights + classifier.trainable_weights\n",
    "\n",
    "# Optimizer and training setup\n",
    "opt = tf.keras.optimizers.Adam(lr=1e-4, epsilon=1e-6, decay=1e-6)\n",
    "# use SGD w/ weight decay\n",
    "# opt = tfa.optimizers.SGDW(learning_rate=0.0001, momentum=0.9, weight_decay=2.0e-4, name='SGDW')\n",
    "\n",
    "updates = opt.get_updates(total_loss, params)\n",
    "train = tf.keras.backend.function(inputs=[X,  Z_s, Y_s, X_labeled, Y_labeled], \n",
    "                                  outputs=[total_loss, wce_loss_labeled, wce_loss_unlabeled], updates=updates)\n",
    "\n",
    "# train = tf.keras.backend.function(inputs=[X, Z_s, Y_s, theta, X_labeled, Y_labeled], \n",
    "#                                   outputs=[total_loss, wce_loss_labeled, wce_loss_unlabeled, matching_loss], updates=updates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/common/software/nersc/shasta2105/tensorflow/2.6.0/lib/python3.8/site-packages/tensorflow/python/framework/indexed_slices.py:447: UserWarning: Converting sparse IndexedSlices(IndexedSlices(indices=Tensor(\"Adam/gradients_1/gradients/GatherV2_grad/Reshape_1:0\", shape=(None,), dtype=int32), values=Tensor(\"Adam/gradients_1/gradients/GatherV2_grad/Reshape:0\", shape=(None, 100), dtype=float32), dense_shape=Tensor(\"Adam/gradients_1/gradients/GatherV2_grad/Cast:0\", shape=(2,), dtype=int32))) to a dense Tensor of unknown shape. This may consume a large amount of memory.\n",
      "  warnings.warn(\n",
      "/global/common/software/nersc/shasta2105/tensorflow/2.6.0/lib/python3.8/site-packages/tensorflow/python/framework/indexed_slices.py:447: UserWarning: Converting sparse IndexedSlices(IndexedSlices(indices=Tensor(\"Adam/gradients_1/gradients/GatherV2_1_grad/Reshape_1:0\", shape=(None,), dtype=int32), values=Tensor(\"Adam/gradients_1/gradients/GatherV2_1_grad/Reshape:0\", shape=(None, 100), dtype=float32), dense_shape=Tensor(\"Adam/gradients_1/gradients/GatherV2_1_grad/Cast:0\", shape=(2,), dtype=int32))) to a dense Tensor of unknown shape. This may consume a large amount of memory.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "# old loss function code\n",
    "# importlib.reload(wasserstein_utils)\n",
    "\n",
    "# Z_s = tf.keras.layers.Input(shape=(img_shape[0], img_shape[1], num_classes,) )\n",
    "# Y_s = tf.keras.backend.placeholder(shape=(None, img_shape[0], img_shape[1], num_classes), dtype='float32') #labels of input images oneHot\n",
    "# lambda2 = 5e-1\n",
    "\n",
    "# loss_function = losses.weighted_ce_loss(num_classes, 0)\n",
    "# wce_loss = loss_function(Y_s, classifier(Z_s), from_logits=False)\n",
    "\n",
    "# # Wasserstein matcing loss\n",
    "# theta = tf.keras.backend.placeholder(shape = (num_projections, num_classes), dtype='float32')\n",
    "# matching_loss = wasserstein_utils.sWasserstein_hd(deeplabv3(X), Z_s, theta, nclass=num_classes, Cp=None, Cq=None,)\n",
    "\n",
    "# # Overall loss is a weighted combination of the two losses\n",
    "# total_loss = wce_loss + lambda2*matching_loss\n",
    "\n",
    "# params = deeplabv3.trainable_weights + classifier.trainable_weights\n",
    "\n",
    "# # Optimizer and training setup\n",
    "# opt = tf.keras.optimizers.Adam(lr=1e-4, epsilon=1e-6, decay=1e-6)\n",
    "\n",
    "# updates = opt.get_updates(total_loss, params)\n",
    "# train = tf.keras.backend.function(inputs=[X,Z_s,Y_s,theta], outputs=[total_loss, wce_loss, matching_loss], updates=updates)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model weights\n"
     ]
    }
   ],
   "source": [
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "print(\"Loaded model weights\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Source clustering & active target sample selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import heapq\n",
    "\n",
    "def calculate_mean_vector(images, labels, vector_size, num_classes):\n",
    "    n = len(images)\n",
    "    full_objective_vector = np.zeros([n, num_classes, vector_size]) \n",
    "    i = 0\n",
    "    for i in range(len(images)):\n",
    "        image = np.expand_dims(images[i], axis=0)\n",
    "        f_E = deeplabv3(image)[0, :, :, :]\n",
    "        F_x = np.zeros((num_classes, vector_size)) # feature map for this image\n",
    "        for c in range(0, num_classes - 1): \n",
    "            F_c_x = tf.math.multiply(f_E[:,:,c], labels[i][:,:,c]) / np.sum(labels[i][:,:,c])\n",
    "            F_c_x_expanded = np.expand_dims(np.expand_dims(F_c_x, axis=0), axis=3)\n",
    "            F_c_x = tf.keras.layers.AveragePooling2D(pool_size=(2, 2), strides=(32, 64), padding='valid')(F_c_x_expanded)\n",
    "            F_c_x = tf.reshape(F_c_x, [-1])\n",
    "            F_x[c] = F_c_x\n",
    "        full_objective_vector[i, :] = F_x\n",
    "        print(i)\n",
    "        i += 1\n",
    "    return full_objective_vector\n",
    "    \n",
    "import faiss\n",
    "\n",
    "def run_kmeans(x, nmb_clusters, ngpu, weighted=False, verbose=False):\n",
    "    \"\"\"Runs kmeans on 1 or more GPUs.\n",
    "    Args:\n",
    "        x: data\n",
    "        nmb_clusters (int): number of clusters\n",
    "        ngpu: number of available gpus\n",
    "        weighted: bool, indicates whether to use weighted K-means\n",
    "    Returns:\n",
    "        list: ids of data in each cluster\n",
    "    \"\"\"\n",
    "    n_data, d = x.shape\n",
    "    \n",
    "    def H(x):\n",
    "        return -(x * np.log2(x, where=x > 0)).sum(1)\n",
    "    \n",
    "    # faiss implementation of k-means\n",
    "    clus = faiss.Clustering(d, nmb_clusters)\n",
    "\n",
    "    # Change faiss seed at each k-means so that the randomly picked\n",
    "    # initialization centroids do not correspond to the same feature ids\n",
    "    # from an epoch to another.\n",
    "    clus.seed = np.random.randint(1234)\n",
    "\n",
    "    clus.niter = 20\n",
    "    clus.max_points_per_centroid = 10000000\n",
    "    res = [faiss.StandardGpuResources() for i in range(ngpu)]\n",
    "    flat_config = []\n",
    "    for i in range(ngpu):\n",
    "        cfg = faiss.GpuIndexFlatConfig()\n",
    "        cfg.useFloat16 = False\n",
    "        cfg.device = i\n",
    "        flat_config.append(cfg)\n",
    "        \n",
    "    if ngpu == 1:\n",
    "        index = faiss.GpuIndexFlatL2(res[0], d, flat_config[0])\n",
    "    else:\n",
    "        indexes = [faiss.GpuIndexFlatL2(res[i], d, flat_config[i])\n",
    "                   for i in range(ngpu)]\n",
    "        index = faiss.IndexReplicas()\n",
    "        for sub_index in indexes:\n",
    "            index.addIndex(sub_index)\n",
    "    if weighted:\n",
    "        weights = H(x)\n",
    "        print(weights.shape)\n",
    "        clus.train(x, index, weights=weights)\n",
    "    else:\n",
    "        clus.train(x, index)\n",
    "\n",
    "    _, I = index.search(x, 1)\n",
    "    stats = clus.iteration_stats\n",
    "    losses = np.array([\n",
    "        stats.at(i).obj for i in range(stats.size())\n",
    "    ])\n",
    "    if verbose:\n",
    "        print('k-means loss evolution: {0}'.format(losses))\n",
    "    centroids = faiss.vector_float_to_array(clus.centroids)\n",
    "\n",
    "    return centroids, [int(n[0]) for n in I], losses[-1]\n",
    "\n",
    "def get_active_XY(images, pred_labels, actual_labels, vector_size, batch_size, sampling_rate, num_classes, strategy = \"ours\"):\n",
    "    ''' calculate active learning criteria and returns vectors for training and list of selected indices'''\n",
    "    \n",
    "    if strategy == \"random\":\n",
    "        selected_idx = np.random.choice(batch_size, int(batch_size * sampling_rate), replace=False)\n",
    "    elif strategy in (\"mada\", \"ours\"):\n",
    "        al_list = []\n",
    "        n = len(images)\n",
    "        full_objective_vector = np.zeros([n, num_classes, vector_size]) \n",
    "        selected_n = int(sampling_rate * n)\n",
    "        \n",
    "        # if strategy == \"ours\":\n",
    "        #     full_objective_vector = calculate_mean_vector(images, pred_labels, vector_size)\n",
    "        #     full_objective_vector = np.nan_to_num(np.reshape(full_objective_vector, (full_objective_vector.shape[0], \n",
    "        #                                              full_objective_vector.shape[1] * full_objective_vector.shape[2])).astype('float32'), nan=0.0)\n",
    "        #     cluster_centroids, cluster_index, cluster_loss = run_kmeans(full_objective_vector, nmb_clusters=10, ngpu=ngpu, weighted=True, verbose=False)\n",
    "        #     cluster_centroids = cluster_centroids.reshape(ncentroids, 20, 256)\n",
    "        \n",
    "        for i in range(len(images)):\n",
    "            # try:\n",
    "            image = np.expand_dims(images[i], axis=0)\n",
    "            f_E = deeplabv3(image)[0, :, :, :]\n",
    "            F_x = np.zeros((num_classes, vector_size)) # feature map for this image\n",
    "            for c in range(0, num_classes - 1): \n",
    "                F_c_x = tf.math.multiply(f_E[:,:,c], pred_labels[i][:,:,c]) / np.sum(pred_labels[i][:,:,c])\n",
    "                F_c_x_expanded = np.expand_dims(np.expand_dims(F_c_x, axis=0), axis=3)\n",
    "                F_c_x = tf.keras.layers.AveragePooling2D(pool_size=(2, 2), strides=(32, 64), padding='valid')(F_c_x_expanded)\n",
    "                F_c_x = tf.reshape(F_c_x, [-1])\n",
    "                F_x[c] = F_c_x\n",
    "            # calculate active learning criteria\n",
    "            loss = []\n",
    "            for centroid in cluster_centroids:\n",
    "                new_loss = np.nanmean((F_x - centroid) ** 2)\n",
    "                loss.append(new_loss)\n",
    "            min_loss = min(loss)\n",
    "            if strategy == \"mada\":\n",
    "                al_list.append(min_loss)\n",
    "            elif strategy == \"ours\":\n",
    "                pred_labels_np = pred_labels.numpy()\n",
    "                entropy_weight = np.mean(-(pred_labels_np * np.log2(pred_labels_np, where=pred_labels_np > 0)).sum())\n",
    "                al_list.append(entropy_weight * min_loss)\n",
    "        \n",
    "        # find idx vals which maximize criteria\n",
    "        selected_idx = list(map(al_list.index, heapq.nlargest(selected_n, al_list)))\n",
    "            \n",
    "        \n",
    "    # return subset of images / labels for training\n",
    "    X_labeled = []\n",
    "    Y_labeled = []\n",
    "    for idx in selected_idx:\n",
    "        X_labeled.append(images[idx])\n",
    "        Y_labeled.append(actual_labels[idx])\n",
    "        \n",
    "    X_eval = np.delete(images, selected_idx, 0)\n",
    "    Y_eval = np.delete(actual_labels, selected_idx, 0)\n",
    "        \n",
    "    return np.asarray(X_labeled), np.asarray(Y_labeled), X_eval, Y_eval, selected_idx"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Clustering on source domain (SYNTHIA) for initial adaptation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "9\n",
      "10\n",
      "11\n",
      "12\n",
      "13\n",
      "14\n",
      "15\n",
      "16\n",
      "17\n",
      "18\n",
      "19\n",
      "20\n",
      "21\n",
      "22\n",
      "23\n",
      "24\n",
      "25\n",
      "26\n",
      "27\n",
      "28\n",
      "29\n",
      "30\n",
      "31\n",
      "32\n",
      "33\n",
      "34\n",
      "35\n",
      "36\n",
      "37\n",
      "38\n",
      "39\n",
      "40\n",
      "41\n",
      "42\n",
      "43\n",
      "44\n",
      "45\n",
      "46\n",
      "47\n",
      "48\n",
      "49\n",
      "k-means loss evolution: [0.00413724 0.00367611 0.0035106  0.00341629 0.00331405 0.00322836\n",
      " 0.00310121 0.00310121 0.00310121 0.00310121 0.00310121 0.00310121\n",
      " 0.00310121 0.00310121 0.00310121 0.00310121 0.00310121 0.00310121\n",
      " 0.00310121 0.00310121]\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(123)\n",
    "ncentroids = 10\n",
    "ngpu = 1\n",
    "seed = 100\n",
    "\n",
    "source_train_data, source_train_labels = utils.sample_batch(source_data_dir, batch_size=50, seed=seed)\n",
    "source_train_labels = tf.keras.utils.to_categorical(source_train_labels, num_classes=num_classes)\n",
    "\n",
    "vector_size = 256\n",
    "full_objective_vector = calculate_mean_vector(source_train_data, source_train_labels, vector_size, num_classes)\n",
    "\n",
    "tensor_sample = np.nan_to_num(np.reshape(full_objective_vector, (full_objective_vector.shape[0], full_objective_vector.shape[1] * full_objective_vector.shape[2])).astype('float32'), nan=0.0)\n",
    "cluster_centroids, cluster_index, cluster_loss = run_kmeans(tensor_sample, ncentroids, ngpu, weighted=False, verbose=True)\n",
    "with open('./anchors/cluster_centroids_full_%d.pkl' % ncentroids,'wb') as f: \n",
    "    pickle.dump(cluster_centroids, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initial SYNTHIA->VIPER adaptation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model weights\n",
      "Loaded means and covariances for replay buffer\n"
     ]
    }
   ],
   "source": [
    "loss = []\n",
    "target_miou = []\n",
    "# load model weights\n",
    "# deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "# classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_adapted.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_classifier_adapted.h5')\n",
    "\n",
    "print(\"Loaded model weights\")\n",
    "\n",
    "# load replay buffer params\n",
    "means = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "means = np.nan_to_num(means)\n",
    "covs = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "print(\"Loaded means and covariances for replay buffer\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mIoU at 19 epochs: 0.45033872586030227, best mIoU: 0.45033872586030227\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-19-0c123131a2e6>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m    130\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mitr\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mepochs2\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    131\u001b[0m         \u001b[0;31m# run for best epoch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 132\u001b[0;31m         \u001b[0mbest_cat_iou\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbest_miou2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mutils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcompute_miou\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcityscapes_data_dir\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbest_combined\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata_utils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlabel_ids_13\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mids_to_ignore\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mall_selected_idx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    133\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mbest_cat_iou\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    134\u001b[0m             \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbest_cat_iou\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda/utils.py\u001b[0m in \u001b[0;36mcompute_miou\u001b[0;34m(data_dir, model, label_ids, ids_to_ignore)\u001b[0m\n\u001b[1;32m    145\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    146\u001b[0m         \u001b[0my_true\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mio\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mread_label\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata_dir\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\"_label.npy\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 147\u001b[0;31m         \u001b[0my_hat\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mravel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    148\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    149\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mlabel\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mlabel_ids\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36m_method_wrapper\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m     86\u001b[0m       raise ValueError('{} is not supported in multi-worker mode.'.format(\n\u001b[1;32m     87\u001b[0m           method.__name__))\n\u001b[0;32m---> 88\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mmethod\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     89\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     90\u001b[0m   return tf_decorator.make_decorator(\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py\u001b[0m in \u001b[0;36mpredict\u001b[0;34m(self, x, batch_size, verbose, steps, callbacks, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1266\u001b[0m           \u001b[0;32mfor\u001b[0m \u001b[0mstep\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msteps\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1267\u001b[0m             \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_predict_batch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1268\u001b[0;31m             \u001b[0mtmp_batch_outputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpredict_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0miterator\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1269\u001b[0m             \u001b[0;31m# Catch OutOfRangeError for Datasets of unknown size.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1270\u001b[0m             \u001b[0;31m# This blocks until the batch has finished executing.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    578\u001b[0m         \u001b[0mxla_context\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mExit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    579\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 580\u001b[0;31m       \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    581\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    582\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mtracing_count\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_tracing_count\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/def_function.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, *args, **kwds)\u001b[0m\n\u001b[1;32m    616\u001b[0m       \u001b[0;31m# In this case we have not created variables on the first call. So we can\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    617\u001b[0m       \u001b[0;31m# run the first trace but we should fail if variables are created.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 618\u001b[0;31m       \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_stateful_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    619\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_created_variables\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    620\u001b[0m         raise ValueError(\"Creating variables on a non-first call to a function\"\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   2418\u001b[0m     \u001b[0;32mwith\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_lock\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2419\u001b[0m       \u001b[0mgraph_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_maybe_define_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2420\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mgraph_function\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_filtered_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# pylint: disable=protected-access\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2421\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2422\u001b[0m   \u001b[0;34m@\u001b[0m\u001b[0mproperty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_filtered_call\u001b[0;34m(self, args, kwargs)\u001b[0m\n\u001b[1;32m   1663\u001b[0m          if isinstance(t, (ops.Tensor,\n\u001b[1;32m   1664\u001b[0m                            resource_variable_ops.BaseResourceVariable))),\n\u001b[0;32m-> 1665\u001b[0;31m         self.captured_inputs)\n\u001b[0m\u001b[1;32m   1666\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1667\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m_call_flat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcaptured_inputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcancellation_manager\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36m_call_flat\u001b[0;34m(self, args, captured_inputs, cancellation_manager)\u001b[0m\n\u001b[1;32m   1744\u001b[0m       \u001b[0;31m# No tape is watching; skip to running the function.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1745\u001b[0m       return self._build_call_outputs(self._inference_function.call(\n\u001b[0;32m-> 1746\u001b[0;31m           ctx, args, cancellation_manager=cancellation_manager))\n\u001b[0m\u001b[1;32m   1747\u001b[0m     forward_backward = self._select_forward_and_backward_functions(\n\u001b[1;32m   1748\u001b[0m         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/function.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, ctx, args, cancellation_manager)\u001b[0m\n\u001b[1;32m    596\u001b[0m               \u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    597\u001b[0m               \u001b[0mattrs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mattrs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 598\u001b[0;31m               ctx=ctx)\n\u001b[0m\u001b[1;32m    599\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    600\u001b[0m           outputs = execute.execute_with_cancellation(\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# w/o  wasserstein loss\n",
    "# Adam loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    # theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda2, label='active learning ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mIoU at 19 epochs: 0.518859441463764, best mIoU: 0.5465702276963454\n",
      "ignore 0.0\n",
      "road 0.8884529720768418\n",
      "sidewalk 0.5858592801531234\n",
      "building 0.6421096840253216\n",
      "traffic light 0.538680884598599\n",
      "traffic sign 0.014632960903552065\n",
      "vegetation 0.34836145079767067\n",
      "sky 0.9209314487803937\n",
      "person 0.33810035465072164\n",
      "rider 0.0\n",
      "car 0.9839621513734922\n",
      "bus 0.0\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.3757922276685512\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# w/o  wasserstein loss\n",
    "# Adam loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    # theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda2, label='active learning ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mIoU at 19 epochs: 0.5367304407633268, best mIoU: 0.5779664608148428\n",
      "ignore 0.0\n",
      "road 0.8927570572276865\n",
      "sidewalk 0.5700248424532656\n",
      "building 0.656345551854552\n",
      "traffic light 0.5547702647577254\n",
      "traffic sign 0.12167058060130395\n",
      "vegetation 0.36339353637753247\n",
      "sky 0.9189963498348828\n",
      "person 0.3685971992311049\n",
      "rider 0.0\n",
      "car 0.981565333375478\n",
      "bus 0.056599243742872576\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.39176571138974314\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# w/o  wasserstein loss\n",
    "# Adam loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    # theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda2, label='active learning ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mIoU at 1 epochs: 0.5210493986423199, best mIoU: 0.5210493986423199\n",
      "Iteration: 2\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31m_FallbackException\u001b[0m                        Traceback (most recent call last)",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name)\u001b[0m\n\u001b[1;32m    925\u001b[0m         \u001b[0;34m\"explicit_paddings\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"data_format\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 926\u001b[0;31m         \"dilations\", dilations)\n\u001b[0m\u001b[1;32m    927\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0m_result\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31m_FallbackException\u001b[0m: Expecting int64_t value for attr strides, got numpy.int32",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-12-7a58995b9c73>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     35\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     36\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mactive_learning\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 37\u001b[0;31m         \u001b[0mpred_target_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdeeplabv3\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     38\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     39\u001b[0m         \u001b[0mX_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mselected_idx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_active_XY\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_target_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_train_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m256\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_classes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstrategy\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    966\u001b[0m           with base_layer_utils.autocast_context_manager(\n\u001b[1;32m    967\u001b[0m               self._compute_dtype):\n\u001b[0;32m--> 968\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcast_inputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    969\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle_activity_regularization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    970\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_set_mask_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_masks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/network.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs, training, mask)\u001b[0m\n\u001b[1;32m    717\u001b[0m     return self._run_internal_graph(\n\u001b[1;32m    718\u001b[0m         \u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtraining\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmask\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmask\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 719\u001b[0;31m         convert_kwargs_to_constants=base_layer_utils.call_context().saving)\n\u001b[0m\u001b[1;32m    720\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    721\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mcompute_output_shape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/network.py\u001b[0m in \u001b[0;36m_run_internal_graph\u001b[0;34m(self, inputs, training, mask, convert_kwargs_to_constants)\u001b[0m\n\u001b[1;32m    886\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    887\u001b[0m           \u001b[0;31m# Compute outputs.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 888\u001b[0;31m           \u001b[0moutput_tensors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlayer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcomputed_tensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    889\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    890\u001b[0m           \u001b[0;31m# Update tensor_dict.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    966\u001b[0m           with base_layer_utils.autocast_context_manager(\n\u001b[1;32m    967\u001b[0m               self._compute_dtype):\n\u001b[0;32m--> 968\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcast_inputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    969\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle_activity_regularization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    970\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_set_mask_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_masks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/layers/convolutional.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m    205\u001b[0m       \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marray_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_compute_causal_padding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    206\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 207\u001b[0;31m     \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_convolution_op\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkernel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    208\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    209\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_bias\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m   1104\u001b[0m           call_from_convolution=False)\n\u001b[1;32m   1105\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1106\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconv_op\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1107\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1108\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m    636\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    637\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m__call__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m  \u001b[0;31m# pylint: disable=redefined-builtin\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 638\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    639\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    640\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m    235\u001b[0m         \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpadding\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    236\u001b[0m         \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 237\u001b[0;31m         name=self.name)\n\u001b[0m\u001b[1;32m    238\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    239\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, data_format, dilations, name, filters)\u001b[0m\n\u001b[1;32m   2012\u001b[0m                            \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2013\u001b[0m                            \u001b[0mdilations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdilations\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2014\u001b[0;31m                            name=name)\n\u001b[0m\u001b[1;32m   2015\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2016\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name)\u001b[0m\n\u001b[1;32m    931\u001b[0m             \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstrides\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0muse_cudnn_on_gpu\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0muse_cudnn_on_gpu\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    932\u001b[0m             \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpadding\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 933\u001b[0;31m             data_format=data_format, dilations=dilations, name=name, ctx=_ctx)\n\u001b[0m\u001b[1;32m    934\u001b[0m       \u001b[0;32mexcept\u001b[0m \u001b[0m_core\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_SymbolicException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    935\u001b[0m         \u001b[0;32mpass\u001b[0m  \u001b[0;31m# Add nodes to the TensorFlow graph.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d_eager_fallback\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name, ctx)\u001b[0m\n\u001b[1;32m   1020\u001b[0m   explicit_paddings, \"data_format\", data_format, \"dilations\", dilations)\n\u001b[1;32m   1021\u001b[0m   _result = _execute.execute(b\"Conv2D\", 1, inputs=_inputs_flat, attrs=_attrs,\n\u001b[0;32m-> 1022\u001b[0;31m                              ctx=ctx, name=name)\n\u001b[0m\u001b[1;32m   1023\u001b[0m   \u001b[0;32mif\u001b[0m \u001b[0m_execute\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmust_record_gradient\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1024\u001b[0m     _execute.record_gradient(\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# w/o  wasserstein loss\n",
    "# Adam loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    # theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda2, label='active learning ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-10-6a2cb7de2034>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     52\u001b[0m         \u001b[0mpred_target_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdeeplabv3\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     53\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 54\u001b[0;31m         \u001b[0mX_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mselected_idx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_active_XY\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_target_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_train_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m256\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_classes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstrategy\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     55\u001b[0m         \u001b[0mY_labeled\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mutils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto_categorical\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mY_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_classes\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mnum_classes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     56\u001b[0m         \u001b[0mall_selected_idx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mselected_idx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-9-f73bc27b5e20>\u001b[0m in \u001b[0;36mget_active_XY\u001b[0;34m(images, pred_labels, actual_labels, vector_size, batch_size, sampling_rate, num_classes, strategy)\u001b[0m\n\u001b[1;32m    120\u001b[0m             \u001b[0;32melif\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m\"ours\"\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    121\u001b[0m                 \u001b[0mpred_labels_np\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpred_labels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnumpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 122\u001b[0;31m                 \u001b[0mentropy_weight\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmean\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpred_labels_np\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpred_labels_np\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mwhere\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpred_labels_np\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    123\u001b[0m                 \u001b[0mal_list\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mentropy_weight\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mmin_loss\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    124\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Error in callback <function flush_figures at 0x15538b537400> (for post_execute):\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/ipykernel/pylab/backend_inline.py\u001b[0m in \u001b[0;36mflush_figures\u001b[0;34m()\u001b[0m\n\u001b[1;32m    119\u001b[0m         \u001b[0;31m# ignore the tracking, just draw and close all figures\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    120\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 121\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    122\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    123\u001b[0m             \u001b[0;31m# safely show traceback if in IPython, else raise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/ipykernel/pylab/backend_inline.py\u001b[0m in \u001b[0;36mshow\u001b[0;34m(close, block)\u001b[0m\n\u001b[1;32m     41\u001b[0m             display(\n\u001b[1;32m     42\u001b[0m                 \u001b[0mfigure_manager\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcanvas\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 43\u001b[0;31m                 \u001b[0mmetadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0m_fetch_figure_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigure_manager\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcanvas\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     44\u001b[0m             )\n\u001b[1;32m     45\u001b[0m     \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/display.py\u001b[0m in \u001b[0;36mdisplay\u001b[0;34m(include, exclude, metadata, transient, display_id, *objs, **kwargs)\u001b[0m\n\u001b[1;32m    311\u001b[0m             \u001b[0mpublish_display_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmetadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmetadata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    312\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 313\u001b[0;31m             \u001b[0mformat_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmd_dict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minclude\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minclude\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexclude\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mexclude\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    314\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mformat_dict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    315\u001b[0m                 \u001b[0;31m# nothing to display (e.g. _ipython_display_ took over)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/formatters.py\u001b[0m in \u001b[0;36mformat\u001b[0;34m(self, obj, include, exclude)\u001b[0m\n\u001b[1;32m    178\u001b[0m             \u001b[0mmd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    179\u001b[0m             \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 180\u001b[0;31m                 \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mformatter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    181\u001b[0m             \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    182\u001b[0m                 \u001b[0;31m# FIXME: log the exception\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<decorator-gen-9>\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, obj)\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/formatters.py\u001b[0m in \u001b[0;36mcatch_format_error\u001b[0;34m(method, self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    222\u001b[0m     \u001b[0;34m\"\"\"show traceback on failed format call\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    223\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 224\u001b[0;31m         \u001b[0mr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmethod\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    225\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mNotImplementedError\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    226\u001b[0m         \u001b[0;31m# don't warn on NotImplementedErrors\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/formatters.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, obj)\u001b[0m\n\u001b[1;32m    339\u001b[0m                 \u001b[0;32mpass\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    340\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 341\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mprinter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    342\u001b[0m             \u001b[0;31m# Finally look for special method names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    343\u001b[0m             \u001b[0mmethod\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_real_method\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_method\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(fig)\u001b[0m\n\u001b[1;32m    246\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    247\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'png'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 248\u001b[0;31m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m'png'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    249\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;34m'retina'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m'png2x'\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mformats\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    250\u001b[0m         \u001b[0mpng_formatter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfor_type\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mFigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mlambda\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mretina_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/IPython/core/pylabtools.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(fig, fmt, bbox_inches, **kwargs)\u001b[0m\n\u001b[1;32m    130\u001b[0m         \u001b[0mFigureCanvasBase\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfig\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    131\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 132\u001b[0;31m     \u001b[0mfig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcanvas\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint_figure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbytes_io\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    133\u001b[0m     \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbytes_io\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgetvalue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    134\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfmt\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'svg'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/backend_bases.py\u001b[0m in \u001b[0;36mprint_figure\u001b[0;34m(self, filename, dpi, facecolor, edgecolor, orientation, format, bbox_inches, pad_inches, bbox_extra_artists, backend, **kwargs)\u001b[0m\n\u001b[1;32m   2191\u001b[0m                            else suppress())\n\u001b[1;32m   2192\u001b[0m                     \u001b[0;32mwith\u001b[0m \u001b[0mctx\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2193\u001b[0;31m                         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrenderer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2194\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2195\u001b[0m                     bbox_inches = self.figure.get_tightbbox(\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/artist.py\u001b[0m in \u001b[0;36mdraw_wrapper\u001b[0;34m(artist, renderer, *args, **kwargs)\u001b[0m\n\u001b[1;32m     39\u001b[0m                 \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstart_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0martist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrenderer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0martist\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_agg_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/figure.py\u001b[0m in \u001b[0;36mdraw\u001b[0;34m(self, renderer)\u001b[0m\n\u001b[1;32m   1862\u001b[0m             \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpatch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrenderer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1863\u001b[0m             mimage._draw_list_compositing_images(\n\u001b[0;32m-> 1864\u001b[0;31m                 renderer, self, artists, self.suppressComposite)\n\u001b[0m\u001b[1;32m   1865\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1866\u001b[0m             \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclose_group\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'figure'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/image.py\u001b[0m in \u001b[0;36m_draw_list_compositing_images\u001b[0;34m(renderer, parent, artists, suppress_composite)\u001b[0m\n\u001b[1;32m    129\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mnot_composite\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mhas_images\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    130\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0martists\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 131\u001b[0;31m             \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrenderer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    132\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    133\u001b[0m         \u001b[0;31m# Composite any adjacent images together\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/artist.py\u001b[0m in \u001b[0;36mdraw_wrapper\u001b[0;34m(artist, renderer, *args, **kwargs)\u001b[0m\n\u001b[1;32m     39\u001b[0m                 \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstart_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0martist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrenderer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0martist\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_agg_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/cbook/deprecation.py\u001b[0m in \u001b[0;36mwrapper\u001b[0;34m(*inner_args, **inner_kwargs)\u001b[0m\n\u001b[1;32m    409\u001b[0m                          \u001b[0;32melse\u001b[0m \u001b[0mdeprecation_addendum\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    410\u001b[0m                 **kwargs)\n\u001b[0;32m--> 411\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minner_args\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0minner_kwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    412\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    413\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mwrapper\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/axes/_base.py\u001b[0m in \u001b[0;36mdraw\u001b[0;34m(self, renderer, inframe)\u001b[0m\n\u001b[1;32m   2745\u001b[0m             \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstop_rasterizing\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2746\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2747\u001b[0;31m         \u001b[0mmimage\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_draw_list_compositing_images\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrenderer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0martists\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2748\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2749\u001b[0m         \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclose_group\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'axes'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/image.py\u001b[0m in \u001b[0;36m_draw_list_compositing_images\u001b[0;34m(renderer, parent, artists, suppress_composite)\u001b[0m\n\u001b[1;32m    129\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mnot_composite\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mhas_images\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    130\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0martists\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 131\u001b[0;31m             \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrenderer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    132\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    133\u001b[0m         \u001b[0;31m# Composite any adjacent images together\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/artist.py\u001b[0m in \u001b[0;36mdraw_wrapper\u001b[0;34m(artist, renderer, *args, **kwargs)\u001b[0m\n\u001b[1;32m     39\u001b[0m                 \u001b[0mrenderer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstart_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 41\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mdraw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0martist\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrenderer\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     42\u001b[0m         \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0martist\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_agg_filter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/axis.py\u001b[0m in \u001b[0;36mdraw\u001b[0;34m(self, renderer, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1164\u001b[0m         \u001b[0mticks_to_draw\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_update_ticks\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1165\u001b[0m         ticklabelBoxes, ticklabelBoxes2 = self._get_tick_bboxes(ticks_to_draw,\n\u001b[0;32m-> 1166\u001b[0;31m                                                                 renderer)\n\u001b[0m\u001b[1;32m   1167\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1168\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mtick\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mticks_to_draw\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/axis.py\u001b[0m in \u001b[0;36m_get_tick_bboxes\u001b[0;34m(self, ticks, renderer)\u001b[0m\n\u001b[1;32m   1090\u001b[0m         \u001b[0;34m\"\"\"Return lists of bboxes for ticks' label1's and label2's.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1091\u001b[0m         return ([tick.label1.get_window_extent(renderer)\n\u001b[0;32m-> 1092\u001b[0;31m                  for tick in ticks if tick.label1.get_visible()],\n\u001b[0m\u001b[1;32m   1093\u001b[0m                 [tick.label2.get_window_extent(renderer)\n\u001b[1;32m   1094\u001b[0m                  for tick in ticks if tick.label2.get_visible()])\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/axis.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m   1090\u001b[0m         \u001b[0;34m\"\"\"Return lists of bboxes for ticks' label1's and label2's.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1091\u001b[0m         return ([tick.label1.get_window_extent(renderer)\n\u001b[0;32m-> 1092\u001b[0;31m                  for tick in ticks if tick.label1.get_visible()],\n\u001b[0m\u001b[1;32m   1093\u001b[0m                 [tick.label2.get_window_extent(renderer)\n\u001b[1;32m   1094\u001b[0m                  for tick in ticks if tick.label2.get_visible()])\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/text.py\u001b[0m in \u001b[0;36mget_window_extent\u001b[0;34m(self, renderer, dpi)\u001b[0m\n\u001b[1;32m    900\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    901\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mcbook\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_setattr_cm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdpi\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdpi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 902\u001b[0;31m             \u001b[0mbbox\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minfo\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdescent\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_layout\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_renderer\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    903\u001b[0m             \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_unitless_position\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    904\u001b[0m             \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/text.py\u001b[0m in \u001b[0;36m_get_layout\u001b[0;34m(self, renderer)\u001b[0m\n\u001b[1;32m    294\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mclean_line\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    295\u001b[0m                 w, h, d = renderer.get_text_width_height_descent(\n\u001b[0;32m--> 296\u001b[0;31m                     clean_line, self._fontproperties, ismath=ismath)\n\u001b[0m\u001b[1;32m    297\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    298\u001b[0m                 \u001b[0mw\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mh\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0md\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/backends/backend_agg.py\u001b[0m in \u001b[0;36mget_text_width_height_descent\u001b[0;34m(self, s, prop, ismath)\u001b[0m\n\u001b[1;32m    235\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    236\u001b[0m         \u001b[0mflags\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_hinting_flag\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 237\u001b[0;31m         \u001b[0mfont\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_get_agg_font\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    238\u001b[0m         \u001b[0mfont\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mflags\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mflags\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    239\u001b[0m         \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mh\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfont\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_width_height\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# width and height of unrotated string\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/backends/backend_agg.py\u001b[0m in \u001b[0;36m_get_agg_font\u001b[0;34m(self, prop)\u001b[0m\n\u001b[1;32m    270\u001b[0m         \u001b[0mGet\u001b[0m \u001b[0mthe\u001b[0m \u001b[0mfont\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mtext\u001b[0m \u001b[0minstance\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcaching\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mefficiency\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    271\u001b[0m         \"\"\"\n\u001b[0;32m--> 272\u001b[0;31m         \u001b[0mfname\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfindfont\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    273\u001b[0m         \u001b[0mfont\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_font\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    274\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/matplotlib/font_manager.py\u001b[0m in \u001b[0;36mfindfont\u001b[0;34m(self, prop, fontext, directory, fallback_to_default, rebuild_if_missing)\u001b[0m\n\u001b[1;32m   1312\u001b[0m             \u001b[0mprop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfontext\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdirectory\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfallback_to_default\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrebuild_if_missing\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1313\u001b[0m             rc_params)\n\u001b[0;32m-> 1314\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mos\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrealpath\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilename\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1315\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1316\u001b[0m     \u001b[0;34m@\u001b[0m\u001b[0mlru_cache\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/posixpath.py\u001b[0m in \u001b[0;36mrealpath\u001b[0;34m(filename)\u001b[0m\n\u001b[1;32m    393\u001b[0m symbolic links encountered in the path.\"\"\"\n\u001b[1;32m    394\u001b[0m     \u001b[0mfilename\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mos\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfspath\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilename\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 395\u001b[0;31m     \u001b[0mpath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mok\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_joinrealpath\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilename\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilename\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    396\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mabspath\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    397\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/posixpath.py\u001b[0m in \u001b[0;36m_joinrealpath\u001b[0;34m(path, rest, seen)\u001b[0m\n\u001b[1;32m    427\u001b[0m             \u001b[0;32mcontinue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    428\u001b[0m         \u001b[0mnewpath\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mjoin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 429\u001b[0;31m         \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mislink\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnewpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    430\u001b[0m             \u001b[0mpath\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnewpath\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    431\u001b[0m             \u001b[0;32mcontinue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/posixpath.py\u001b[0m in \u001b[0;36mislink\u001b[0;34m(path)\u001b[0m\n\u001b[1;32m    169\u001b[0m     \u001b[0;34m\"\"\"Test whether a path is a symbolic link\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    170\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 171\u001b[0;31m         \u001b[0mst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mos\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlstat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpath\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    172\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mOSError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mAttributeError\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    173\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# active learning\n",
    "# .7 active learning .3 replay buffer\n",
    "# w/o  wasserstein loss\n",
    "# Adam loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "# load model weights\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "print(\"Loaded model weights\")\n",
    "\n",
    "# load replay buffer params\n",
    "means = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "means = np.nan_to_num(means)\n",
    "covs = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "print(\"Loaded means and covariances for replay buffer\")\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "loss = []\n",
    "target_miou = []\n",
    "\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.1, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda2, label='active learning ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mIoU at 19 epochs: 0.27941320034173817, best mIoU: 0.34360525699762196\n",
      "ignore 0.0\n",
      "road 0.7947118816096556\n",
      "sidewalk 0.16950489943124855\n",
      "building 0.5743483505290082\n",
      "traffic light 0.14107658747600407\n",
      "traffic sign 0.008417780324936393\n",
      "vegetation 0.09634747227249862\n",
      "sky 0.8953033105324258\n",
      "person 0.17757578623261971\n",
      "rider 0.0\n",
      "car 0.8397632815376592\n",
      "bus 0.007852661616903418\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.26463585796878286\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA34AAAJcCAYAAACmOnadAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAADb30lEQVR4nOzdd3xUVfrH8c+ZSSOkQAq9I70kdBAQUFGwABZExF5Q7OvqquuubXUt62/Xde0FRcGKDQQROx3pvUPoJSGk98z5/TGTMIQktCST8n3zmtfMvffcM8+dmxnmmXPuOcZai4iIiIiIiFRfDl8HICIiIiIiIuVLiZ+IiIiIiEg1p8RPRERERESkmlPiJyIiIiIiUs0p8RMREREREanmlPiJiIiIiIhUc0r8REREREREqjklfiIiIiUwxsQZY873dRwiIiJnSomfiIiIiIhINafET0RE5BQYYwKNMS8bY/Z5bi8bYwI926KMMd8ZY5KMMYnGmLnGGIdn28PGmL3GmFRjzCZjzHm+PRIREalJ/HwdgIiISBXzGNAXiAUs8C3wN+DvwJ+BPUC0p2xfwBpj2gF3A72stfuMMS0AZ8WGLSIiNZla/ERERE7NOOBpa+0ha2088BRwnWdbLtAQaG6tzbXWzrXWWiAfCAQ6GmP8rbVx1tptPoleRERqJCV+IiIip6YRsNNreadnHcC/gK3AbGPMdmPMIwDW2q3A/cCTwCFjzKfGmEaIiIhUECV+IiIip2Yf0NxruZlnHdbaVGvtn621rYARwAMF1/JZaz+21g7w7GuBFyo2bBERqcmU+ImIiJTO3xgTVHADPgH+ZoyJNsZEAY8DkwGMMZcYY84yxhggGXcXT5cxpp0x5lzPIDBZQCbg8s3hiIhITaTET0REpHQzcSdqBbcgYCmwGlgDLAee8ZRtA/wEpAELgdettb/ivr7veSABOADUAx6tuEMQEZGazrivORcREREREZHqSi1+IiIiIiIi1ZwSPxERERERkWpOiZ+IiIiIiEg1p8RPRERERESkmvPzdQBlJSoqyrZo0cLXYYiIiIiIiPjEsmXLEqy10cVtqzaJX4sWLVi6dKmvwxAREREREfEJY8zOkrb5pKunMWaYMWaTMWarMeaREspcZYxZb4xZZ4z5uKJjFBERERERqS4qvMXPGOMEXgOGAnuAJcaYadba9V5l2uCe2La/tfaIMaZeRccpIiIiIiJSXfiixa83sNVau91amwN8CowsUuY24DVr7REAa+2hCo5RRERERESk2vDFNX6Ngd1ey3uAPkXKtAUwxswHnMCT1tpZRSsyxowHxgM0a9asXIIVERERESkrubm57Nmzh6ysLF+HIlVYUFAQTZo0wd/f/6T3qayDu/gBbYDBQBNgjjGmi7U2ybuQtfZt4G2Anj172gqOUURERETklOzZs4fQ0FBatGiBMcbX4UgVZK3l8OHD7Nmzh5YtW570fr7o6rkXaOq13MSzztseYJq1NtdauwPYjDsRFBERERGpsrKysoiMjFTSJ6fNGENkZOQptxr7IvFbArQxxrQ0xgQAVwPTipT5BndrH8aYKNxdP7dXYIwiIiIiIuVCSZ+cqdP5G6rwxM9amwfcDfwAbAA+t9auM8Y8bYwZ4Sn2A3DYGLMe+BV4yFp7uKJjFRERERERqQ58Mo+ftXamtbattba1tfZZz7rHrbXTPI+ttfYBa21Ha20Xa+2nvohTKp+DKVms3ZuMtbqkU0RERORUJSUl8frrr5+wXFxcHB9/fOKptOPi4ujcufNJrxff8UniJ3I6rLXcMmkJl/xvHv2f/4Wnpq9jSVwiLpeSQBEREZGTUdaJn1QdSvzKWXZevq9DqDaW7jzC2r0pXNWzCR0bhTNl8S5Gv7mQPs/9zN++WcP8rQnk5bt8HaaIiIhIpfXII4+wbds2YmNjeeihh7DW8tBDD9G5c2e6dOnCZ599Vlhu7ty5xMbG8p///Ie4uDgGDhxI9+7d6d69OwsWLDjp58zKyuKmm26iS5cudOvWjV9//RWAdevW0bt3b2JjY+natStbtmwhPT2diy++mJiYGDp37lwYj5y5yjqdQ7WwYGsCE6Ys54s7+tG2fqivw6nyPpgfR3iQH09e2oHgwADSsvP4ZeMhZq3dz5fL9jJ50S7qBvsztGN9hnduSP+zogjw028bpYlPzWbjgRT6tYrEz6nXSkREpCI9NX0d6/ellGmdHRuF8cSlnUrc/vzzz7N27VpWrlwJwJdffsnKlStZtWoVCQkJ9OrVi3POOYfnn3+el156ie+++w6AjIwMfvzxR4KCgtiyZQtjx45l6dKlJxXTa6+9hjGGNWvWsHHjRi644AI2b97Mm2++yX333ce4cePIyckhPz+fmTNn0qhRI2bMmAFAcnLymb0gUkiJXzlq3zCMrNx83p+/g+cu7+rrcKq0/cmZzFp3gGkNJhL8n9ug/SWEdBzFiE6DGRHTiMycfH7fHM+stfv5fs0BPl+6h9BAP87vWJ9hnRswqG00Qf5OXx+GzyWm57B4+2EWbj/Mwm2H2XIoDYB/jOzEdf1a+DY4ERERqXDz5s1j7NixOJ1O6tevz6BBg1iyZAlhYWHHlMvNzeXuu+9m5cqVOJ1ONm/efErPcc899wDQvn17mjdvzubNm+nXrx/PPvsse/bs4fLLL6dNmzZ06dKFP//5zzz88MNccsklDBw4sEyPtyZT4leOImoHcHn3xny1fC8PXdieiNoBvg6pypq8aCdN7AE6Jv4E9TvDhu9g5RQIDIf2F1Gr4yiGtR/CsM4NyM7LZ/7WBL5fc4AfNxzk6xV7CQ5wMqRdPYZ1bsC57etRO7Bm/OknZ+Yek+htPJAKQHCAk54tIri8exO+Wr6HL5btUeInIiJSwUprmats/vOf/1C/fn1WrVqFy+UiKCjojOu85ppr6NOnDzNmzOCiiy7irbfe4txzz2X58uXMnDmTv/3tb5x33nk8/vjjZXAEUjO+/fpKdip/qjOPqXmN+eSPXdw15CxfR1QlZeXm88kfu3m+3nxMqhPGfQ7BkbD9N1j3DWyaAas+gcAwaDecwI6jOLf1uZzbvj65+S4Wb0/k+7X7+WHdQWas2U+An4NBbaMZ3rkB53WoT3gtf18fYplJy85jyY5EFmxLYOH2w6zbl4K1EOjnoEfzuvx5aFvOPiuSrk3q4O/p2hng5+Af361n88FUdUkWERGp5kJDQ0lNTS1cHjhwIG+99RY33HADiYmJzJkzh3/961/s3bv3mHLJyck0adIEh8PBpEmTyM8/+XEsBg4cyJQpUzj33HPZvHkzu3btol27dmzfvp1WrVpx7733smvXLlavXk379u2JiIjg2muvpU6dOrz77rtlevw1mRK/8rRrMfXmPMoDjR7m/QW1uW1gK11zdhqmr9pHVnoK5zp+gA4jIKyRe0PbC923vBzY8bs7Cdz4Haz+DAJCod0w/DuOYsBZ5zGgTReeHtmZpXGJfL/2AD+sO8CP6w/i7zSc3TqK4Z0bcEGnBlWuVTYjJ4+lcUcKW/TW7E0m32UJcDqIbVaHe89tQ7/WkXRrVodAP6+urjnpsGMhxM3jur2r+cRxMV8u28OjF3Xw3cGIiIhIuYuMjKR///507tyZ4cOH8+KLL7Jw4UJiYmIwxvDiiy/SoEEDIiMjcTqdxMTEcOONN3LnnXdyxRVX8OGHHzJs2DBq16590s955513MmHCBLp06YKfnx8ffPABgYGBfP7553z00Uf4+/vToEED/vrXv7JkyRIeeughHA4H/v7+vPHGG+X4atQsprrMh9azZ097sheYVhhr4bU+pLj86brvUV4e041R3Rr7OqoqxVrLJf+bxwUZM7gv6w24+Qdo1rfkHfJzj00CM49AQIg7Qew4CtoMBf9auFyWVXuSmLX2AN+vPcCuxAwcBvq2imR45wZc2KkB9cLOvAtDWcvKzWf5zqOJ3qo9SeTmW/wchpimdejXKpJ+rSPp3qwutQK8E70M2L0Y4uZB3FzYuwxceeDwA2v5NfRi/pJ5AwsfOVeDvIiIiJSjDRs20KGDfmiVM1fc35IxZpm1tmdx5ZX4lbelE+G7P3FfrefYUbsr397VH2OMr6OqMpbGJXLlmwtYGfU4dUJDYPzvcLKvX36uO8kpSAIzDoN/bWh7gScJvAACgrHWsm5fiicJ3M+2+HSMgR7N6jKscwOGd2lI4zq1yvMwS5ST52Ll7iQWbjvMwu0JLN+VRE6eC4eBLo3D6ds6kn6tIunVIuLY6xZzs2DPEvfx75gLe5dCfg4YJzTqBi0HQosB0LQvTL+P3E0/0in1Fd668WyGtK/nk2MVERGpCZT4SVlR4lfZ5GTAfzqyM6wHg3bezBd39KNXiwhfR1Vl3PXxcrI3/8q7PA0jX4du406vovw82DnPnQRumA4ZCeAf7E7+Oo50twgGuLssbDmYyveelsAN+91DLMc0CWdY54YM79yAFlEn37XhVOXmu1i9J5lFnha9pTsTycp1YQx0bBhW2KLXq2UEYUFe1ybmZbtb8XbMdSd7u/+A/GwwDmgY407yWpzjbi0NOnaULrb+BJOv4CHHn8lofQmvjetebscnIiJS0ynxk7JyqomfrvErbwHB0OMmms1/mY5Bo5g4b4cSv5O0PzmTWWsPMLPeb5ATCZ2vOP3KnH7QarD7dtFLsGvB0SRw/TfgV8vdDbTTKNq0uZA257Xh3vPaEJeQzvdrDzBr7X5emLWRF2ZtpH2DUIZ3bsjwLg1oUy/kjFpw812WdfuSWbjtMAu2HWZpXCLpOe6Lpds3COXqXs3o1zqSPi0jqBPsdf1hfi7sWgxxc9zJ3u4/IC8TMNCgM/S61d2q16wf1KpTehCthkBoQ24zC7lkfS+SM3IJD64+A96IiIiIiBK/itHrVsyCV3iy/gKuXhfB7sQMmkYE+zqqSm/Kol00tIdomzwP+t8P/mV0zZ3TD1qe475d9C/YtdCTBE5z3/yC4KzzodNltGh7IRMGt2bC4NbsTXInorPW7uflnzfzn5820yq6Nhd1bsiwzg3o1CjshEmgy2XZcCCFhdsOs2j7YRbvSCQ1Kw+A1tG1uax7Y/q1iqJvqwgiQwKP7pifB3uWHu26uWsR5Ka7t9XrBD1ugBYDofnZEHyKPyw4nNB1DG0W/I/w/ESmrd7HdX2bn1odIiIiIlKpKfGrCOGNoeMoem6eTog5n0kL4vjbJR19HVWl5p7CYRfP1puPSTHQ65byeSKH09MNcgAMf8E9AEpBErjxO3AGepLAUTRuO4xbBrTklgEtOZSSxQ/r3N1B3/h9G6/+upWmEbUY7kkCY5vUweEwWGvZfDCNhZ7pFRbvSCQpIxeAFpHBXNK1IX1bua/TO2YwGVc+7FtxtOvmzoWQ4xlSObo9xF7jbtFr3h9qR5356xB7DWb+y9xWZylTl7VQ4iciIiJSzSjxqyh978SxdiqPN1nBU0tCuH9oW0JqyCTip+O71ftJT0/lPOcP0OESCG9S/k/qcLpbzJqfDcOedyeB67913zbNAGcAtD4POo2iXrvhXNevBdf1a0Fieg4/rncnge/P38Hbc7bTICyIzo3DWLEricPpOQA0qVuLoR3q0691JH1bRdLIe8AYlwv2rz7aordzAWQnu7dFtoGuoz0J6kAIKYfBV6LbQeMeXJk0h3/uPo+th1I5q57m9BMRERGpLpR5VJQmPaBpHy5NmsZfsvswdelubuzf0tdRVUrWWiYtiGN8naX4ZyVD79srPgiHA5r3c98u/Kd7hMyCJHDz9+Dwh9bnQqdRRLS7iDG9mjGmVzOSM3P5ZeNBZq45wJaDqZzTNrpwQJZjuve6XHBwnVeL3nz31BMAEa2g00j3YCwtBkBYw4o55thriJjxZ7o6dzJ12V4eGd6+Yp5XREREqpTBgwfz0ksv0bNnsWOInLQWLVqwdOlSoqJOvvfSQw89xMyZM7nooov4y1/+wiWXXEJOTg6vvPIKAwcOPKN4TlVZvQ4VRYlfRep7J4Ff3MBt9Tfx/oIQru/XAodDUzsUtXzXEdbsTeLDqNlQv4u7Bc6XHA5o1sd9u+AZ9+iZ679xJ4FbfnAnga0GQ6dRhLe7iMu6NeGybkVaKK2FQxvdSV7cXPd8ehmH3dvqNId2Fx+dYqEiWjeL0+lymPUo94Yt4bEV7XjownY49fcpIiJSbVlrsdbicFSdOXzffvttEhMTcTqdfPrpp3Tp0oV33333pPfPz8/H6XSeuGA1VHXOcnXQ/hIIb8pt/j+w83AGP2885OuIKqX358cxJGgLddO2QJ/xJz9vX0VwOKBpL7jwWbh/Ddz6C/SdAAmb4Nu74KU2MPkKWP4hHFjjnsfxi5vgpbbweh+Y+SDsWeaeRmLk63Dfarh/NYx6DWKu9l3SB+5BYdpdxDlZv5KYks7cLfG+i0VERETKRVxcHO3ateP666+nc+fO7N69m3/961/06tWLrl278sQTTxSWa9++PePGjaNDhw5ceeWVZGRkHFffhAkT6NmzJ506dSrc95dffmHUqFGFZX788Ucuu+yyYuN58cUX6dKlC71792br1q0A3HjjjUydOrWwTEhICAAjRowgLS2NHj168MILL/CXv/yFb7/9ltjYWDIzM5k9ezb9+vWje/fujB49mrS0NMDdsvjwww/TvXt3vvjii2Oe/+DBg1x22WXExMQQExPDggULAJg8eTK9e/cmNjaW22+/nfz8/FJf108++YQuXbrQuXNnHn74YcCdZN5444107tyZLl268J///AeAV155hY4dO9K1a1euvvrqUustS2rxq0hOP+g9nqgf/86g0P1MnLeDoR3r+zqqSuVAchaz1h5ger3fILsudBnt65BKZoy7C2+THjD0afdgLOu/cQ8OM+2eo+VCG7pbBAta9Oq2rFzJrLfYcQSs/4ZLaq1h6rKmDG6nydxFRETKzfePuH8oLksNusDw50stsmXLFiZNmkTfvn2ZPXs2W7Zs4Y8//sBay4gRI5gzZw7NmjVj06ZNvPfee/Tv35+bb76Z119/nQcffPCYup599lkiIiLIz8/nvPPOY/Xq1QwZMoQ777yT+Ph4oqOjef/997n55puLjSU8PJw1a9bw4Ycfcv/99/Pdd9+VGPe0adMICQlh5cqVANSvX5+lS5fy6quvkpCQwDPPPMNPP/1E7dq1eeGFF/j3v//N448/DkBkZCTLly8/rs57772XQYMG8fXXX5Ofn09aWhobNmzgs88+Y/78+fj7+3PnnXcyZcoUrr/++mLj2rdvHw8//DDLli2jbt26XHDBBXzzzTc0bdqUvXv3snbtWgCSkpIAeP7559mxYweBgYGF6yqCWvwqWvfrwb82f438jYXbD7N+X4qvI6pUpizeSX0bT/vkOdD9BvCvdeKdKgNjoHF3dwJ43yoY/zuMehPuXgYPbIAr3nGf+4hWlTfpA/d1iyH1uT1sEbPXHyTZMwKpiIiIVB/Nmzenb9++AMyePZvZs2fTrVs3unfvzsaNG9myZQsATZs2pX///gBce+21zJs377i6Pv/8c7p37063bt1Yt24d69evxxjDddddx+TJk0lKSmLhwoUMHz682FjGjh1beL9w4cLTPqZFixaxfv16+vfvT2xsLJMmTWLnzp2F28eMGVPsfr/88gsTJkwAwOl0Eh4ezs8//8yyZcvo1asXsbGx/Pzzz2zfvr3E516yZAmDBw8mOjoaPz8/xo0bx5w5c2jVqhXbt2/nnnvuYdasWYSFhQHQtWtXxo0bx+TJk/Hzq7h2OLX4VbRadaDbONou+4Cm/pcwcf4OXhod4+uoKoXsvHw+XryLf9RbgEmh/KZwKG/GQKNY962qcfpB16tot/ANQvLG8d2afYzro6kdREREysUJWubKS+3atQsfW2t59NFHuf32YwfTi4uLO25+4qLLO3bs4KWXXmLJkiXUrVuXG2+8kaysLABuuukmLr30UoKCghg9enSJCY53nQWP/fz8cLlcALhcLnJyck54TNZahg4dyieffHLCYz6Zum644Qaee+65k96nOHXr1mXVqlX88MMPvPnmm3z++edMnDiRGTNmMGfOHKZPn86zzz7LmjVrKiQBVIufL/S5A5Ofy1ONFjNt5T7iU7N9HVGl8N2q/aSlpzE0cxa0uwjqNPN1SDVTzDUYm8dtdZYxddkeX0cjIiIi5ejCCy9k4sSJhdfD7d27l0OH3ONQ7Nq1q7AV7uOPP2bAgAHH7JuSkkLt2rUJDw/n4MGDfP/994XbGjVqRKNGjXjmmWe46aabSnz+zz77rPC+X79+gPuavGXLlgHu7p25uSfugdS3b1/mz59feJ1geno6mzdvPuF+5513Hm+88QbgviYvOTmZ8847j6lTpxa+DomJice0HhbVu3dvfv/9dxISEsjPz+eTTz5h0KBBJCQk4HK5uOKKK3jmmWdYvnw5LpeL3bt3M2TIEF544QWSk5MLX/vyphY/X4hsDW2HMWj3t5j8c5iyeCf3n9/W11H5lLWWDxbEcVudZfhnJUEfH0zhIG71O0LDWEanzuGFXUPYeiiNs+qF+DoqERERKQcXXHABGzZsKEy6QkJCmDx5Mk6nk3bt2vHaa69x880307Fjx8IukQViYmLo1q0b7du3P6ZbaIFx48YRHx9Phw4dSnz+I0eO0LVrVwIDAwtb62677TZGjhxJTEwMw4YNO6nWuujoaD744APGjh1Ldra7UeWZZ56hbdvSv2P/97//Zfz48bz33ns4nU7eeOMN+vXrxzPPPMMFF1yAy+XC39+f1157jebNi+8F1bBhQ55//nmGDBmCtZaLL76YkSNHsmrVKm666abC1svnnnuO/Px8rr32WpKTk7HWcu+991KnTp0THl9ZMNbaCnmi8tazZ0+7dOlSX4dx8rb/Dh+OYGLkg7ye3Jf5j5xLoF/NHFoWYNnOI1zxxnyWR/+DiFoOmLCgcl8LV90tfgu+/wsX5z7POQOH8PAwzeknIiJSFjZs2FBqIlRZxMXFcckllxQOTHI67r77brp168Ytt1TRy3cqueL+lowxy6y1xU4sqK6evtLyHKjfmTH500lIy2bayn2+jsinJi2I45ygbUSkboTelWwKh5qo85Xg8OfeyCV8vXwv+a7q8QORiIiIVIwePXqwevVqrr32Wl+HIh5K/HzFGOg7gdpJm7gqcgcT58dRXVpfT9XBlCxmrtnPI3V/g6Bw6HqVr0OS2pHQ9kIGZ/9GQkoa87cm+DoiERERqUAtWrQ4o9a+ZcuWMWfOHAIDA8swKjkTSvx8qfOVUDua+2r/xIb9KSzanujriHxiyqKdRNsEOiT97p7yIODkR12SchQ7jsDsw1wUtE6DvIiIiIhUcUr8fMk/CHreQqNDvxMTfJj35u3wdUQVLjsvn4//2MVj9RZirAt63errkKRAm6EQHMXtdRbzw7oDJGdqTj8RERGRqkqJn6/1vBnj9OfJenP4eeNB4hLSfR1RhZqxej+paWlcmPU9tBsOdVv4OiQp4PSHrlfRIWU+tfKSmbF6v68jEhEREZHTpMTP10LrQ+criTk8g7qODD5YEOfriCpMwRQON9dZgX/2EU3hUBnFjMXhyuWWOsv5crm6e4qIiIhUVUr8KoO+E3DkZvBk42V8sXQ3KVk1o0vdit1JrN6TxK0BP0J0e2g5yNchSVENu0L9Lozxn8uynUfYHl8xE4yKiIhI5fDbb7+xYMGCwuU333yTDz/88IzrffLJJ3nppZfOuJ4TOfvss8v9OU6koo71RHyS+BljhhljNhljthpjHilm+43GmHhjzErPrXpf+NWwK7QYyLCMaWTl5PD5kt2+jqhCTFoQx4DA7USmrIfet2kKh8oq9hrqpa6nnWOPWv1ERERqmKKJ3x133MH111/vw4iOlZeXV+p279hrugpP/IwxTuA1YDjQERhrjOlYTNHPrLWxntu7FRqkL/SdQEDaXu5qsIH358eRl+/ydUTl6lBKFjNW7+eRiDkQGA5dr/Z1SFKSLqPB4cd9UUv5SnP6iYiIVHmjRo2iR48edOrUibfffrtw/axZs+jevTsxMTGcd955xMXF8eabb/Kf//yH2NhY5s6dW9h6tXHjRnr37l24b1xcHF26dAHcUzkMGjSIHj16cOGFF7J/f+njBGzbto1hw4bRo0cPBg4cyMaNGwGYPn06ffr0oVu3bpx//vkcPHgQcLegXXfddfTv35/rrruOJ598kptvvpnBgwfTqlUrXnnllcK6Q0JCAHcCO3jwYK688krat2/PuHHjCqdSmzlzJu3bt6dHjx7ce++9XHLJJcfFmJ+fz4MPPkjnzp3p2rUr//vf/07rWFeuXEnfvn3p2rUrl112GUeOHAHglVdeoWPHjnTt2pWrr3Z/L/7999+JjY0lNjaWbt26kZqaWmrdJ+J3Rnufnt7AVmvtdgBjzKfASGC9D2KpPNoOg7otucnxPa8kdeanDQcZ1rmhr6MqN1MW7yLSJtIp+VfofTsEhvg6JClJSDS0uYBzd/7KPcmXsnDbYQa0ifJ1VCIiIlXeC3+8wMbEjWVaZ/uI9jzc++FSy0ycOJGIiAgyMzPp1asXV1xxBS6Xi9tuu405c+bQsmVLEhMTiYiI4I477iAkJIQHH3wQgJ9//tn9PO3bk5OTw44dO2jZsiWfffYZY8aMITc3l3vuuYdvv/2W6OhoPvvsMx577DEmTpxYYjzjx4/nzTffpE2bNixevJg777yTX375hQEDBrBo0SKMMbz77ru8+OKL/N///R8A69evZ968edSqVYsnn3ySjRs38uuvv5Kamkq7du2YMGEC/v7+xzzPihUrWLduHY0aNaJ///7Mnz+fnj17cvvttxce99ixY4uN8e233yYuLo6VK1fi5+dHYmLiaR3r9ddfz//+9z8GDRrE448/zlNPPcXLL7/M888/z44dOwgMDCQpKQmAl156iddee43+/fuTlpZGUFBQqef1RHyR+DUGvPsy7gH6FFPuCmPMOcBm4E/W2uP6PxpjxgPjAZo1a1YOoVYghxP63EHdWQ8zNHw3782rW20Tv+y8fKYs3sXf6y3CJOdD7+rdk7daiBlL0KaZXBi0nqnLmirxExERqcJeeeUVvv76awB2797Nli1biI+P55xzzqFly5YAREREnLCeq666is8++4xHHnmEzz77jM8++4xNmzaxdu1ahg4dCrhbyho2LPk7bVpaGgsWLGD06NGF67KzswHYs2cPY8aMYf/+/eTk5BTGBjBixAhq1apVuHzxxRcTGBhIYGAg9erV4+DBgzRp0uSY5+rdu3fhutjYWOLi4ggJCaFVq1aFdY8dO/aYVtACP/30E3fccQd+fn6Fr8/atWtP6ViTk5NJSkpi0CD3uBY33HBD4XF37dqVcePGMWrUKEaNGgVA//79eeCBBxg3bhyXX375ccdzqnyR+J2M6cAn1tpsY8ztwCTg3KKFrLVvA28D9OzZs+r3P+s2Dn59lofr/Mr5cU1ZvSeJrk3q+DqqMjdzzX5S0tIY7jcT2lwAEa18HZKcSNthUCuCCYGLGb0uhpSsXMKC/E+8n4iIiJToRC1z5eG3337jp59+YuHChQQHBzN48GCysrJOq64xY8YwevRoLr/8cowxtGnThjVr1tCpUycWLlx4UnW4XC7q1KnDypUrj9t2zz338MADDzBixAh+++03nnzyycJttWvXPqZsYGBg4WOn01nstX8nU+ZUWGtP6VhLM2PGDObMmcP06dN59tlnWbNmDY888ggXX3wxM2fOpH///vzwww+0b9/+tJ/DF4O77AWaei038awrZK09bK3N9iy+C/SooNh8KzAUul9P6/ifaB2YzPvz43wdUbn4YMFObqyzkoCsw5rCoarwC4AuV9IpdT4BuSnM1Jx+IiIiVVJycjJ169YlODiYjRs3smjRIgD69u3LnDlz2LFjBwCJiYkAhIaGlnhtWevWrXE6nfzjH/9gzJgxALRr1474+PjCZCg3N5d169aVGE9YWBgtW7bkiy++ANzJ1KpVqwpjbdy4MQCTJk0600MvVrt27di+fTtxcXEAfPbZZ8WWGzp0KG+99VZhspiYmHjKxxoeHk7dunWZO3cuAB999BGDBg3C5XKxe/duhgwZwgsvvEBycjJpaWls27aNLl268PDDD9OrV6/Cax9Ply8SvyVAG2NMS2NMAHA1MM27gDHGu410BLChAuPzrd7jMdbFkw0XMn3VPg6mnN4vMJXVil1HWLU7idsCf4TINtBqiK9DkpMVew2O/GxuCl/O1GUa3VNERKQqGjZsGHl5eXTo0IFHHnmEvn37AhAdHc3bb7/N5ZdfTkxMTGEid+mll/L1118XDu5S1JgxY5g8eTJXXXUVAAEBAUydOpWHH36YmJgYYmNjTziy5pQpU3jvvfeIiYmhU6dOfPvtt4B7EJfRo0fTo0cPoqLK5zKTWrVq8frrrxcOLhMaGkp4ePhx5W699VaaNWtG165diYmJ4eOPPz6tY500aRIPPfQQXbt2ZeXKlTz++OPk5+dz7bXX0qVLF7p168a9995LnTp1ePnllwsHk/H392f48OFndKymYDSbimSMuQh4GXACE621zxpjngaWWmunGWOew53w5QGJwARrbakpbs+ePe3SpUvLOfIK8tl15O+YS+fkf3PL4E48eGE7X0dUZu7/dAUHN8znE/MYDP8X9Bnv65DkZFkLr/fjQLYffQ89ym8PDqZFVO0T7yciIiKFNmzYQIcOHXwdhnhJS0sjJCQEay133XUXbdq04U9/+pOvwzqh4v6WjDHLrLU9iyvvk3n8rLUzrbVtrbWtrbXPetY9bq2d5nn8qLW2k7U2xlo75ERJX7XT906cWUd4rPEqpizeSVZuvq8jKhOHUrOYsWY/j0b+DgGhEFv8qElSSRkDsdfQIGUNrR37NKefiIiIVAvvvPMOsbGxdOrUieTkZG6/vXpeiuSTxE9OoFlfaBjLFbnTScrI5psVe0+8TxXw8eJd1Mk/QpekX90D2QSG+jokOVVdrwLj4P6opXy5bA8uzeknIiIiVdyf/vQnVq5cyfr165kyZQrBwcG+DqlcKPGrjIyBfndRK3kb10ZtY+L8HfiiS25ZyslzMXnRLh6rvwjjyoVet/k6JDkdoQ3grPM5P+dXDiRnsHD7YV9HJCIiUuVU9e914nun8zekxK+y6jgKQhpwV63ZbD6YxrytCb6O6IzMXLOf5LR0hmd/D2cNhaizfB2SnK6YsdTKOsj5QRv5UoO8iIiInJKgoCAOHz6s5E9Om7WWw4cPn/KE7pV1Hj/xC4Det9Lgl2foVXs0E+ftYGCbaF9Hddo+WBDHDeGrCMyK1xQOVV27iyAonDuDFnP12s48NbIToZrTT0RE5KQ0adKEPXv2EB8f7+tQpAoLCgo65QndlfhVZj1uhjkv8XjU71y6qR5bD6VxVr0QX0d1ylbuTmLl7iTerv8j1G4Nrc/zdUhyJvyDoPMVdF3xMf65Y/h+zQGu6tX0xPuJiIgI/v7+tGzZ0tdhSA2krp6VWe1I6DqGzgnfU88vnQ8W7PB1RKdl0oI4+gTupF7yauh9Gzj0Z1flxY7DkZ/FDeErNKefiIiISBWgb+CVXd87MXlZPNXoD75ctpekjBxfR3RKDqVm8d3qfTwaOQcCQiD2Gl+HJGWhcQ+Iass1gfP5Iy6RnYfTfR2RiIiIiJRCiV9lV689tD6X89OmkZebzadLdvs6olPyyeLdhOUnEZP0M8SMhaBwX4ckZcEYiBlLo5SVtDAH+HJ59ZhyRERERKS6UuJXFfS9E/+Mg9zfcB2TFsSRm+/ydUQnJSfPxeTFO/lr/cUYVw70Hu/rkKQsxVztntMvWnP6iYiIiFR2SvyqgtbnQVRbrmMG+5MzmbX2gK8jOinfr93PkdR0Lsn5HlqfC9FtfR2SlKWwRtBqMBfk/sq+pHQW7dCcfiIiIiKVlRK/qsDhgD53EHZkLZfU2cXE+VVjkJcPFsRxXfgaAjMPQm9N4VAtxY4jOHM/5wZu1iAvIiIiIpWYEr+qIuZqCKrDg2E/s2JXEst3HfF1RKVatTuJFbuSuKPWj1C3BbQZ6uuQpDy0vxgCw7gr4g9mrT1AenaeryMSERERkWIo8asqAmpDz5toHv8L7YISmTivcrf6TVoQR4+AXdRPWum+ts/h9HVIUh78a0Gny4hJ/R2Tk8bMNft9HZGIiIiIFEOJX1XS6zaMcfBUgwV8v/YA+5IyfR1RseJTs5m+eh+PRc0F/2CIHefrkKQ8xV6DMy+T68JXqbuniIiISCWlxK8qCW8MHUfRO3E6wTaDSQvjfB1RsT75Yxch+cnEJv/k7qJaq46vQ5Ly1LQPRLTiuqD5LN6RyK7DGb6OSERERESKUOJX1fS9E0dOKn9vsoJPFu8iI6dyXVOVk+di8qKdPFr/Dxz52ZrCoSYwBmKuoXHyMpqaQ3y5XK1+IiIiIpWNEr+qpkkPaNKbEVnTScvKqXQTZ89ad4DDqRmMyP0eWg6Ceh18HZJUhJirAcP90cv4aoXm9BMRERGpbJT4VUX97iQodSe3RG/i/fk7KtWX7A/m72Bc+BqCMvZDH03hUGPUaQotBzIs71f2JKbzR1yiryMSERERES9K/Kqi9pdCWBPGB85me3w6v2+O93VEAKzek8TyXUlMqPUz1GkGbYf5OiSpSLHjqJ2xh3MCt2qQFxEREZFKRolfVeT0gz7jiU5YzICQfZVmQvcPFsTRLWAPDZOWQa/bNIVDTdPhUggI4e66S5i5Zr/m9BMRERGpRJT4VVXdrwf/2jwW+TtztySw+WCqT8NJSMvmu1X7eSx6LvjVgm7X+jQe8YGA2tBxFN3TfsPmpDNr7QFfRyQiIiIiHkr8qqpadSH2Gton/EAT/xSfT+j+yeJdBOcn0z3pR+h6FQRH+DQe8ZHYa3DmpTMufLW6e4qIiIhUIkr8qrK+EzD5OTzVaDFfrdjL4bRsn4SRm+9i8uKdPFJ/CY78LA3qUpM16wd1mnNDrQUs3H6Y3Yma009ERESkMlDiV5VFtoa2wxiUMh2Tl8Unf+zySRiz1h4gISWDUbkzocVAqN/JJ3FIJeBwQOw1NElaQmOTwFeVbLoRERERkZpKiV9V13cCfpkJPNhoDR8u3ElOnqvCQ/hgQRxjw9cRlLFPE7YLxFyNwXJv9DK+XL4HayvPdCMiIiIiNZUSv6qu5SCo14mxrhkcSs1ixpp9Ffr0a/Yks2znEe4M/hnCmkC7iyr0+aUSqtsCmg/govzf2JWYzpK4I76OSERERKTGU+JX1RkDfScQkrSRKyJ28N68HRXawvLBgji6Buyl0ZEl0PtW91QTIrFjCU3fSf+AbUxdttvX0YiIiIjUeEr8qoMuoyE4ij+F/MjavSks3VkxLSwJadlMX7WPv0XPA78g6H5DhTyvVAEdR4J/MPdGLmHG6v1k5GhOPxERERFfUuJXHfgHQa9baHxoDl1rJfDe3IqZ2uHTP3YRlJ9Cz+TZ0OVKTeEgRwWGQocR9Ej7lbycTM3pJyIiIuJjSvyqi563YBx+PFFvLrPXHyj3YfRz8118tGgnj9RfiiMvE3prCgcpIvYa/HLTuCZsNV8u15x+IiIiIr7kk8TPGDPMGLPJGLPVGPNIKeWuMMZYY0zPioyvSgqtD12upFviDMJNBpMWxJXr0/2w7gDxKZlcljcTmp0NDbuW6/NJFdRiIIQ35YbghSzYdpg9RzSnn4iIiIivVHjiZ4xxAq8Bw4GOwFhjTMdiyoUC9wGLKzbCKqzvBBy5GTzeeBmfLdlNWnb5XVf1wfw4xoStp1b6HuijKRykGA4HxFxN8+TF1LOJfK05/URERER8xhctfr2Brdba7dbaHOBTYGQx5f4BvABkVWRwVVrDGGg+gIszppGRnc0XS8tnNMW1e5NZuvMId4f8DGGNof0l5fI8Ug3EjMVYF/dFL2eq5vQTERER8RlfJH6NAe+MZI9nXSFjTHegqbV2RmkVGWPGG2OWGmOWxsfHl32kVVG/OwlI38sd9TfwwYI48l1l/0X7gwVxdPbfT+PExdDzZnD6l/lzSDUR2Rqa9uUS+xs7D6dX2IizIiIiInKsSje4izHGAfwb+POJylpr37bW9rTW9oyOji7/4KqCtsOgbgtu8ZvFzsMZ/LLxUJlWfzgtm2mr9vH3enPBGQg9bizT+qUair2GsLTt9AmI48tlGuRFRERExBd8kfjtBZp6LTfxrCsQCnQGfjPGxAF9gWka4OUkOZzQ5w4iDi/nvLA9vDdve5lW/+mS3QTmpdEreTZ0vgJqR5Vp/VINdRoFfkHcF7mE71bvJzMn39cRiYiIiNQ4vkj8lgBtjDEtjTEBwNXAtIKN1tpka22UtbaFtbYFsAgYYa1d6oNYq6Zu10JgGI/U/ZVF2xNZty+5TKrNzXfx0cKd/KX+Uhx5GRrURU5OUDh0uJTe6b+Qm53BD+s0p5+IiIhIRavwxM9amwfcDfwAbAA+t9auM8Y8bYwZUdHxVEuBodDtOs6K/5EWAcm8Pz+uTKqdve4gB1MyuCJ/JjTtA426lUm9UgPEjMUvJ4WrwtYyVd09RURERCqcT67xs9bOtNa2tda2ttY+61n3uLV2WjFlB6u17zT0GY+xLp5uuIBpK/dxKPXMB0f9YMEORodvIDhtF/RWa5+cglaDIbQRN9VeyPxtCexLyvR1RCIiIiI1SqUb3EXKSN0W0P5i+id9hyM/kymLdp1RdWv3JrMk7gj31P4VQhtCx+Jm4BApgcMJMWNombSIKHuEr1doTj8RERGRiqTErzrreyfO7CP8tfEqpizeSVbu6Q+qMWlBHB38D9I0cYGmcJDTE3MNxuZzb/QKpi7TnH4iIiIiFUmJX3XWrB80jOXK3OkcTsti+qp9p1VNYnoO367ax+P15oEzQFM4yOmJbguNe3Ipv7MjIY3luzSnn4iIiEhFUeJXnRkDfe8kOGUb10Ru4715O06rleWTP3YRkJdGn5RZ0OlyCKlXDsFKjRB7DXVSt9DDf5cGeRERERGpQEr8qrtOl0FIA+4Ons3GA6ks3H74lHbPy3cxedFOHmqwHEduuqZwkDPT+XJwBnJ/9BK+W7X/jLofy5lLysjxdQgiIiJSQZT4VXd+AdD7VhrGz6dn8EEmzos7pd1nrz/IgeQMrsybCY17QuMe5ROn1Ay16kL7i+ib/itZ2Vma08+H/m/2Jno88xOLTvHHIBEREamalPjVBD1uAr8gHo+ew88bDxKXkH7Su34wP47LwzZROy0O+txefjFKzRFzDf7ZR7gydJ26e/rId6v38b9ftmKt5fFv15Kb7/J1SCIiIlLOlPjVBLWjoOtVdDn8PVGOND5YEHdSu63bl8wfcYncF/orhNSHjqPKNUypIVqfCyH1uSVkIfO2JrA/WXP6VaR1+5J58ItV9Gxel1ev6c7mg2lMOsnPBBEREam6lPjVFH3vxORl8WTjJXy+dDfJmbkn3GXSgjja+x+i2eF5nlbDgAoIVKo9px90HUPr5AVE2GTN6VeBEtKyGf/hMuoGB/DWiPoMP/whF7QJ4+WftnAwJcvX4YmIiEg5UuJXU9TrAK2GcEHaNHJzsvli6e5Siyem5/Dtyn38vf58cPhBz5sqKFCpEWKvwbjyuCd6peb0qyA5eS7unLychLRs3r26A5HfXof57Z+82PAXcvJd/HPmBl+HKCIiIuVIiV9N0u8u/DMOck+Ddbw/P468Uq7r+XTJLpx5GfRNmeUeGTS0QQUGKtVevQ7QqBsjze9sj09nxe4kX0dU7T01fR1/xCXy4hWd6bT4IYjfAI17UGfZazzcO5BvV+5j4TYN9CIiIlJdKfGrSVqfB5FtuMHMYG9SBj+uP1hssbx8F5MX7uTB+stx5qRCbw3qIuUg5hrqpmwkVnP6lbvJi3YyZfEu7hjUmpHJU2Djd3DBszBmCjj9uTH5dZrUCeKJaRroRUREpLpS4leTOBzQ9w7Cj6zlovBdTJy/o9hiP64/yL7kTK5yzYRG3aBJzwoOVGqELleCw58/RS9j+qp9mtOvnCzafpgnp63j3Pb1eKjZJvjtOYi5BvpOgLCGMPhRnNt+5NUeBzTQi4iISDWmxK+miRkLQXV4KPxnlsQdYfWepOOKvL8gjlGhmwlJ3Q597gBjKj5Oqf6CI6DdMPql/0xmVlaJLdBy+vYcyeDOKctpFhnMK+cG4Pxmgns+zkv+c/R93ed2qNeRmLXPMaxtGP/5cbMGehEREamGlPjVNAG1oceNtIj/hbaBiUycd2yr3/p9KfyxI5E/hf8KtaPd1/eJlJfYcQRkJ3J5yHp19yxjGTl53PbhMnLzXUwc3YqQr66FoDC4egr4Bx0t6PSHi17CJO/m+fo/keuyGuhFRESkGlLiVxP1vg2D4akG8/lu9f5jft2ftCCOs/zjaZYwF3rcCH6BvotTqr+zzofgKG4NW8zcLfEcSFZLU1mw1vLQF6vZdCCFV8d0psUvd0LqQfc1fcUN1NSiP3QdQ53lr/NIb38N9CIiIlINKfGricKbQKdR9Dkyg1o2gw8XxgFwJD2Hb1bu5Yn6CzAOJ/S82bdxSvXn9IeuY2iTNI8wm6o5/crIa79uZcaa/Tw8rD2DdrwMcXNhxCvQpEfJOw39B/gFcUPSazSpE8Tj32qgFxERkepEiV9N1fdOHDkpPNZ4BR8v3kVmTj6fLtmNIy+Ds1NmQYcRENbI11FKTRA7FuPK5e7olXy5XHP6nakf1x/kpdmbuaxbY8aHzIM/3oZ+d0PM1aXvGFofhvwV5/ZfeK37PrYc0kAvIiIi1YkSv5qqSU9o0otR2dNJyshm6vI9TF60kz/XX4kzJ9k94INIRWjQBRp04XLzO1sPpbFqT7KvI6qyNh9M5f5PV9C1STjP90rHzPgztD4Xhj59chX0ug3qdSJm3QsMaxuqgV5ERESqESV+NVnfOwlK3cmNUZv454wN7E3KYIz9Hhp0haZ9fB2d1CQx1xCRsp4u/nuZumy3r6OpkpIycrjtw6XUCvDjnVENCPzyBqjTDK6cCA7nyVXi9IOLX4Lk3bxQ70dyXZZnZ2igFxERkepAiV9N1mEEhDVhQtAPZObmc2noVkJTtmgKB6l4XUaDw4/7o5cybaXm9DtVefku7v54BfuTsnh7bEfqz7gZcrNg7CdQq+6pVdb8bIgZS/jyN3i0tx/TVmmgFxERkepAiV9N5vSDPuOpl/AHIxsm8nDE7xAcCZ2v8HVkUtOEREObCxmY8QvpWdn8tEFz+p2K577fyLytCTwzshPdV/wN9q+GK9+D6HanV+HQp8G/FtcnvUbTuhroRUREpDpQ4lfTdb8e/IP5b92pNDn0G3S/4dg5vkQqSuxYArLiGRmykS81p99Jm7psD+/N28GNZ7fgqpwvYe2XcN7j0PbC0680pB6c+zec23/ltW572XIojQ/mx5VZzCIiIlLxlPjVdLXqQuw1sP03wECvW3wdkdRUbS6EWhGMD1/M75vjOaRBRU5o+a4j/PWrNZzdOpK/t90FPz3lbrEf8Kczr7znLVC/C13XPs/wtqG8/JMGehEREanKlPgJ9Jngvu9wiXuOPxFf8AuALqNpd2QOITZNc/qdwMGULO74aBkNwoN448IQnF+Ph4ZdYcSrZXONrtMPLv4/SNnL89E/aKAXERGRKk6Jn0DUWTBuKgx7wdeRSE0Xew3GlcNdUauYukxz+pUkKzef8R8tIy07j/fGnEX4N9eDXyCMmQIBwWX3RM36QOy1hK94i7/2cjJt1T4WbEsou/pFRESkwijxE7c2QyGsoa+jkJquYQzU68gVzrlsOZTGmr2a068oay1//WoNq3Yn8e/RXWgz5z5I2gVXfQR1mpb9E57/JAQEc92RV2laN4gnvl2ngV5ERESqoDNO/IwxtY0xDs/jtsaYEcYY/zMPTURqHGMg9hqiklfTwW8/UzXIy3Hem7eDr1bs5U/nt2XY/jdg28/uufea9yufJwyJhnP/jjPud16P3aWBXkRERKqosmjxmwMEGWMaA7OB64APyqBeEamJulwFxskD9Zbx7cp9ZOdpTr8Cv2+O558zNzC8cwPuiVoKC/4HvW6DHjeW7xP3vBkadKXL2he4uG0IL/+0mQPJGuhFRESkKimLxM9YazOAy4HXrbWjgU5lUK+I1ESh9eGs8zkn82dSM7P5ecMhX0dUKexISOeej5fTtn4o/+6fj2P6fdBiIAx7rvyf3OGEi/8Nqft5LmqWe6CXmRroRUREpCopk8TPGNMPGAfM8KxznmCHYcaYTcaYrcaYR4rZfocxZo0xZqUxZp4xpmMZxCkiVUXsWAIzD3JJyCZ19wRSs3K57cOlOB2G965oSq2vrncnyKMngbOCetY37QXdriNs5dv8rbdhugZ6ERERqVLKIvG7H3gU+Npau84Y0wr4taTCxhgn8BowHOgIjC0msfvYWtvFWhsLvAj8uwziFJGqou1wCKrDHQVz+qXW3G6F+S7L/Z+uZEdCOm9c3ZnGP9wGWSlw9SdQO7Jigzn/SQgIYdxh90Avj2ugFxERkSrjjBM/a+3v1toR1toXPIO8JFhr7y1ll97AVmvtdmttDvApMLJInSlei7UBjekuUpP4B0HnK2ifNIdgVzrfrtjn64h85t8/buLnjYd44pIO9F3/DOxZApe9AQ06V3wwtaPgvMdx7pzLG7FxbD2Uxvvzd1R8HCIiInLKymJUz4+NMWHGmNrAWmC9MeahUnZpDOz2Wt7jWVe03ruMMdtwt/gVm0gaY8YbY5YaY5bGx8ef/kGISOUTOw5HfhZ3RK2usXP6TV+1j9d+3cbY3k25znwPK6fAoIeh48gT71xeetwIDWPpvOZFLm4Xwn9/2qKBXkRERKqAsujq2dHTQjcK+B5oiXtkzzNirX3NWtsaeBj4Wwll3rbW9rTW9oyOjj7TpxSRyqRxd4hqy2i/uWw6mMravSkn3qcaWbs3mYemrqJn87o83TkBM/tv0P4SGHTcZdEVq2Cgl7SDPBcxUwO9iIiIVBFlkfj5e+btGwVMs9bmUnrXzL2A9yzDTTzrSvKpp24RqUk8c/rVS1pBG7+DfLm85gzykpCWzfgPlxIRHMBbF9fF/6ubILodXPYWOMriY/sMNekB3a8nbOU7/L2XdQ/0slUDvYiIiFRmZfEN4i0gDve1eHOMMc2B0n6aXwK0Mca0NMYEAFcD07wLGGPaeC1eDGwpgzhFpKrpOgaMgwfqLeeblXtrxJx+OXkuJkxexuH0HN69uh2R0290J8FXfwyBIb4O76jznoCgMMYdfpVmdWvx+LR15ORpoBcREZHKqiwGd3nFWtvYWnuRddsJDCmlfB5wN/ADsAH43DMa6NPGmBGeYncbY9YZY1YCDwA3nGmcIlIFhTWCVkMYkvUzyRnZ/LDuoK8jKndPTl/HkrgjvHhFZzoueggStrinbYho6evQjlU7Es5/Eseu+bwRs42th9L4YIEGehEREamsymJwl3BjzL8LBlkxxvwf7ta/EllrZ1pr21prW1trn/Wse9xaO83z+D5rbSdrbay1doi1dt2ZxikiVVTsNQRl7GNE+Dbu+3QFt3+0lKVxidVysJePFu3k48W7mDC4NSOPTIJNM90TtLca5OvQitftemjcg05r/8Wl7YJ5+act7E/O9HVUIiIiUoyy6Oo5EUgFrvLcUoD3y6BeERFofzEEhvNC67XcNfgsFm1P5Mo3F3LZ6wuYsXo/edVkHrlF2w/z1LR1nNu+Hg822QBz/gXdroPe430dWskcDrjoJUg7xD/rziDfZXl2hgZ6ERERqYzKIvFrba19wjMv33Zr7VNAqzKoV0QE/GtB58sI2jKDBwc3YuGj5/L0yE4cycjhro+XM+T/fuP9+TtIz87zdaSnbXdiBndOWU7zyGBeGeLE+e2d0LQPXPx/7uv7KrPG3aHnTYSumsjjvSzfrd6vgV5EREQqobJI/DKNMQMKFowx/QH19RGRshNzDeRmwJJ3CfZzcH2/Fvzy58G8eW0P6oUG8dT09fR77mdemLWRgylVa065jJw8bvtwKbn5Lt4b3ZKQr26AWnXhqo/AL9DX4Z2cc/8OQeGMTfivBnoRERGppMoi8bsDeM0YE2eMiQNeBW4vg3pFRNya9obGPeGnJ+H1PrB0Is68TIZ1bsCXE87mqzvPZkCbKN76fRsDXviFBz5fyYb9lX/eP2stD36xis0HU3n1qk60+HkCpB+Cq6dAaH1fh3fygiNg6FM4di/iza5b2Hoojffna6AXERGRysSU1QAJxpgwAGttijHmfmvty2VS8Unq2bOnXbp0aUU+pYhUpLwcWPc1LHwVDqyGWhHQ82bofRuENgBg1+EMJs7fwedLd5ORk8+As6K4dWBLBrWNxlTCLpP/+3kL//fjZv56UXvGp7wKSyfC5e9C19G+Du3UuVzw3lBI2sk9ke/yc1wWP/95EA3Da/k6MhERkRrDGLPMWtuz2G3lMTKeMWaXtbZZmVdcCiV+IjWEtbBzPix8DTZ9Dw4/6HIl9L0TGnYFIDkjlyl/7OSD+XEcSs2mXf1QbhnYkpGxjQj0c/r4ANxmrzvA+I+WcVm3xvy71TLMjAeg/30w9Glfh3b69q2EtweTGnMzPZcNZWjH+rx6TXdfRyUiIlJj+CLx222tbVrmFZdCiZ9IDXR4Gyx+E1ZMdl8D2GIg9LsL2lwIDgc5eS6mr9rHO3O3s/FAKtGhgdx4dgvG9WlGneAAn4W9+WAql702n9b1QvhiuCVwyihofS6M/RQclSMxPW0z/gxLJzIlZjKPLYIpt/ah/1lRvo5KRHD/KJZvLRG1fff5JyLlSy1+IlK9ZR6BZZPgj7chZS9EtIa+EyD2GgiojbWWeVsTeGfuDuZsjqeWv5PRPZtwy4CWNI8sddrRMpeUkcPI1+aTkZPPjOubU++TC93dVm/7GYLCKzSWcpF5BP7XA1fEWQw5/DD+/n7MvHcgAX5lcUm5iJwOay1fr9jLU9PXE+TvYPo9A6gXGuTrsESkHJRL4meMSQWK29kAtay1fqdV8WlS4ici5OfC+m/d1wHuWwFBdaDnTe658MIaAbDxQArvzt3Btyv3kueyXNCxPuPPaUWP5hHlHl5evosb31/CHzsS+eymrnT7cQwk7XInfVFtyv35K8yKyfDtXazv8zwX/d6MR4e35/ZBrX0dlUiNdCA5i8e+XsPPGw8R27QOmw6k0rlxGFNu7asfZESqoQpv8fMFJX4iUsha2LUIFr0GG2eAcUCny9zXATZ2X3N2KCWLSQvjmLxoF8mZuXRrVofbBrbiwk4NcDrKZyCYp6evZ+L8Hbx4RReu2vF3d5I67gtoM7Rcns9nXC6YeCEkbueeqPc00IuID1hrmbpsD09/t57cfBcPXdieG89uwYw1+7n3kxXc0K85T43s7OswRaSMKfETkZorcQcsfgtWfAQ5adDsbOh3J7S7CBxOMnLymLpsD+/O3cGuxAyaRtTi5v4tuapnU2oHll3HhS+W7uahqau5qX8LngifCb884x7Ipf99ZfYclcr+1fD2IFK73EjP5Rdwfsf6vKaBXkQqxP7kTB79ag2/bYqnd4sIXryyKy1CLfz+PDgDeDbzCt6ZF8dLo2O4skcTX4crImVIiZ+ISFYyLP/QnQQm74a6LaDPBOg2DgJDyXdZflx/gHfm7mDZziOEBfkxrm9zbjy7BfXDzuxamGU7jzD27UX0almXD89OwPn5OOhyFVz+NlTCaSbKzMy/wJJ3+CT2Qx5d6NBALyLlzFrL50t388x3G8hzWR4e1o7r+7XAsXMufHs3JO0EIH/4v7hudVeW7jzCl3ecTZcm1eD6YhEBlPiJiByVnwcbp8PC12HPHxAYDj2uh963Qx33YMTLdh7h3bnb+WHdAZwOw6UxjbhtYCs6NAw75ac7kJzFpa/Oo5a/k+ljIgmfMtx9Pd9N34N/Ne/6mJkEr/bEVac55x75K06nk+/vO0fXFYmUg71JmTzy5WrmbkmgT0t3K1/zEAs/PQFL3oW6LWHkq7Dgf7D1J5LHfMNFX+dirWX6PQOIDAn09SGISBlQ4iciUpzdS9zXAa6f5l7uONI9HUQT9+dl0QnhB7aJ4taBrTinTdRJTQiflZvPmLcWsvVQGt/c3JE2346AnHQY/xuENy7HA6tEVn4C39zBht7/ZPicFjwyvD13aKAXkTJjreXTJbt5dsYGXNby6PD2jOvT3NPKdxck7XaPcnzu3yEg2P2DzDtDICeDDSOmM/LDbfRoVpePbumNn1M/yohUdUr8RERKk7TL3QV0+YeQnQJN+7gHgml/CTj9TmtCeGstD3y+iq9X7OWdcTEMXXEX7FwAN86Apr0r+AB9yFp4fzgkbOa+6Hf5cUeOBnoRKSO7EzN49Ks1zNuaQL9Wkbx4ZVea1nYdbeWLaAUjX4fm/Y7d8eB6ePc8aNCVr7q+wQNfbuDWAS352yUdfXMgIlJmlPiJiJyM7FT3VASL3nBfCxPeDPreAd2ug6AwcvJcTFu1j3dPYkL4d+Zs59mZG3hgaFvuzZ0Ii16Hka9Bt2t9dHA+dGAtvHUOaZ2vpceK4RroReQMuVyWKX/s4vmZGwB49KIOXNO7GY64OTDt7uNb+Yqz9kuYejP0Hs8TuTcwaeFO/nt1LCNja0hvBJFqSomfiMipcOXDppmw8DXYtRACQqH7ddDndqjbonBC+LfnbGfuloTjJoT/fXM8N73/Bxd2asBrHTfgmHYX9LkDhr/g6yPzne8fgcVvugd6WeRk8i19GNBGA72InKrdiRn8ZepqFm4/zICzonj+ii40Cc6HH5+Ape9BRGv3j0xFW/mK88NjsPBV8ka8wTV/tGT13iS+mtCfjo1O/XpmEakclPiJiJyuvcvdrXXrvgbrcnf/7He3u7umMcdNCD+0Q30Wbj9M4zq1+HqEP7WmjIBm/eDar8BZdtNDVDlZyfC/nrjCm3Bu0t800IvIKXK5LJMX7+T57zfiMIa/XdyBMb2aYnbMcY/YmbzbfY3ykMdKbuUrKj8PPhoFe5aQePV3DP88mQA/B9PvHnBcLwYRqRqU+ImInKnkvfDH27DsfXcS07iH+zrAjiPB6c/BlCwmLYhjyuJdOB2G725oRaPPh7tH7rztVwiO8PUR+N6qz+Dr8Wzs9SzD5rbUQC8iJ2nn4XT+MnU1i3ckck7baJ67vAuNa+XBj4/D0onuVr5Rr0OzvqdeeVo8vD0IHE5WXfQtoydtok+rCD64qTdORzWebkakmlLiJyJSVrLTYNUn7lbAxO0Q1gT6jIfuN0CtOmTm5JOTmU74ZyMhYTPc+hPU6+DrqCsHa+H9iyB+I/fXe5fZO3I10ItIKVwuy6SFcbw4axN+DsPfL+nI6J5NMDt+h2/vOb1WvuLsXQYTh0GLAXzW9t88/PV6JgxuzcPD2pfdwYhIhVDiJyJS1lwu2DzLnQDGzQX/2u6BW/rcDr+/AKs/g6s/hvYX+zrSyuXgOnhzIGmdrqHHyos10Es5yMzJ50BKFvuTMzmYksWB5GwOJGdyICWLA8lZHEjJwmkMF3dtyOXdm5zW/JRS/nYkpPOXqatYEneEIe2i+eflXWgYlAez/+7ueRB5lnvEzmZ9yuYJl38I0+6BAQ/waMrlfPLHLl4f152LujQsm/pFpEIo8RMRKU/7V7knhF/7Jbhy3euGPAaD/uLbuCqrWX+FRa/zaewHPLLIXwO9nCRrLUkZuexPznIndClZ7sfJWexPcd8fSMkiOTP3uH1Dg/xoEBZEg/AgGoQFcSQjl983HyI339K+QShXdG/CyNhG1AsL8sGRibd8l+X9+Tv41w+bCPRz8Pilnbiie2PM9t/ciVnyHncr37l/c3clL0vT74NlH5B75SSumhPNpgOpfHNXf9rWDy3b5xGRcqPET0SkIqTsd8+dZfPhvCfgJCZ5r5GyUuDVXrhCG3Je8t8wTj9m1fCBXvLyXcSnZR9N5Iomd57Wuuw81zH7GQNRIYHHJHXH3Hse1w48fmChxPQcvlu9j6+W72Xl7iQcBga0ieaK7o25oGMDagUUP0ellJ9t8Wk89MUqlu9K4rz29fjn5V2oH5ADP/4dln1Q9q18ReVlu+fdjN9EwjWzGDb5IKFBfnxzV3/Ca/mXz3OKSJlS4iciIpXLmqnw5S1s7PUPhs11X0s0YXD1HOglIyevsIvlQe9WOq/kLj41G1eR/44DnA7qhxckdbVoEBbouT+a1NULDcTfeRIJc1YKpOyDlL3um3FC+4ugVl3AnXB8vXwvX6/Yy96kTGoHOBnepSGXd29M35aRODTIR7nKd1nem7ed/5u9mSB/J0+O6Mio2MaY7b/CtHvd56zgWr6ybuUrKnkvvHUOBEew7IKpjPlgHYPaRvPO9T31dyBSBSjxExGRysVamHQpHFzLA/XeY9aOXH56YBCN6lTNgV4S03OYvzWBLYfSjul6uT85k5SsvOPKF+166d06V3AfUTsAc6JWY2shK8mT1BUkdp775L1H1+ekHr+vMwDaDYeYsXDW+eD0x+Wy/BGXyFfL9zBzzQHSsvNoFB7EqG6Nubx7Y86qpy5/ZW3roVQemrqaFbuSGNqxPs+O6ky9Y1r52rhH7Gzau+KCipsHk0ZAu+F82OwZHp+2nvvOa8OfhratuBhE5LQo8RMRkcrn0AZ4cwBpHa6ix6oRnN+hPq+NqxoDveTmu1i+8whztyQwZ0s8a/YmY+2Zdb08jrWQeaSEZM4rqctNL7KjgdAGENbIc2vsde95nJnonl5jzReQkQDBUdDlSoi5GhrGgjFk5uTz44aDfLV8D3O3JJDvsnRtEs7l3RpzaUwjIkMCy+OlrTHy8l28M3cH//lpM8EBTp4a0YkRMY2KtPLdDUP+Wv6tfMVZ+Dr88Cj23Md56OD5TF22h3eu78nQjvUrPhYROWlK/EREpHKa/TdY8D8+j3mfvywOrNQDvew8nM6czfH8vjmBRdsPk5adh9NhiG1ah3PaRHNO2yg6NQo/uWsVrYWMw+6BOoomcgXdMVP2QV7WsfsZB4Q29Erqmhyf4IU2AOdJXo+Vnwtbf4ZVH8Om7yE/B6I7uBPArle56wMOpWYxbaX7esD1+1PwcxgGt4vm8u5NOLd9PYL8dT3gqdh8MJWHvljFqj3JDOvUgH+M6ky0f7b7/bB8EkS1dV/L17SX74K0Fr68FdZ+Sc7YL7hidi3iEtL55u7+tI4O8V1cIlIqJX4iIlI5Zae6B3qpXY/zUx6HSjTQS2pWLgu2HWbulnjmbE5gV2IGAE3q1uKcttGc0yaKfq2jjh/0wuWC9Pgiydye47tj5uccu5/DD0IbeSVyXslcuCfBq10PnCfRWng6Mo/Auq9h1aewezFgoNVgiL3GPS1JQG0ANh5IKbwe8FBqNmFBflzctRFXdG9Mj+Z1T9w9tQbLy3fx1pzt/PenLYQE+fH0yE5c3KUhZtsv7la+1H1w9j0w+K/gXwlGWM1Jh3eHQuo+Dlw9i4s+3E3dYH++uas/oUEa7EWkMlLiJyIildfaL2HqzWzq+SQXzmvrs4Fe8l2WtXuTmbM5nrlbEli+6wh5LktwgJN+rSLdyV7baFpEBh9Nbqx1d1nd/its/w3iN7pHd3UVmVLB4X80kQtvXKQLpudx7WhwVJKWs8Pb3HNRrvoEknZBQAh0HOluCWw+ABwO8l2WBdsS+Gr5XmatPUBmbj7NIoK5zHM9YPPI2r4+ikpl44EUHvpiNWv2JnNxl4Y8NbITUX5Znla+D92tfKPegCbFfl/zncTt8PZgqNOMRUM+YdykNZzfoR5vjOuhwV5EKiElfiIiUnlZCx+OgP2reKD+RL7fnsfPf66YgV4OJGcxZ0s8czbHM39rAkcy3Alb58ZhnNMmmoFtounRvO6xLZCpB9xJ3jZPspd2wL0+sg006uZJ7IpcVxccCQ7ft2KeMpcLdi10J4DrvnEPEhPe1N0NNGYsRLUBID07j1lrD/DVij0s2HYYa6FH87pc3r0xl3RpRHhwzW0dys138cZv2/jfL1sIC/LnH6M6uydF3/qTp5VvP5x9Lwx+tHK08hVn82z4+CroOoZ3o/7CMzM38tCF7bhryFm+jkxEiqh0iZ8xZhjwX8AJvGutfb7I9geAW4E8IB642Vq7s7Q6lfiJiFRh8ZvgjbNJaz+anmtGcl778hnoJSs3n8U7Epm7OZ45W+LZfDANgOjQQAa2iWJQ22j6nxVFlPfAJTkZsHOBu1Vv269waJ17fXCkuytkqyHQeoi7O2Z1lpMBm2a6u4Ju+xmsCxr3dLcCdr4CgiMA2J+cyTcr9vHV8j1sOZRGgNPBeR3qcXn3JgxqG10puvFWlPX7Unho6irW7Uvh0phGPDWiExHOTPjhMVjxEUS187Ty9fB1qCf2+4vw67PY4f/i/h29mLZqHxNv7MWQdvV8HZmIeKlUiZ8xxglsBoYCe4AlwFhr7XqvMkOAxdbaDGPMBGCwtXZMafUq8RMRqeJ+fBzm/5cvYiby0OIgPrqlNwPbRJ9RldZaNh9MY44n0Vu8I5GcPBcBfg56t4hgYJsozmkbTfsGoUe7b7pcsH/l0URv92L39XjOQGjW153ktT4X6nepmq14ZSH1gHtE0JWfuBNhhz+0vdB9PeBZQ8EvAGst6/al8OXyPUxbuY/D6TnUDfZnREwjLu/ehK5Nwqvt9YA5eS5e/20rr/6ylTrBATwzqjPDOjeALT/BdE8rX//7YNAjlbeVryiXCz69Brb+SPa4aYz6zsXeIxlMv2eAuvWKVCKVLfHrBzxprb3Qs/wogLX2uRLKdwNetdb2L61eJX4iIlVcdpp7oJfgSM5PffK0B3pJTM9h7hb3dXpzt8RzMCUbgLPqhRSOvtmnZSS1Aryup0va5em66em+mXnEvb5+F2g92N2q16wfBASXyaFWKwfWuFsBV38O6YegVsTRqSEadQdjyM13MXdLPF8u38uP6w+Sk+eiVXRtrujehFHdGtO4is7fWJy1e5N58ItVbDyQyqjYRjxxaSfqOjPhh7/CiskQ3d49YmdVaOUrKisZ3h4COWnsvWoWF3+wlQZhQXx159kEB5TToEMickoqW+J3JTDMWnurZ/k6oI+19u4Syr8KHLDWPlPMtvHAeIBmzZr12Lmz1N6gIiJS2a37Gr64kc09nuCC+e1OaqCXnDwXK3YdYY4n2SuYUy+8lj8D2kRxTpsoBraJPvaawaxk2DH3aKte4jb3+tCGR7tuthoMIerGdtLy82DbL+7rATfOgPxsd1fGgqkhPF1hkzNz+X7Nfr5avpc/4hIB6Nsqgsu7N2F45wZVdrTI7Lx8Xv1lK6//to2I2gE8O6ozF3RqAFt+hOn3eVr57odBD1edVr7iHFwP754PDToz9+z3ueHDlVzUpSH/G9ut2rbgilQlVTbxM8ZcC9wNDLLWZpdWr1r8RESqAWvho1GwdwV/bvAeM7fnFzvQS1xCOnO3uOfUW7gtgfScfJwOQ7emdQpH3+zSOBxnwaiD+bmwd5k7ydv2i/uxzQf/2tBigCfRGwLR7dyzsMuZyUyC9d+6k8BdCwEDLc9xDwjT4VIIdM8Dtzsxg69XuKeG2JGQTpC/gws6NuDy7o0ZcFYUfk7fdKW11pKbb8l3WfJcLvJdpS8npOXwzxkb2HQwlcu7N+bxSzpSx2TA7MeOtvKNeh0aV8FWvuKs/Qqm3gS9buON2hN4YdZG/npRe8afU/Gj8YrIsSpb4ndSXT2NMecD/8Od9B06Ub1K/EREqomELfB6P9LbXU6PtZdxbvt6vHBFVxZsO1w41ULBnHpNI2oVjr559lmRhBW0FlnrnpJg2y/uVr0dc90jUhqHe+TNgla9Jr3BL8CHB1sDJG53dwNd9QkciQP/YOgwwt0S2PIccDix1rJidxJfLd/D9FX7Sc7MJTo00NMC6EdeviXPdTTR8l7OzXd51he/nJfvKqWsJd/l8irrrt91Gl+N6ocF8s/LunBeh/ruVr5p90LaQRhwv7uVzy/whHVUKbP/DgtewY58nbvWt2fW2gN8eHMfBrSJ8nVkIjVaZUv8/HAP7nIesBf34C7XWGvXeZXpBkzF3TK45WTqVeInIlKN/PQUzPs3U2Pe5cHFwTgdhnyXpXaAk36t3XPqDWxTZE699MOw47ej0ywk73avr9P86IAsLc+BWnV9dVQ1m7XugXJWfQJrv4bsZPdUFwVTQ0S3A9xdJn/dGM9Xy/fw26Z4XNbidBj8HAanw+DvdBQu+zkdheuP32bwc5Sw7Cyoz+G1rfhlf6fxen6H17ajz+fvNPRsEUGYTXeP2LlyMkR38LTylf3otJVCfh5Mvgx2/0HmdTMZ+VUa8anZTLt7AE0jdC2siK9UqsQPwBhzEfAy7ukcJlprnzXGPA0stdZOM8b8BHQB9nt22WWtHVFanUr8RESqkZx0eLU3rqBwHot+lbqhwZzTNpruzbzm1MvNgt2Ljg7Ksn81YCEo3J3gFbTqRbTy6aFIMXKzjk4NsfUnd7fbRt3cCWDnK6F2pK8jPHWbZ7uv5avOrXxFpSfAW4PAONh55QwueW8DzSKC+XLC2QT5O0+8v4iUuUqX+JUHJX4iItXM+mnw+XUw7AXoe4e7xejguqMDsuxcAHmZ4PBzd9ksuE6vUTdwaoTBKiPtEKyZCqs+do8Q6vCDNhe6u4K2vfD45Mla9xyC+bngygNXLrjyj1925XmtyzuD5XxPnaUspx9yJ7D1OsLI16pvK19x9i6HicOgeT9+6fkGt3y0glGxjfn3VTEa7EXEB5T4iYhI1WMtTL4C9iyBtsPc3TfTPZd8R7V1d91sNQRa9IfAUJ+GKmXkwFpY7ZkaIu2g+3pAv8DjEztfcPh5bv7gcILT/+g6p797EvtzHqr+rXzFWf4RTLsbBvyJV8w4/v3jZp64tCM39W/p68hEahwlfiIiUjUd3gZvDnAnAK0GH23VC2/s68ikPOXnua/X3Dzb3brn8HO34hYmXmW97J3IlbCs1qvSTb8flr2Pa/Qkxi9twq+bDjHl1j70bVUFu+2KVGFK/EREpOrKTnMnfg7fDO0vIichLxs+uBgObSDt+tmM+CyelMxcpt8zgIbhtU68v4iUidISP/0vKiIilVtgiJI+kcrOLxCu+hD8axHy9Q28M6YtmTn53DF5OVm5+b6OTkRQ4iciIiIiZSGsEYyeBInbaT33Qf5vdFdW7U7iiW/XUV16mIlUZUr8RERERKRstOgPFz4Lm2Yw7MjH3D3kLD5bupuP/9jl68hEajwlfiIiIiJSdvrcAV1Gwy/P8KeWuxncLponp61j2c4jvo5MpEZT4iciIiIiZccYuPQVqN8Z51e38MqFETSqU4sJk5dxKCXL19GJ1FhK/ERERESkbAUEw5iPAEvYtzfyztWdSM3KY8KU5eTkuXwdnUiNpMRPRERERMpeREu44j04uJa2fzzGv67swrKdR/jHd+t9HZlIjaTET0RERETKR5uhMOQxWPM5l2R9x+3ntOKjRTv5fMluX0cmUuMo8RMRERGR8jPwz9DuIvjhrzzU4TADzorib9+sZdXuJF9HJlKjKPETERERkfLjcMBlb0LdFvhNvYlXL21IdGggd0xeRkJatq+jE6kxlPiJiIiISPkKCocxUyAnnTrTb+Hta7qQmJ7DXVOWk5uvwV5EKoKfrwMQERERkRqgXnsY9Tp8cQOdVj/H81c8wJ8+W8VzMzfy+KUdfR3dScl3WfYeyWRbfBrb4tPYlZiBwxhqBTip5e++BQU4CfJzHLeu4HGtACdBfk6CAhwEOB0YY3x9WFJDKPETERERkYrRaRTsuw/m/5fLGndndf9uTJy/gy5NwrisWxNfR1coNSuX7fHpbE9IY9uhdLbFp7E9Pp0dh9OPmY4iNMgPA2Tlusg5jZZLh+FoMliQJB6TLDqO2V6w7fhk0uFJJoskmJ7H/k6jBFOU+ImIiIhIBTr3cdi3Er57gL/eNIv1+yJ49Ks1tKkXSufG4RUWhstl2ZuUWZjUed8fSj167aHTYWgeEUyr6NoMbhdN6+gQWkXXplV0CBG1AwrL5eW7yMpzkZmTT1au+5aZm09mjvv+6LKr8HGW1/aC5axcdx0pmbkcSjm2juzTTDCdDkNYkB+RIYFE1g4gKiSQyJAAImu776NCAgq3RYYEEhbkp0SxGjLWWl/HUCZ69uxply5d6uswRERERORE0g/D24MAODxuNpe8twGnwzD97gHU9UqmyuSpsvO8Wu/S2JaQzrZDaexISCfbq/UuvJY/rT0JXUFy1zo6hGYRwQT4lTAsRl4OOJzuWwUpmmAWJIbHJJOeBNM7uUzKzOFwWg6H03M4nJbN4fQckjJyi30Of6chsnYgEbUDPInh0aSwMFEsTBoDCfKvuOOX0hljlllrexa7TYmfiIiIiFS4fSvgvQuhWV9WDZ7I6HeW0LtFBB/c1As/56mNP+hyWfanZLE93pPceXXTPJCSVVjOYaBZRLAnuTs2yYusHVByK1f6YUjYBPGbIGHz0ftkz3yEDn/wrwV+QeAfBH61TnDvuRXuU/T+BHX4BUIZtMjl5rs4kp5DQloOh9OzOZyWQ4InKTycln00UfRsy8jJL7ae2gFOIkPciaJ3UhgZElhkOYCI4IBTPr9y8pT4iYiIiEjls2IKfHsn9L+Pz+vcxl++XM3tg1rx6PAOxRbPyClovXO32m33ar3LzD2alIQG+RUmd6297ptFBhPoV0LrlLWQstcrudsI8ZvdCV/G4aPl/GpBVBuIbg+RrQEDeZmQm3US91mQm3n03hafSJ2YOYkEMehoGf9gqFUXQupBSH0IaeB5XM+dbJ6kjJy8Y1sN03JI8CSFBa2ICWk5JHrW5bmKzzPqBvuXmChGhwQQHRpEvdBA6oUFlny+pFilJX66xk9EREREfKPbONi7DOb/l6tGd2N137N46/fttIqqTZO6wcddf7c3KbNwV2OgaV33tXf9WkcWds1sFV2b6JDAklvv8vPgyA5PgrfpaHKXsAVy0o6Wq1UXotpB+4vd99HtIKothDd1z01YFvJzj00EC++zTyGZLCapzEgsUmcGZKUAxSRigeHuBDC0gVdiWM8rOazvvgVHEhzgR3CEH00jgk94aNZaUjLzjkkME45pScwmIS2HTQdSSUw/zJESup3WCfanXmgg9cOCiA4NpF5oEPXD3Pf1wgKp77lXd9MTU4ufiIiIiPhOXg58cBEcXE/OzT8x9psklu08Urg5JNDvaFIXVZvW9dzJXYvI2qV/2c/NdCdzBV0z4ze6Hx/eBi6vJCO0EUS3PZrcRbdzP64dVSbdKSuN/DxIj4e0g5B2yHPvffOsSz0IuenH72+cR1sJC5LBwptnXahnOaD2KYeXm+/iSEYO8anZHErN5lBKFodS3I8PpmRxKDXbsy2L3Pzj85fQID/qh3laCr0TxbAg6nvu64UGUjuwerd7qauniIiIiFReKfvhrXMgKIzka2fz845MGoQH0To6hHqhpbTeAWQeOdpq530NXtIuClu4jAPqtjzaaleQ3EW1gaCwCjnEKiU77eQSxPT44rurBoQU02pY0KrolSgGR4Hz1BIxl8uSlJnLodQsDqZ4EsRUr3uvRNF76o0CIYF+1AsNJNqTHBZ0KS3aohgSWDVHNlXiJyIiIiKV284FMOlSaHMhjJl8bHdKayH1wLFdM+M9t/RDR8s5A93JXGFy1/botXh+gRV/TNWdK9/drbS05LBgOTv5+P2Nw538Hddq2MDr3nM7hWsR4WhX04OpBS2HnkQx9dhE8WBKFlm5xyeItfydhV1Jo8OOtiLW80oOm0fWLnnEVx9R4iciIiIild/it+D7v0Dv8RDW+NgRNLNTjpYLDPfqnul1X6d5hU6tIKcgN7OYFsRiEsS0g8d2xS0QGH60K2low+MTw4JkMTD0lMKy1pKanVeYHHrfH0w9tkUxvcioprP/dA5t65/a85U3De4iIiIiIpVf7/HuwV7+eNu9HFLf3WrX9aqjyV10e/f6KtgNr0bzrwV1m7tvpXG53N13U/dD2gFPUui5T93vTgx3L3bf52Udv79/bXcCGNrQkyQ2KCZZrA9BdcAYjDGEBfkTFuTPWfVCSg0tLTvvmC6lTeueeJCbykQtfiIiIiJSebhc7oFYwhq6R9YUKY61kJV0bGKYdsDdJTj1gKcl0fO4uMFq/IKKJIZFWg4LWhNrRZTdKK4VQC1+IiIiIlI1OBxQv6Ovo5DKzhj3DwO16kK99qWXzU4tJjH0ShbjN8L234u/DtHhf/y1hwWtiR0uheCI8jm+cqDET0REREREqq/AUPct6qzSy+VkHL3+MHX/8a2JR3bAroWQmegu32KAEj8REREREZEqJSAYIlq6b6XJy3YPRBPasGLiKiM+6bBqjBlmjNlkjNlqjHmkmO3nGGOWG2PyjDFX+iJGERERERGR4/gFQp2mpzwHoa9VeOJnjHECrwHDgY7AWGNM0Y7cu4AbgY8rNjoREREREZHqxxdpam9gq7V2O4Ax5lNgJLC+oIC1Ns6z7fjZFEVEREREROSU+KKrZ2Ngt9fyHs+6U2aMGW+MWWqMWRofH18mwYmIiIiIiFQ3VWdSimJYa9+21va01vaMjo72dTgiIiIiIiKVki8Sv71AU6/lJp51IiIiIiIiUg58kfgtAdoYY1oaYwKAq4FpPohDRERERESkRqjwxM9amwfcDfwAbAA+t9auM8Y8bYwZAWCM6WWM2QOMBt4yxqyr6DhFRERERESqC59MPmGtnQnMLLLuca/HS3B3ARUREREREZEzVKUHdxEREREREZETU+InIiIiIiJSzSnxExERERERqeaU+ImIiIiIiFRzSvxERERERESqOSV+IiIiIiIi1ZwSPxERERERkWpOiZ+IiIiIiEg1p8RPRERERESkmlPiJyIiIiIiUs0p8RMREREREanmlPiJiIiIiIhUc0r8REREREREqjklfiIiIiIiItWcEj8REREREZFqTomfiIiIiIhINafET0REREREpJpT4iciIiIiIlLNKfETERERERGp5pT4iYiIiIiIVHNK/ERERERERKo5JX4iIiIiIiLVnBI/ERERERGRak6Jn4iIiIiISDWnxE9ERERERKSaU+InIiIiIiJSzSnxExERERERqeaU+ImIiIiIiFRzfr4OQER8z1qLxeKyLiwWa92PXdYF4H6My13OWly4tx2zn2f9cWVOsr6CMqXVZwr+GYPDOHAYB4ajj4Gj643BQenrjTHH7F/w+Lj1J1lXZWWtLXV7ZY7d17z/Lgv+3r3/Rgv/vj1/qwUK/oYM5pjl47Z7XvvCsp4i3n/rhctV9DwVvC4WW/wyFs/DY9eVwPt1LG7Zs/KU9inutT2ZMjVF0c/6fJt/zOe19/pi1xX5f8NlXceULfw/wlOmYF3BZ33B33/hZ3SRZeCYz3AMODj6GY0pst2rfLH1FvOZf0xdOI57D9cE3n8Hx90ocm69/t/3Pq8l/Q1438rKic5NsZ8dp1imTd02BPkFnVJcvqTErxwtObCE22bfVrh8zB/PMQ9N8Y+9/sMvuq6oksqUVLf3lwvv/Yp+mJX2YVvcB3LRD82TqaO0uop+gBf8K5pUFPynguWYDx/guOSj4EPGe1txZYt+uTumrNfzFP3i512+QGlfYop+KfcuW+q2onUet1hyWe96S4tNTo3333hxSnqtS1xfQsJWXues8L3o+XJzzHu6mCTEO1kp7ovYicoVV6f36+ddrjDlKua9ecz7j+Pfg0U/K4r+yOG9XPS9XVl5v44Fy97nDI5PLr33K1DccZaWqBWWLSFRK7pvTVL0dS9cVySZP65MCefHu3zBuS2u/uL+zy+tjuL+PzzmS3kJ66RkJ3o/lvZeLPr5CJT6GVn4uITP1IL9j3sOY06cjOHC5Sp+W0GSJseaNmoaLcNb+jqMk+aTxM8YMwz4L+AE3rXWPl9keyDwIdADOAyMsdbGVXScZ6pB7Qbc3Pnm49aX9MW+2C/qXu+xU9qPE39pLFpHSV+aivuSVdIXp4J6S/yF3OvLV0GiVvTXweO+hBVJziz22NaZYr5wOowDp8N5/K9zXglp0W2l/QpY7C+KJZQtmtQW8H4MJSfonhUl7+ddpyl5W1GllS2p9euY18t7nXEc95oW1xrm/Zp4v+4l1YfnV1nvRN/79Yejf6vF/drs/YW9uPXePwoUbV081bqKW+9dV2nn4mTOz4nO6SmXL2F90YSq4HOh8P3qeR8WfT8fkwwU+Ww4Zv8TlCtaZ0nl4AS/0Hv9LRX3mVDc+9P7b7jg38n+wl/S83t/sS76eeydJJW63es18exwzHLRc1Vcnd7nrbg6C1pTCl6b4r4weh4cs1z0y+cxf2NF1h/3RbiE+o4rewotKMX9P+f9uhZX5rjtFPN/ZSk/pJX2HMWdl4LtRc9N0fJFz493HcXVX9rfV2l1wNHP6aK9JorrzVBSDwencR7XU8JhHDiNs9iyRdcV1l10HQ4cDscxx1XsjznF/eBbsHwyP+oWLBfzw9ExPxiXsq1oXUVf76KfocccUynntbTPgxN9phb3HAUxHPO6e53/wvPjva2Y81/0PBZ3K+7vqehzlfQ34F2uNMW+Z0+xTEnfkU/1eeoH1z9hmcqkwhM/Y4wTeA0YCuwBlhhjpllr13sVuwU4Yq09yxhzNfACMKaiYz1TTUObcm/3e30dhoiIiIiI1HC+GNylN7DVWrvdWpsDfAqMLFJmJDDJ83gqcJ6pSZ2oRUREREREypAvEr/GwG6v5T2edcWWsdbmAclAZNGKjDHjjTFLjTFL4+PjyylcERERERGRqq1KT+dgrX3bWtvTWtszOjra1+GIiIiIiIhUSr5I/PYCTb2Wm3jWFVvGGOMHhOMe5EVEREREREROkS8SvyVAG2NMS2NMAHA1MK1ImWnADZ7HVwK/2JMZfkdERERERESOU+Gjelpr84wxdwM/4J7OYaK1dp0x5mlgqbV2GvAe8JExZiuQiDs5FBERERERkdPgk3n8rLUzgZlF1j3u9TgLGF3RcYmIiIiIiFRHVXpwFxERERERETkxJX4iIiIiIiLVnKkuY6YYY+KBnb6OoxhRQIKvg5AT0nmqGnSeqgadp8pP56hq0HmqGnSeKr+adI6aW2uLneeu2iR+lZUxZqm1tqev45DS6TxVDTpPVYPOU+Wnc1Q16DxVDTpPlZ/OkZu6eoqIiIiIiFRzSvxERERERESqOSV+5e9tXwcgJ0XnqWrQeaoadJ4qP52jqkHnqWrQear8dI7QNX4iIiIiIiLVnlr8REREREREqjklfiIiIiIiItWcEr8yYowZZozZZIzZaox5pJjtgcaYzzzbFxtjWvggzBrNGNPUGPOrMWa9MWadMea+YsoMNsYkG2NWem6P+yLWms4YE2eMWeM5B0uL2W6MMa943k+rjTHdfRFnTWWMaef1HllpjEkxxtxfpIzeSz5gjJlojDlkjFnrtS7CGPOjMWaL575uCfve4CmzxRhzQ8VFXfOUcJ7+ZYzZ6PlM+9oYU6eEfUv9fJSyU8J5etIYs9frs+2iEvYt9XuhlI0SztFnXucnzhizsoR9a9x7Sdf4lQFjjBPYDAwF9gBLgLHW2vVeZe4Eulpr7zDGXA1cZq0d45OAayhjTEOgobV2uTEmFFgGjCpyngYDD1prL/FNlALuD2Ogp7W22MlWPf/R3gNcBPQB/mut7VNxEUoBz+ffXqCPtXan1/rB6L1U4Ywx5wBpwIfW2s6edS8Cidba5z1fQOtaax8usl8EsBToCVjcn489rLVHKvQAaogSztMFwC/W2jxjzAsARc+Tp1wcpXw+Stkp4Tw9CaRZa18qZb8Tfi+UslHcOSqy/f+AZGvt08Vsi6OGvZfU4lc2egNbrbXbrbU5wKfAyCJlRgKTPI+nAucZY0wFxljjWWv3W2uXex6nAhuAxr6NSk7TSNwf8tZauwio40nspeKdB2zzTvrEd6y1c4DEIqu9//+ZBIwqZtcLgR+ttYmeZO9HYFh5xVnTFXeerLWzrbV5nsVFQJMKD0yOUcL76WSczPdCKQOlnSPP9+yrgE8qNKhKTIlf2WgM7PZa3sPxCUVhGc8HezIQWSHRyXE8XW27AYuL2dzPGLPKGPO9MaZTxUYmHhaYbYxZZowZX8z2k3nPScW4mpL/U9V7qXKob63d73l8AKhfTBm9pyqXm4HvS9h2os9HKX93e7rkTiyh67TeT5XDQOCgtXZLCdtr3HtJiZ/UOMaYEOBL4H5rbUqRzcuB5tbaGOB/wDcVHJ64DbDWdgeGA3d5unJIJWOMCQBGAF8Us1nvpUrIuq/v0DUelZgx5jEgD5hSQhF9PvrWG0BrIBbYD/yfT6OR0oyl9Na+GvdeUuJXNvYCTb2Wm3jWFVvGGOMHhAOHKyQ6KWSM8ced9E2x1n5VdLu1NsVam+Z5PBPwN8ZEVXCYNZ61dq/n/hDwNe5uM95O5j0n5W84sNxae7DoBr2XKpWDBV2hPfeHiimj91QlYIy5EbgEGGdLGIThJD4fpRxZaw9aa/OttS7gHYp//fV+8jHPd+3Lgc9KKlMT30tK/MrGEqCNMaal5xfwq4FpRcpMAwpGSbsS9wXc+tW1Ann6er8HbLDW/ruEMg0Krr00xvTG/R5Rgl6BjDG1PYPvYIypDVwArC1SbBpwvXHri/vC7f1IRSvx11S9lyoV7/9/bgC+LabMD8AFxpi6nq5rF3jWSQUxxgwD/gKMsNZmlFDmZD4fpRwVuZ78Mop//U/me6GUr/OBjdbaPcVtrKnvJT9fB1AdeEbguhv3f5JOYKK1dp0x5mlgqbV2Gu6E4yNjzFbcF6Fe7buIa6z+wHXAGq+hff8KNAOw1r6JOymfYIzJAzKBq5WgV7j6wNeenMEP+NhaO8sYcwcUnqeZuEf03ApkADf5KNYay/Mf5VDgdq913udI7yUfMMZ8AgwGoowxe4AngOeBz40xtwA7cQ92gDGmJ3CHtfZWa22iMeYfuL+wAjxtrT2dQS3kJJRwnh4FAoEfPZ9/izwjgTcC3rXWXkQJn48+OIQaoYTzNNgYE4u7y3Qcns9A7/NU0vfCij+C6q+4c2StfY9irj/Xe0nTOYiIiIiIiFR76uopIiIiIiJSzSnxExERERERqeaU+ImIiIiIiFRzSvxERERERESqOSV+IiIiIiIi1ZwSPxEREcAYk2+MWel1e6QM625hjKn2c0SJiEjlpXn8RERE3DKttbG+DkJERKQ8qMVPRESkFMaYOGPMi8aYNcaYP4wxZ3nWtzDG/GKMWW2M+dkY08yzvr4x5mtjzCrP7WxPVU5jzDvGmHXGmNnGmFqe8vcaY9Z76vnUR4cpIiLVnBI/ERERt1pFunqO8dqWbK3tArwKvOxZ9z9gkrW2KzAFeMWz/hXgd2ttDNAdWOdZ3wZ4zVrbCUgCrvCsfwTo5qnnjvI5NBERqemMtdbXMYiIiPicMSbNWhtSzPo44Fxr7XZjjD9wwFobaYxJABpaa3M96/dba6OMMfFAE2tttlcdLYAfrbVtPMsPA/7W2meMMbOANOAb4BtrbVo5H6qIiNRAavETERE5MVvC41OR7fU4n6PX2V8MvIa7dXCJMUbX34uISJlT4iciInJiY7zuF3oeLwCu9jweB8z1PP4ZmABgjHEaY8JLqtQY4wCaWmt/BR4GwoHjWh1FRETOlH5VFBERcatljFnptTzLWlswpUNdY8xq3K12Yz3r7gHeN8Y8BMQDN3nW3we8bYy5BXfL3gRgfwnP6QQme5JDA7xirU0qo+MREREppGv8RERESuG5xq+ntTbB17GIiIicLnX1FBERERERqebU4iciIiIiIlLNqcVPRERERESkmlPiJyIiIiIiUs0p8RMREanijDEfGGOe8cHzfm+MuaGin1dERE6dEj8RETktxpg4Y0yOMSaqyPoVxhhrjGnhWT4mKTHGBBpjnjPG7DLGZBpjthhjHjLGmCJ1n1+k3huNMfPK+bAwxjxpjJlc3s9TWZ//VFhrh1trJ/k6DhEROTElfiIiciZ2cHReO4wxXYDgE+zzBXAecBEQClwHjAf+W04xVihjjObIFRGRSkeJn4iInImPgOu9lm8APiypsDHmPOAC4Apr7VprbZ61dhFwLXCXMeas0w3EGHObMWarMSbRGDPNGNPIa5s1xtzhaV1MMsa85t3C6FVuGPBXYIwxJs0Ys8qz/iZjzAZjTKoxZrsx5navfQYbY/YYYx42xhzAPal7LWPMJGPMEc9+fzHG7PHap5Ex5ktjTLwxZocx5t7Snr+YOLsZY5Z74vkMCDrF1+JOz2uRaoz5hzGmtTFmgTEmxRjzuTEmwFO2rjHmO0+cRzyPm3jV9Zsx5lbP4xuNMfOMMS95yu4wxgw/lXMoIiLlR4mfiIiciUVAmDGmgzHGCVwNlNZNcSiw2Fq723ultXYxsAd3S+ApM8acCzwHXAU0BHYCnxYpdgnQC+jqKXdh0XqstbOAfwKfWWtDrLUxnk2HPPuHATcB/zHGdPfatQEQATTH3Xr5BNACaOU55mu9YnUA04FVQGPPMd9vjLmwlOf3PtYA4BvcSXcE7hbUK07xtbgQ6AH0Bf4CvO2JsSnQmaOtuA7gfc9xNQMygVeLxuSlD7AJiAJeBN4rLsEWEZGKp8RPRETOVEGr31BgA7C3lLJRwP4Stu33bD8d44CJ1trl1tps4FGgX8F1hh7PW2uTrLW7gF+B2JOt3Fo7w1q7zbr9DswGBnoVcQFPWGuzrbWZuJOuf1prj1hr9wCveJXtBURba5+21uZYa7cD7+BOmk9GX8AfeNlam2utnQos8dp+Mq/Fi9baFGvtOmAtMNtau91amwx8D3TzHPdha+2X1toMa20q8CwwqJTYdlpr37HW5gOTcCee9U/yuEREpBzpOgQRETlTHwFzgJaU0s3TIwFoU8K2hp7tAHm4kxtv/kBuCfs2ApYXLFhr04wxh3G3qMV5Vh/wKp8BhJwg1kKeLotPAG1x/2gaDKzxKhJvrc0qEo93q6b34+ZAI2NMktc6JzD3JMNpBOy11lqvdTuLbD/Ra3HQq3xmMcsNAIwxwcB/gGFAXc/2UGOM05PcFVX4GltrMzyNfSf9OouISPlRi5+IiJwRa+1O3IO8XAR8dYLiPwF9jDFNvVcaY/rg7mb4i2fVLtxdJb215NgEx9s+3AlVQX21gUhKb30siXdChTEmEPgSeAmob62tA8wETEn74G69bOK17H28u4Ed1to6XrdQa+1FJdRV1H6gcZEulM28Hpfla/FnoB3Qx1obBpxTUO1p1CUiIj6kxE9ERMrCLcC51tr00gpZa38Cfga+NMZ0MsY4jTF9cV8X+Ia1doun6Ge4r3trb9x6Ajdz/LVqBT4BbjLGxHoStX/ivpYw7jSO5SDQwnMtHkAAEAjEA3me1r8LTlDH58CjnsFRGgN3e237A0j1DAZTy/MadDbG9Crh+YtaiLtF9F5jjL8x5nKgt9f2snwtQnG3ACYZYyJwt3qKiEgVpMRPRETOmOf6t6UnWfwK3NfYzQLScCd97wH3eJV5B/egItOBZNxdSB/zDH5S3PP/BPwdd8vcfqA1J3/NXFFfeO4PG2OWe65tuxd3MncEuAaYdoI6nsY9WM0O3K2cU4FsT6z5uAeKifVsTwDeBcKLe/6iFVtrc4DLgRuBRGAMXi2tZfxavAzU8sS4CPc5ExGRKsgce4mAiIiIlDVjzATgamttaQOjiIiIlBu1+ImIiJQxY0xDY0x/Y4zDGNMO97VyX/s6LhERqbk0qqeIiEjZCwDewj0gTRLuaxNf92VAIiJSs6mrp4iIiIiISDWnrp4iIiIiIiLVXLXp6hkVFWVbtGjh6zBERERERER8YtmyZQnW2ujitlWbxK9FixYsXXqyI4mLiIiIiIhUL8aYnSVtU1dPERERERGRak6Jn4iIiIiISDWnxE9ERERERKSaU+InIiIiIiJSzSnxExERERERqeaU+ImIiIiIiFRzSvxERERERESqOSV+IiIiIiIi1ZwSPxGRGmjroTRunbSE5MxcX4ciIiIiFUCJn4hIDfTl8j38tOEQ01ft83UoIiIiUgGU+ImI1EBzt8QD8NXyPT6ORERERCqCEj8RkRrmcFo2a/em0CAsiOW7ktiRkO7rkERERKScKfETEalh5m87DMBTIzvhMPC1Wv1ERESqPSV+IiI1zNzN8YTX8uf8DvXpf1YUX63Yi8tlfR2WiIiIlKNyTfyMMcOMMZuMMVuNMY8Us/0OY8waY8xKY8w8Y0zHItubGWPSjDEPlmecIiI1hbWWuVsS6H9WJE6H4YruTdhzJJMlcYm+Dk1ERETKUbklfsYYJ/AaMBzoCIwtmtgBH1tru1hrY4EXgX8X2f5v4PvyilFEpKbZFp/GgZQsBpwVDcAFnepTO8DJV8v3+jgyERERKU/l2eLXG9hqrd1urc0BPgVGehew1qZ4LdYGCvsaGWNGATuAdeUYo4hIjTJncwIAA9tEARAc4MfwLg2ZsWY/Wbn5vgxNREREylF5Jn6Ngd1ey3s8645hjLnLGLMNd4vfvZ51IcDDwFOlPYExZrwxZqkxZml8fHyZBS4iUl3N3RJPy6jaNI0ILlx3effGpGXnMXv9QR9GJiIiIuXJ54O7WGtfs9a2xp3o/c2z+kngP9batBPs+7a1tqe1tmd0dHQ5RyoiUrVl5+WzaHsiA86KOmZ935aRNAoP0px+IiIi1ZhfOda9F2jqtdzEs64knwJveB73Aa40xrwI1AFcxpgsa+2r5RGoiEhNsHxnEpm5+YXdPAs4HIbLujfmjd+2cSg1i3qhQT6KUERERMpLebb4LQHaGGNaGmMCgKuBad4FjDFtvBYvBrYAWGsHWmtbWGtbAC8D/1TSJyJyZuZuicfpMPRrHXnctsu6NcFlYdrKfT6ITERERMpbuSV+1to84G7gB2AD8Pn/t3ff8VVX9x/HXyd7kpDFSMIMyB5JQPZwD3CguEfdOOrosLZWba221VptrXvvBaKi4lYgLCEJe4eZhJUwQgaZ9/z+SPAXkU3u/d57834+HvdB8r3f8Q439+Z+7lnW2mXGmAeNMec07HabMWaZMWYh8BvganflERFp7mbmFdM/NZbosOBf3JeWFEXf1Fg+1OyeIiIifsmdXT2x1k4Fpu637f5GX99xBOf4S9MnExFpXnaVV7OksIQ7T+560H0uSE/m/k+WsXzzHnq0beHBdCIiIuJujk/uIiIi7jdrbTHWwvCuCQfdZ0yftgQHGj5aoEleRERE/I0KPxGRZiBrdTHRYUH0SY456D5xkSGMPiGJjxduprbO5cF0IiIi4m4q/ERE/Jy1lqw1RQztnEBQ4KFf9selp1BUWsXMvGIPpRMRERFPUOEnIuLn1hWXs7mk8pDdPPcZ3S2R2IhgJmuSFxEREb+iwk9ExM9lrS4CYHha4mH3DQ0KZGyftny1bCullTXujiYiIiIeosJPRMTPZa0ppn18BO3iI45o/3HpyVTVuvhiyVY3JxMRERFPUeEnIuLHqmtdzF23g+FdDt/Nc59+qbF0Sojkw1zN7ikiIuIvVPiJiPixBZt2UV5dx7Aj6Oa5jzGGcenJ/Lh+J/k7K9yYTkRERDxFhZ+IiB/LWlNMYIBhcOf4ozruvP7JAHy0QJO8iIiI+AMVfiIifiwrr5h+qbHEhAcf1XEpLSMY1CmOybkFWGvdlE5EREQ8RYWfiIif2l1RzeKC3QxLO/LxfY2NS09hw44KcjftbtpgIiIi4nEq/ERE/NSsvB1YCyOOYP2+AzmzV2vCggOYrEleREREfJ4KPxERPzUzr4jo0CD6psQe0/HRYcGc3rM1ny7aTFVtXdOGExEREY9S4Sci4oestcxYXczgzvEEBR77S/249BT2VNby/YrtTZhOREREPE2Fn4iIH1pfXE7h7r0M73rkyzgcyLC0BJKiQ/kwV7N7ioiI+DIVfiIifmhmXjEAw49xYpd9AgMM5/dPZtqq7ewoq2qKaCIiIuIAFX4iIn5oxupiUuPCaR8fcdznGpeeQq3L8umizU2QTERERJygwk9ExM/U1LmYu24Hw7skYow57vOd0Dqanm1bMFmLuYuIiPgsFX4iIn5mYf5uyqpqj7ubZ2Pj0lNYXFDCmm2lTXZOERER8RwVfiIifiZrdREBBoZ0brrC75y+bQkMMGr1ExER8VEq/ERE/ExWXjF9U2OJiQhusnMmRocysmsiHy8opM5lm+y8IiIi4hkq/ERE/EhJRQ2L8nc3aTfPfcalJ7OlpJK563Y0+blFRETEvVT4iYj4kdlri3FZjnv9vgM5pXsrosOC+DC3oMnPLSIiIu6lwk9ExI9k5RUTFRpEv9TYJj93WHAgY/q04culWymvqm3y84uIiIj7qPATEfET1lpmrC5iUKd4ggPd8/I+Lj2Fiuo6vlq21S3nFxEREfdQ4Sci4ic27qigYNdeRnRt+vF9+2S2b0lqXDiTczW7p4iIiC9R4Sci4iey8ooBGN6l6cf37WOMYVz/FGatLWZLyV63XUdERESalgo/ERE/kbW6iOTYcDrER7j1OuPSk7EWPl6w2a3XERERkaajwk9ExA/U1rmYs3YHI7omYIxx67Xax0eS2b4lk3MLsFZr+omIiPgCFX4iIn5gUcFuSqtq3drNs7Fx6Sms2V7G0sI9HrmeiIiIHB8VfiIifmDG6mKMgSGd4z1yvbN7tyEkKEBr+omIiPgIFX4iIn5gZl4xfVJiiY0I8cj1YiKCObV7K6Ys2kxNncsj1xQREZFjp8JPRMTHleytYWH+bkZ0cd8yDgcyLj2ZneXVTF9V5NHrioiIyNFT4Sci4uPmrN1BncsyLM2zhd+IronER4YweYG6e4qIiHg7FX4iIj5uZl4RkSGB9G/X0qPXDQ4M4Jx+bfl2+XZKKmo8em0RERE5Oir8RER8XNaaYgZ1iickyPMv6Rekp1Bd5+KzJVrTT0RExJu59V2CMeYMY8wqY0yeMeaeA9w/wRizxBiz0Bgz0xjTo2H7qcaYnIb7cowxJ7kzp4iIr9q0o4KNOyoY7uHxffv0bNuCrq2imJxb6Mj1RURE5Mi4rfAzxgQCTwNnAj2AS/cVdo28Y63tba3tBzwKPN6wvRgYa63tDVwNvOmunCIiviwrr35ileFdPbN+3/6MMYxLTyFn4y42FJc7kkFEREQOz50tfgOBPGvtOmttNfAecG7jHay1jVf+jQRsw/YF1tp9/YaWAeHGmFA3ZhUR8UlZq4tpGxNGp4RIxzKc1y8ZY2DyArX6SfP0/cptbC2pdDqGiMghubPwSwbyG31f0LDtZ4wxtxpj1lLf4nf7Ac5zAZBrra06wLE3GmOyjTHZRUWaTlxEmpfaOhez1hYzvEsixhjHcrSOCWNYWgKTcwtwuaxjOUScsHpbKde+ls3fp65wOoqIyCE5PrmLtfZpa21n4A/AnxvfZ4zpCTwC3HSQY1+w1mZaazMTE53p5iQi4pTFhSWUVtYyvKsz4/saG5eeTMGuvWRv3OV0FBGPem76WgC+XLZVs9uKiFdzZ+FXCKQ2+j6lYdvBvAect+8bY0wK8BFwlbV2rTsCioj4sqzVxRgDQzs7X/id3rM1ESGBTM7Vmn7SfBTsqmDKws0M6RxPda2LKYs1u62IeC93Fn7zgS7GmI7GmBDgEmBK4x2MMV0afXs2sKZheyzwOXCPtXaWGzOKiPisrDVF9E6OoWVkiNNRiAgJ4sxebfh88RYqa+qcjiPiES9lrQfgsfF96dY6mknZ+Yc5QkTEOW4r/Ky1tcBtwFfACuADa+0yY8yDxphzGna7zRizzBizEPgN9TN40nBcGnB/w1IPC40xSe7KKiLia0ora1iQv9uxZRwO5IL0ZEqravlm+Tano4i43c7yat6bv4nz+ifTNjacCzNSWFRQwuptpU5HExE5ILeO8bPWTrXWdrXWdrbWPtyw7X5r7ZSGr++w1va01vaz1o621i5r2P6QtTayYfu+23Z3ZhUR8SVz1u6gzmUZluY945sHdYqnbUyYuntKs/Da7A1U1riYMLITAOf3TyYowDBRrX4i4qUcn9xFRESOXtaaYiJCAklvH+t0lJ8EBBjO65/MjDXFbC/V1Pbiv8qranl99gZO69GKtKRoAOKjQjm5exIfLSikps7lcEIRkV9S4Sci4oNm5hUzqFM8oUGBTkf5mXHpydS5LFMWapIL8V/vzttEyd4aJozq/LPt4zNSKS6rZtoqLTElIt5HhZ+IiI/J31nB+uJyhqV5z/i+fdKSoumbEsPkXC3mLv6putbFS1nrGdQpjvR2LX9236gTEkmIClV3TxHxSir8RER8zMy8YgBGeMH6fQcyLj2F5Vv2sGLLHqejiDS5jxcWsnVPJTePSvvFfUGBAYxLT+b7ldspLqtyIJ2IyMGp8BMR8TFZa4poExNG58Qop6Mc0Ni+bQkKMHy0QK1+4l9cLstz09fSo00LRhxkRt3xGSnUuiwf6/dfRLyMCj8RER9S57LMytvBsLQEjDFOxzmguMgQRnern+SiVpNciB/5evk21hWVc/Oozgd9/nVpFU3f1FgmZhdgrfVwQhGRg1PhJyLiQ5YUllCyt4bhXb1nGYcDuSA9maLSKmat3eF0FJEmYa3l2elraR8fwZm9Wh9y34syU1i1rZQlhSUeSicicngq/EREfEjW6iKMwSsndmlsdLckYsKDtaaf+I0563awKH83N47oRFDgod8+je3bltCgACZm6/dfRLyHCj8RER+StaaYnm1bEBcZ4nSUQwoNCmRs3zZ8tWwrpZU1TscROW7PTltLQlQoF6SnHHbfFmHBnNGrNZ8sLKSyps4D6UREDk+Fn4iIjyirqiV30y6Gd/Hubp77jEtPobLGxRdLtzodReS4LC0sIWtNMdcN60hY8JGtnTk+I5U9lbV8s3ybm9OJiBwZFX4iIj5i7tod1Losw728m+c+/VNj6ZgQqe6e4vOenb6W6NAgLh/U7oiPGdI5nuTYcCbm6PdfRLyDCj8RER+RtaaIsOAAMjq0PPzOXsAYw7j+ycxdt5OCXRVOxxE5JuuLy/liyRauGNyeFmHBR3xcQIDhgvRkstYUsXn3XjcmFBE5Mir8RER8RFZeMYM6xRMadGRdzbzBef2TAbSmmfisF2asIygwgGuGdjjqYy/MSMVa1OotIl5BhZ+IiA8o2FXBuqJyr5/Nc3+pcRGc2DGOybmFWtNMfM72PZV8mFPA+IwUkqLDjvr4dvERDOoUx6QcreknIs5T4Sci4gNmrikGYISXr993IBekp7CuuJyF+budjiJyVF6etZ5al4sbR3Q65nOMz0hlw44K5m/Y1YTJRESOngo/EREfkJVXTKsWoXRJinI6ylE7s3drQoMCmJyr7p7iO0r21vD23E2c3act7eMjj/k8Z/ZuTVRoEBOz85swnYjI0VPhJyLi5epclll5xQxLS8QY43ScoxYdFszpPVvz6eLNVNVqTTPxDW/N3UhZVS0TRh57ax9AREgQZ/duw+dLtlBeVdtE6UREjp4KPxERL7dscwm7K2oY0dW3xvc1Ni49md0VNfywssjpKCKHVVlTx6uz1jOyayI928Yc9/nGZ6ZQUV3H1CVbmiCdiMixUeEnIuLlshrG9w31sYldGhuWlkBidKhmNxSfMDGngOKyam4e1blJzpfRviWdEiKZmK3ffxFxjgo/EREvN2N1ET3atCAhKtTpKMcsKDCA8/q15YdV29lZXu10HJGDqq1z8cKMtfRvF8uJHeOa5JzGGC7ISGHehp1sKC5vknOKiBwtFX4iIl6svKqW3E27GO7D3Tz3GZeeQk2d5bPFm52OInJQny/ZQv7Ovdw8snOTjqm9ID2FAAOTctTqJyLOUOEnIuLFfly/g5o6y4guvreMw/66t2lB9zYt+FCze4qXstby7LS1pCVFcUr3Vk167tYxYYzomsiHuQXUubSmn4h4ngo/EREvNmN1MaFBAWS0b+l0lCZxQXoyi/J3k7e9zOkoIr8wbXURK7eWMmFkZwICmn4G3fEZqWwpqWRWXnGTn1tE5HBU+ImIeLGZecWc2CmesOBAp6M0iXP6tSXAwEcL1N1NvM+z09bSNiaMc/q2dcv5T+mRRGxEMBPV3VNEHKDCT0TES23evZe87WWM6OL74/v2SYqu7+72UW4hLnV3Ey+Ss3En89bv5PrhnQgJcs/bo9CgQM7t25avlm2lpKLGLdcQETkYFX4iIl5qZsMyDsP8qPCD+kleNpdUMnf9DqejiPzk2WnriI0I5pKBqW69zvjMVKprXUxZpLGuIuJZKvxERLxUVl4xidGhnNAq2ukoTeq0Hq2IDg1isiZ5ES+xelsp367Yxq+GdCAiJMit1+rZtgXdWkeru6eIeJwKPxERL+RyWWauKWJ4l4QmnVLeG4QFB3JW7zZ8sWQLFdW1TscR4bnpawkPDuTqwR3cfi1jDBdlprK4oIRVW0vdfj0RkX1U+ImIeKFlm/ewq6KG4X7WzXOfcenJlFfX8fWybU5HkWauYFcFUxZu5tKB7WgZGeKRa57XP5ngQMPE7HyPXE9EBFT4iYh4pay8IgCGpvln4TegQxwpLcP5MFfd3cRZL2WtB+D64R09ds24yBBO7taKjxcWUlPn8th1RaR5U+EnIuKFslYX071NC5Kiw5yO4hYBAYZx/ZOZlVfM1pJKp+NIM7WzvJr35m/ivP7JtI0N9+i1x2emUFxWzQ8rt3v0uiLSfB208DPG/Ga/213GmCuNMZ77SExEpBmqqK4lZ+Muv+3muc/56Sm4LHyyUJO8iDNem72ByhoXE0Z28vi1R3ZNJDE6VJO8iIjHHKrFL3q/WwsgE/jCGHOJB7KJiDRLP67fSXWdy+8Lv44JkaS3i+XD3AKs1Zp+4lnlVbW8PnsDp/VoRVqS52fODQoMYFz/ZL5fuZ2i0iqPX19Emp+DFn7W2r8e4HYHMAS423MRRUSal6zVxYQEBTCgQ5zTUdxuXHoKq7eVsWzzHqejSDPz7rxNlOytYcKozo5lGJ+ZQp3L8vECtXqLiPsd9Rg/a+1OwL/mFhcR8SIz84o4sWMcYcGBTkdxuzF92hASGKA1/cSjqmtdvJS1nkGd4khv19KxHGlJ0fRvF8vEnHy1eouI2x114WeMGQ3sckMWEZFmb2tJJau3lfl9N899YiNCOLl7ElMWaXZD8ZyPFxaydU8lN49KczoK4zNSWb2tjMUFJU5HERE/d6jJXZYYYxbvdysAHgFu8VxEEZHmI2tN/TIOw9ISHU7iOePS62c33Pezi7iTy2V5bvpaerRpwQgv+IBlTN82hAUHMDFHa/qJiHsdqsVvDDC20W0McIK1dqC1duWRnNwYc4YxZpUxJs8Yc88B7p/QUGAuNMbMNMb0aHTfHxuOW2WMOf3ofiwREd80M6+YhKhQurX2/GQTThnZNZG4yBA+VHdP8YCvl29jXVE5N4/qjDHOj1xpERbMGT1bM2XhZipr6pyOIyJ+7FCTu2y01m4EOlFf+J0DDDzSExtjAoGngTOBHsCljQu7Bu9Ya3tba/sBjwKPNxzbA7gE6AmcATzTcD4REb/lcllmrilmeJcEAgKcf0PqKSFBAZzTty3fLN9Gyd4ap+OIH7PW8uz0tbSPj+DMXq2djvOT8Zmp7Kms5evl25yOIiJ+7FBdPZONMT8Cf6G++OsE/MUYM88Yk3wE5x4I5Flr11lrq4H3gHMb72CtbTyNWySwb2TzucB71toqa+16II+jKDpFRHzR8i172FFezbA057ufedq49GSqa11MXbLF6Sjix+as28Gi/N3cOKITQYFHPc2B2wzuFE9ybDgTs9XdU0Tc51Cvek8Bz1prR1prf9NwG9mw/ZkjOHcy0PgVrKBh288YY241xqylvsXv9qM89kZjTLYxJruoSGNDRMS3zcwrBmg2E7s01js5hrSkKCbnajFrcZ9np60lISqUC9JTnI7yMwEBhgsyUpiZV0zh7r1OxxERP3Wowq+Htfa1/Tdaa98AujVVAGvt09bazsAfgD8f5bEvWGszrbWZiYnNZyIEEfFPWWuK6NY6mqQWYU5H8ThjDOPSk5m/YRcbd5Q7HUf80NLCErLWFHPdsI5euVTK+IwUrIXJOfrwQ0Tc41CF3wHvM8YEAEfyilkIpDb6PqVh28G8B5x3jMeKiPi0vdV1zF+/q1l289znvH7JGAMfaTFrcYNnp68lOjSIywe1czrKAaXGRTC4UzyTcgu0pp+IuMWhCr/PjDEvGmMi921o+Po5YOoRnHs+0MUY09EYE0L9ZC1TGu9gjOnS6NuzgTUNX08BLjHGhBpjOgJdgHlHcE0REZ80b8NOqutcDO/afHsvtI0NZ0jneCbnFuqNrzSp9cXlfLFkC5cPak+LsGCn4xzU+MwUNu6oYN76nU5HERE/dKjC726gBNhojMkxxuQCG4A9wO8Od2JrbS1wG/AVsAL4wFq7zBjzoDHmnIbdbjPGLDPGLAR+A1zdcOwy4ANgOfAlcKu1VnMci4jfylpdREhQAAM7xDkdxVHj+qewaWcFORt3OR1F/MgLM9YRFBjAtUM7OB3lkM7s1Yao0CAmqruniLhB0MHusNbWAL8zxtwHpDVsXmutrTjSk1trp7Jf66C19v5GX99xiGMfBh4+0muJiPiyrDXFDOjQkvAQ7xt75Eln9GrNnz9eyoe5hWQ28yJYmsb2PZV8mFPAhZkpXj9+NjwkkDF92jBl0Wb+ek5PIkMP+jZNROSoHWo5h3HGmHHUr8PXpeF2RqPtIiLSBLbvqWTVtlKGd2m+3Tz3iQwN4sxerflssRazlqbx8qz11Lpc3Di8k9NRjsj4zBQqquv4XEubiEgTO9RHSWMPcZ8FJjdxFhGRZilrTfNdxuFAxqWnMHlBId+u2MaYPm2djiM+rGRvDW/P3cRZvdvQISHy8Ad4gfR2LemUGMnE7Hwuykw9/AEiIkfoUF09r/FkEBGR5mpmXjHxkSF0b93C6SheYXDneFq3CGNybqEKPzkub83dSFlVLRNGdnY6yhEzxjA+I5VHvlzJ+uJyOvpIwSoi3u9Qk7uIiIibuVyWrDXFDOuSQECAcTqOVwgMMJzXP5npq4soKq1yOo74qMqaOl6dtZ4RXRPplRzjdJyjMi49mQADk3LynY4iIn5EhZ+IiINWbi2luKxK4/v2My49mTqXZcqizU5HER81MaeA4rJqbvah1r59WrUIY2TXRD7MKaTOpaVNRKRpqPATEXHQzLwigGa9cPuBdG0VTe/kGCbnalp7OXq1dS5emLGWfqmxDOrkm7PDjs9MZeueSmbmFTsdRUT8xBEVfsaYIcaYy4wxV+27uTuYiEhzkLWmmK6tomgd493TzDthXHoyyzbvYeXWPU5HER/z+ZIt5O/cy82jOmOMb3ahPrl7ErERwUzMVndPEWkahy38jDFvAo8Bw4ABDbdMN+cSEfF7lTV1/Lh+p7p5HsTYvm0JCjB8lFvodBTxIdZanp22ls6JkZzavZXTcY5ZaFAg5/VL5uvl29hdUe10HBHxA0fS4pcJDLXW3mKt/XXD7XZ3BxMR8XfzN+ykutbFMC3jcEAJUaGMOiGRjxZonJMcuWmri1i5tZQJIzv7/IRJF2akUF3r0lhXEWkSR1L4LQVauzuIiEhzk7WmmJDAAE7s6JtjkDxhXHoK20urmLZqu9NR/MaM1UXc+9ESv50x9dlpa2kTE8a5/ZKdjnLceiXH0KNNCyZma6yriBy/Iyn8EoDlxpivjDFT9t3cHUzkQKpq69hbXed0DJEmMWN1ERntWxIRctAlVZu9k7ol0T4+gnsmL2FrSaXTcXzeqq2lTHgrh7d/3MRpT0znUz9rScrZuJN563dy/fBOhAT5x/x14zNTWFJYorGuInLcjuRV8S/AecDfgX83uol43E1v5nD+M7OorXM5HUXkuGwvrWTl1lKGd1U3z0MJCw7khSszKa+q5aa3cqis0Qc/x2p3RTU3vJFNZGgQ71x/Iu3iI/n1uwu49e1cdpT5R+vfs9PWERsRzCUDUp2O0mTO7ZdMcKBRq5+IHLfDFn7W2ukHunkinEhjK7bsYdqq+rEb78zb5HQckeMyq2GK9hGa2OWwTmgdzeMX9WNR/m7+/PFSrNV4v6NVW+fi1+8uYGtJJc9dkcGQtAQ+nDCYu884gW+Wb+O0J2bw5dItTsc8Lqu3lfLtim1cPbgDkaH+04oeFxnCKd1b8fGCQmr0oaeIHIcjmdVzkDFmvjGmzBhTbYypM8aov4F43Kuz1hMeHEh6u1ge/2Y1JRU1TkcSOWZZq4uJiwyhR5sWTkfxCWf0as3tJ3dhUk4Br83e4HQcn/PIlyvJWlPMQ+f1IqN9SwCCAgO4ZVQan/56GG1iw5jwVi63v7uAXeW+OYPkc9PXEh4cyNVDOjgdpcmNz0xhR3k136/UWFc5NhuKyzU7rBxRV8+ngEuBNUA4cD3wtDtDieyvuKyKjxdu5oKMZB4+vzd79tbwn+9WOx1L5JhYa8nKK2ZoWoLPzzroSXee3IVTe7Tioc9XMFuLWh+xjxYU8GLWeq4e3J6LDtAF8oTW0Xx0y1B+c2pXpi7ZwqlPzOCb5dscSHrsCnZVMGXhZi4ZmEpcZIjTcZrciC6JJEWHqrunHJV1RWU8+d0aTn18OqMem8agf3zH/Z8sZeOOcqejiUOOaOSztTYPCLTW1llrXwXOcG8skZ97e+4mqmtdXDO0I93btOCSge14c85G8raXOR1N5Kit2lZKUWkVw7WMw1EJCDA8flFfOiZEcss7ueTvrHA6ktdbXLCbP3y4hEGd4vjzmB4H3S84MIDbT+7CJ7cNJTE6lBveyOY3Hyz0mZ4VL2WtB+D64Z0cTuIeQYEBnJ+ezA+rtrO9VJMcycHl76zg2WlrOeu/WZz07+k88e1qWkaE8MDYHpzTty3vzctn1GPTmPBmDjkbdzkdVzzsSAq/CmNMCLDQGPOoMeauIzxOpElU1dbx5tyNjD4hkc6JUQD85tSuhAcH8vDnyx1OJ3L0Zq6pb61S4Xf0osOCefGqTFwuyw1vZFNRXet0JK9VVFrFTW/mkBgVytOXpRMcePg/3T3bxvDJrUO5/aQ0Plm4mdP+M50fvHwpjZ3l1bw3fxPn9ksmOTbc6ThuMz4jlTqX5eMFhU5HES+zpWQvL2Wt49ynZzH80R945MuVhAQFcN+YHsy552Q+mDCYa4Z25NEL+zLzD6O5ZVRn5qzbwQXPzmbcM7P4cukWrZXaTBxJAXdlw363AeVAKnCBO0OJNPbpoi0Ul1Vx7bCOP21LiArl9pO78MOqIq3vJT5nxppi0pKiaBPjv29S3aljQiRPXtqf1dtK+f3ExZrs5QCqa13c/FYOuyqqeeGqDOKjQo/42JCgAH5z2gl8fMtQYsKDuebV+fxh0mL2VHpn699rszdQWeNiwkj/bO3bJy0pivR2sUzMLtDvvLC9tJLXZ29g/HOzGfyP73no8xXU1rm458xuZN09mo9vHcp1wzrSOibsZ8cltQjj96d3Y/Y9J/GXsT0oKqtiwlu5nPTvabw5Z4OWzPJz5khePIwx4UA7a+0q90c6NpmZmTY7O9vpGNLErLWc/eRMal0uvrpzBMb8/3io6loXpz0xnaDAAL64Y/gRfZot4rTKmjr6/vVrLjuxHQ+M7el0HJ/2/PS1/OOLlfz+9BO4dXSa03G8yp8+WsI7P27if5f2Z2zftsd8nqraOv777Rqem76W1i3CeOTCPgz3oploy6tqGfLP7xnYMY4Xr8p0Oo7bvTtvE3+cvISPbx1Kv9RYp+OIh+0sr+bLpVv5bPFm5q7bgctC11ZRjO3TlrP7tKFTQ6+oo1Hnsny1bCsvzFjHwvzdxEYEc+Wg9lw1uAOJ0Uf+gZF4D2NMjrX2gC+IRzKr51hgIfBlw/f9tIC7eMrcdTtZvmUP1w7t+LOiD+o/lb737B7kbS/j7bkbHUoocnRyNu6iqtalbp5N4MYRnTinb1se+3oV36/0rclI3OntHzfyzo+bmDCy83EVfQChQYHcfUY3Prx5COEhgVz58jz+9NESyqq8o4vtu/M2UbK3hptHdXY6ikeM6dOGsOAAJmbnOx1FPKRkbw0Ts/O56pV5DHj4W/700RK2llRy2+g0vr5rBF/fNZJfn9zlmIo+gMAAw1m92/DRLUOYNGEwAzvE8dQPeQx95Hv+MGkxa7aVNvFPJE46koVu/gIMBKYBWGsXGmM6HuoAkabyyqz1xEWGcF7/5APef0r3JIalJfDEt2s4t18yLf1wNjfxLzPWFBEcaDixY7zTUXyeMYZHLujD2qIy7nh3IR/dOpS0pGN78+Mv5m/YyQOfLGPUCYn8/vQTmuy8/du15PPbh/P4N6t5MWsdM1YX8eiFfRjS2bkPMKprXbyUtZ4TO8aR3q6lYzk8KTosmDN7tWHKos3cN6YHYcGBTkcSNyirquXb5dv4bPFmZqwuprrORUrLcG4Y3okxfdrQs22LX3wYfryMMWR2iCOzQxzrisp4eeZ6JuUU8H52Pid1S+KG4Z0Y1Cmuya8rnnUkfeNqrLUl+21T53Jxuw3F5Xy7YhuXn9juoH/cjDH8eUx3Sitr+O93azycUOToZa0uJqN9S79aYNpJ4SGBvHBVJiFBAdz4RjYle71zHJonbN69l5vfyiE1LoL/XtKfwCZeKiQsOJA/ndWdiTcNJijAcNmLP/LAJ0sdm2Dn44WFbN1T2Wxa+/YZn5FCaWUtXy3b6nQUaUJ7q+v4fPEWbn4rh4y/fcOd7y9kaeEerhzcno9vHUrW3aO558xu9EqOcXvx1SkxiofP783se07irlO6sih/N5e+OJexT83kk4WF1NS53Hp9cZ8jKfyWGWMuAwKNMV2MMf8DZrs5lwivzd5AUIDhykHtD7lft9YtuOzEdrw5d6O6JIhXKy6rYvmWPV41RsofJMeG88zl6WzaWcGd7y1olrPTVdbUcdObOVTWuHjxqgxiwoPddq3MDnF8cccIrhnagdfnbOTM/2Yxb/1Ot13vQFwuy/PT19K9TQtGdm1ez6dBneJJaRmuNf38QFVtHV8v28rt7y4g46FvuPWdXOZv2MnFA1KZOGEws+85ifvG9KBfaqwjLW3xUaHccUoXZt1zEv8Y15uK6jrueG8ho/41jZey1lHqpRM+ycEdSeH3a6AnUAW8C+wB7nRjJhH2VNb3aR/bpy1JLcIOu/9dp3QlIiSQhz5f4YF0IsdmVp6WcXCXEzvF88DYHvywqoh/f+2185C5hbWWP05ewtLNJfzn4n6kJUW7/ZrhIYE8MLYn7904CJe1XPzCHP722XIqazwzI+A3K7axtqicm0d1bnZdzwICDBdmpDBrbTGFu/c6HUeOUk2dix9Wbee3Hywi82/fcuObOcxYU8S5/dryzvUn8uOfTuHBc3sxoEMcAU3can+swoIDuXRgO769ayQvXZVJcstwHvp8BUP+8T3/mLqCLSX6PfQVh+1rZK2tAO5tuIl4xAfz8ymvrvvZEg6HEh8Vyh0nd+Ghz1fww8rtjO6W5OaEIkdvxupiWkYE07NtjNNR/NIVg9qzfMsenpm2lh5tWzCmz/FNbOIrXp65no8WFPLbU7tySo9WHr32oE7xfHnHCP75xUpenrmeH1Zu51/j+5LR3n1j7qy1PDNtLe3iIjirV2u3XcebXZCewn++XcOHOQXcfnIXp+PIYdTWufhx/U4+XbSZL5dtZXdFDdFhQZzeqzVj+rRhaFqCT8xMHhBgOKVHK07p0YpF+bt5MWsdL2at4+WZ6xnbty3XD++ov29e7qCF3+Fm7rTWntP0cUTqXyBfnbWBgR3j6JV85C8gVw3uwDs/buJvny9nWBffeBGV5sNay8y8IoakJTT52CupZ4zhr+f0YvW2Mn4/cTGdEqLo0baF07HcKmtNEX+fuoIze7XmtpOcWdIiMjSIv53XizN6tebuSYsZ/9xsbhjRibtO6eqWyUfmrtvJovzdPHReL4Ka6et8alwEQzrHMymngNtGp3lNy5D8P5fLMn/DTj5bvIUvlm6huKyaiJBATu3RijF92jKiawKhQb47OU/f1Fieuiyd/J0VvDprA+/N38RHCwoZlpbA9cM7MrJrYrNrjfcFB13HzxhTBORT373zR+Bnj561drrb0x0FrePnP75YsoWb387luSsyOOMoP839bsU2rns9m/vH9Dji1kIRT1i9rZTTnpjBIxf05uIB7ZyO49e2l1Zyzv9mERhg+PTXw4jz09l+N+4o55ynZtEmJowPbx7iFRMGlVbW8PepK3l33ibSkqL49/i+9G3i9eauemUeyzfvYeYfRjfrWS0/WlDAXe8v4r0bBzGok2YJ9gbWWhbk7+azRVuYumQLW/dUEhoUwMndkxjTpy2jT0giPMQ/f2dLKmp4Z94mXp21nu2lVZzQKprrh3fknH5tfbrA9UXHuo5fa+BPQC/gv8CpQLG1drq3FX3iX16ZtZ7UuHBOPYYuSyd1S2J4lwT+8+1qdpZXuyGdyLGZsboIgGGa2MXtkqLDeP7KDIrKqrj17Vy/nIGurKqWG97Ixhh44cpMryj6oH65gX+M683r1w6krLKWcc/O5rGvVlFV2zRj/5YWljBjdRHXDuvQrIs+gDN6tiE6NEiTvDjMWsvSwhL+8cUKhj/6A+Oemc1bczfSKzmG/17Sj5z7TuWZyzM4q3cbvy36AGIigrl5VGdm/uEkHhvfF2Pg95MWM+yRH3j6hzxKKjQRjDc4aOFnra2z1n5prb0aGATkAdOMMbd5LJ00O4sLdjN/wy5+NaTjMXWHM8Zw35gelFfX8Z9vV7shocixmZlXTKfESJJjw52O0iz0TY3lH+f3Zs66HTzsZ5M+uVyW336wkLztZTx1aTrt4iOcjvQLI7sm8tVdIzi/fzJP/ZDHuU/NYmnh/itDHb3npq8lOjSIKw4z23NzEB4SyJi+bZi6ZAtlVc4sqdHcrdpayqlPzGDM/2byctZ6OidG8dj4vsz/8ym8dHUm5/ZLJspLPpTxlJCgAC7MSOGLO4bzxrUD6dY6mn99tYrB//yOv0xZxqYdFU5HbNYO2TneGBNqjBkHvAXcCjwJfOSJYNI8vTJzPVGhQVyUmXLM5+jaKprLT2zH2z9uYrWWdxAvUFVbx9x1Oxih1j6PuiAjhWuHduS12Rv4IDvf6ThN5n/f5/HVsm386azuDPPiGWJjwoN5bHxfXr46k53l1Zz39Cz+8+3qY26B3VBcztQlW7h8UHtahLlvuQpfcmFGKntr6vh88WanozQ7uZt2cdHzcyitrOEf43oz/95TeP3agVyYkeLW5VR8hTGGEV0TefO6E5l6+3DO6NWat+ZuZNRjP3Dr27ks2LTL6YjN0kELP2PMG8AcIB34q7V2gLX2b9baQo+lk2Zla0klny3ewkWZqUQf5x/1u07pSmRIIH/7bDkHG8cq4ik5G3ZRWeNiWJr3vkn3V386qxtD0+L580dLyfWDNxpfL9vKE9+uZlz/ZK7zkXHMJ3dvxdd3jWBs37b859s1nPvULFZs2XPU53khax1BgQFcO7RD04f0UentYumcGKnunh42c00xV7z0I7ERwUyaMIRLB7ajpZ+OJW4KPdq24PGL+pH1h9HcMKITM9YUcf4zsxn/3Gy+XrYVVzNce9Uph2rxuwLoAtwBzDbG7Gm4lRpjjv4VW+Qw3py7AZe1XNMEf9RbRoZw5yldyVpTzPcrtx9/OJHjkJVXTFCAYVBnTcDgaUGBATx1aTqtYkKZ8GYO2/ZUOh3pmK3ZVspd7y+kT0oMfx/X26dmzIuNCOGJi/vx/JUZ9ZPvPDWTp75fQ+0Rtv5t31PJpOwCLsxIOaK1XZsLYwzjM1PJ3riLdUVlTsdpFr5cuoVrX5tPu7gIJt40mNQ47+tq7a3axITzxzO7M+ePJ3PfmB5s3l3JjW/mcPLj03lr7kaPrQPanB1qjF+AtTa64dai0S3aWuvf82OLx+2truOdHzdxao9WTfYieuXg9nRKjOShz1dQXet/kzuI78haU0R6+5bNbqyHt2gZGcKLV2VSVlXLhLdymmyiEU8qqajhhjeyCQ8J4vkrM3x2YpPTe7bm67tGcnrP1jz29WrGPTubNUfQJf+VWRuodbm4cXgnD6T0LeP6JxMYYJiUo1Y/d/tgfj63vJ1Lr+QWvH/jYH0IcYyiQoO4blhHpv9+FP+7tD9RoUH8+eOlDPnn9zz+zWqKy6qcjui3mucCOOJ1PlpQyK6KGq4b1nR/1IMDA7jv7B6sLy7njTkbmuy8IkdjR1kVyzbvYbi6eTqqW+sW/Ht8XxZs2s19Hy/1qS7gdS7L7e8toHD3Xp67Ip02Mb49QVBcZAhPXZbO05elU7BrL2c/OZPnpq+l7iDdvfZU1vD23I2c1bsNHRIiPZzW+yW1CGNk10Qm5xYe9P9Qjt9LWeu4+8PFDE1L4K3rTyQmQuP4jldQYABj+7Zlym1Dee/GQfRPjeXJ79Yw8tH6mUDVAtj0VPiJ46y1vDJrPb2SWzCgQ8smPffobkmM7JrIf79bww59giQOmLV2B9bC8K6a2MVpZ/Zuw69PSuOD7ALemLPR6ThH7NGvVjJ9dREPntuLzA5xTsdpMmf3acPXd43gpG5J/POLlVz43GzWHqC74ltzN1JaVcuEkZ0dSOkbxmeksHVPJVlripyO4nestTz21Soe+nwFZ/duw0tXZxIRot4bTckYw6BO8bz8qwF8+5sRDElL4F9freLUJ6bzxZItPvVBnbdT4SeOm7GmmLztZVw7tKNbxqzcN6Y7FdV1PKHlHcQBWauLiAkPpndyjNNRhPqJn07pnsSDny1n9tpip+Mc1icLC3l++jquGNSOSwe2czpOk0uICuXZK9L57yX9WFdUzln/zeKlrHU/tVxV1tTxyswNjOiaSC89hw7q5O6taBkRzER192xSLpfl/k+W8dQPeVwyIJUnL+2vxcjdLC0pmhevyuTt608kIjiIm9/O5eIX5jbJcjDi5sLPGHOGMWaVMSbPGHPPAe7/jTFmuTFmsTHmO2NM+0b3PWqMWWaMWWGMedL40ih2OSqvzFxPYnQoY/q0dcv505KiuXJQe975cRMrt2peIvEcay0z84oZmhZ/TOtSStMLCDA8cXE/OsRHcOvbueTv9N41pZYWlvCHDxczsEMc94/p6XQctzHGcG6/ZL65awTDuyTw0OcruOSFOWwoLmdSTgHFZVXcrNa+QwoJCqj/P1y2jd0V1U7H8Qs1dS7u+mAhb87dyE0jOvGPcb31Ou5BQ9MS+Pz2YTx8fi/ytpcx9qmZ/GHSYraX+u4EXd7AbYWfMSYQeBo4E+gBXGqM6bHfbguATGttH2AS8GjDsUOAoUAfoBcwABjprqzinLztpUxfXcRVg9oTEuS+zyHuPKUL0WHBWt5BPGptURlbSioZrvX7vEp0WDAvXpVJrcty45s5VFR73+LXxWVV3PRmDnERITxzRbpbXx+9RVKLMF68KpN/j+/Lyq2lnPHfGTzxzWr6pcYyqJP/dHF1l4syU6muc/HJQq3pd7z2Vtdx05s5fLJwM3efcQJ/PKu7T82i6y+CAgO4/MT2/PC7UVw3tCMf5hZw0mPTeXbaWo3/O0bu/EsyEMiz1q6z1lYD7wHnNt7BWvuDtXbfx61zgX2rdlsgDAgBQoFgYJsbs4pDXpm1gdCgAC470b1dmGIjQrjrlC7MytvBtyu0vIN4xozV9V0JtX6f9+mUGMWTl/Zn5dY93D1psVd9IFRT5+KWt3MpLqvi+SszSYgKdTqSxxhjuCAjhW/uGsmgTvHsKK/mttFpetN9BHq0bUHPti2YmJPvdBSftqeyhqtfmccPq7bz8Pm9uGVUmtORmr2Y8GD+PKYHX981gkGd4nnky5Wc+sR0vlyq8X9Hy52FXzLQ+NWnoGHbwVwHfAFgrZ0D/ABsabh9Za1dsf8BxpgbjTHZxpjsoiINaPY1u8qrmZxbwPn9k4n3wBubywe1Jy0pioc/X+6T07mL75mZV0zHhEit8+SlRp+QxN2nd+OzxVt4dvpap+P85MFPlzNv/U4evbAPvVOa57i21jFhvPqrAWTdPZpTerRyOo7PGJ+RwtLCPazYomENx2JHWRWXvTiX3E27ePKS/lx+YvvDHyQe0ykxipeuzuTN6wYSHhzIhLdyueSFuSzbrPF/R8or+o4YY64AMoF/NXyfBnSnvgUwGTjJGDN8/+OstS9YazOttZmJiepK5WvembeJyhoX1w7r6JHrBQcGcN+YHmzYUcEbs31nRj/xTVW1dcxZu4PhXdTa580mjOzE2L5t+ddXq/hhpfO9Ad6bt4k3527kxhGdOLffoT4r9X/GGH1ocpTO7ZdMSGAAE7M1ycvRKty9l/HPzyFvexkvXp3J2L7umXdAjt/wLolMvX04fzuvF6u3lTLmfzO558PFFJVq9vbDcWfhVwikNvo+pWHbzxhjTgHuBc6x1u57xM4H5lpry6y1ZdS3BA52Y1bxsJo6F2/M2cDwLgl0bRXtseuO7JrI6BMSefK7NVogVNwqd+Nu9tbUaXyflzPG8OgFfejeugW3v7fggMsJeErOxp3c98lShndJ4A9ndHMsh/iulpEhnNIjiY8XFlJd63I6js9YW1TG+GdnU7SnijeuPZHRJyQ5HUkOIygwgCsHtWfa70dz7dCOTMopYPRj03hu+lr16joEdxZ+84EuxpiOxpgQ4BJgSuMdjDH9geepL/oaf9S6CRhpjAkyxgRTP7HLL7p6iu+aumQL2/ZUeay1r7F7z+7B3po6/v21lncQ95mZV0RggNGkFD4gPCSQF67KIDgwgBveyGZPZY3HM2wtqWTCW7m0jQ3nqUvTNXugHLPxGansLK/mey9owfYFSwtLuOi5OVTVunj3xkEM7KjXbF8SEx7MfQ3j/07sGMc/v1jJqY/P4MulWzX+7wDcVvhZa2uB24CvqC/aPrDWLjPGPGiMOadht38BUcBEY8xCY8y+wnASsBZYAiwCFllrP3VXVvEsay0vz1xPp8RIRjrQGpKWFMWVg9vz/vxNLN+scRDiHllriklvF0t0WLDTUeQIpLSM4OnL0tm4o4K73luIy+W5NwyVNXXc9GY2FVW1vHhVJjER+p2RYze8SwJJ0aFM0iQvh/Xjuh1c+sJcwoIDmThhsNaK9GGdEqN4+VcDeOPagYQFBzDhrRwufXGu3uftx61j/Ky1U621Xa21na21Dzdsu99aO6Xh61Osta2stf0abuc0bK+z1t5kre1ure1hrf2NO3OKZ+Vs3MXighKuGdqRAIc+1b7z5K60CNfyDuIeu8qrWVJYom6ePmZw53geGNuD71Zu5/FvPNMjwFrLvR8tZVFBCY9f3M+jXd/FPwUFBnBBRgo/rCrSmmeH8P3KbVz1yjySWoQyccJgOiVGOR1JmsCIrg3j/87tyaqtpZz9vyz+OHmxhvc08IrJXaR5eWXWemLCg7kg3bmJC2IigvnNqV2Zs24HXy/XSiHStGatLcZaGKaJXXzOlYPac3FmKk/9kMfUJVvcfr1XZ23gw9wC7jylC6f3bO3260nzMD4jhTqX5aPcX0ytIMAnCwu58Y0curaK5oObBtM2NtzpSNKEggIDuHJwB6b9bjTXDOnIxOwCRv1rGs9r/J8KP/Gs/J0VfLl0K5cObEdESJCjWS4b2I6uraL4+9QVzf6FQJpW1upiWoQF0UfdhnyOMYYHz+tJertYfvvBIrdOiz8rr5iHp67gtB6tuP2kLm67jjQ/nRKjyGjfkok5BerVsp83527kzvcXktG+Je/ccKJHlpMSZ8REBHP/2B58eecIBnaM4x9frOS0J2bw1bLmO/5PhZ941BtzNmCM4eohzq+NE9SwvMPGHRW8NmuD03HET1hrmZlXzNC0BIIC9RLri0KDAnnuigxahAdxwxvZ7CyvbvJr5O+s4NZ3cumcGMnjF/dzrNu7+K/xGSnkbS9jYf5up6N4BWstT32/hvs+XsrJ3ZJ4/dqBGoPdTKQlRfHKrwbw+rUDCQ4M4KY3c7j8pR+b5XqXelciHlNWVct78/I5q3cb2sR4R7eK4V0SOblbEv/7Pk/rv0iTWFdcTuHuverm6eOSWoTx3BUZbN9TxW3v5FJb13RT41dU13LDG9m4XJYXrswkKtTZ3g/in87u04aw4AAm5mhNP2stf5+6gse+Xs35/ZN59ooMwoIDnY4lHjayayJf3jGcB8/tyfItezj7ySz+OHlJsxr/p8JPPGZSdj6lVbVc58ASDody79ndqayp499fr3I6iviBrNVFAIzQxC4+r3+7ljx8fi9mr93Bw1ObZkUhay2/m7iI1dtKeeqydDokRDbJeUX2Fx0WzFm92vDpos1U1jTf4Qx1Lss9Hy7hxaz1XD24Pf8e35dg9cZotoICA7hqcAem/W4UVw3uwAfZ+Yz+1zRenLGuWax9qd988QiXy/Lq7A2kt4ulX2qs03F+plNiFFcP6cD72fks21zidBzxYdW1Lt6cu5G0pChS4yKcjiNNYHxmKr8a0oFXZ21gUhO0nDz9Qx5Tl2zlnjO7MaKrPhwQ97owM4XSylq+WrbV6SiOqKqt47Z3cnk/O5/bT0rjL+f0VLdqASA2IoS/nNOTr+4cTmaHlvXjrZ+Yztd+Pv5PhZ94xHcrt7NxRwXXDevkdJQDuv3kLrSMCOHBT7W8gxy7F7PWsbaonHvP6u50FGlC957dnSGd4/nTR0uOa7zUdyu28e9vVnNev7bcMNw7XwvFvwzqGE9qXDgfZDe/Nf3Kq2q5/vVsvli6lT+f3Z3fnHYCxqjok59LS4rm1WsG8to1AwgMMNz4Zg5XvPwjK7f65/g/FX7iEa/MXE/bmDBO79nK6SgHFBNev7zDj+t3NttPRuX45O+s4H/fr+GMnq0Z3S3J6TjShIIDA3jqsnSSokO56c1stu85+rXR8raXccd7C+nZtgX/vKCP3oCKRwQEGC5MT2X22h0U7KpwOo7H7K6o5oqXf2RWXjGPXtiH6/VBixzGqBOS+PLOEfxlbA+WFu7hrP9mce9HS9jhZ+P/VPiJ2y3bXMKcdTu4ekgHr57l8JIBqZzQKpqHp65o1uMh5Nj89dPlBBjD/WN7OB1F3CAuMoQXr8pkz95aJryVc1RLwJTsreHGN7IJDQrg+SszNamEeNQFGfVr5n6Y0zzW9Nu+p5KLn5/LssI9PHN5BhdlpjodSXxEcGAAvxrakem/rx//9978fEY9No2Xsvxn/J/3vgsXv/HqrA1EhARyyYB2Tkc5pKDAAO4f24P8nXt5ZdZ6p+OID/lm+Ta+XbGNO07uooWA/Vj3Ni14bHxfcjft5v6Plx1Rt/A6l+XO9xawaWcFz16RQbJ+P8TDUlpGMKRzPJNy83G5/Hsow6YdFVz43Bzyd1Xw6jUDOKNXa6cjiQ9qPP4vvV1LHvq8fvzfN8u3+fxwIBV+4lbbSyuZsnAzF2akEBPh/evlDE1L4JTurXj6+zy2lx59dy5pfiqqa/nLlGV0bRXFtV42Y600vbP7tOHW0Z15PzufN+duPOz+//56FT+sKuKBc3oysGOcBxKK/NL4jFTyd+7lx/U7nY7iNqu2lnLhc7Mp2VvD29efyNA0LakjxyctKZrXrx3Iqw3j/254I5srX57n0+P/VPiJW709dxPVdS6uGeo7b4jvPbs71XUuHvtKyzvI4T31fR6Fu/fy0Hm9NUV4M/HbU0/gpG5JPPjpcuau23HQ/T5bvJlnpq3l0oHtuOJE7+7xIP7t9J6tiQ4NYmKOf07ysmDTLi56fg4AH9w0mP7tWjqcSPzJ6Ibxfw+M7cHigt2c9d8s/vyxb47/07sUcZvKmjremruRk7sl0dGH1qrqmBDJNUM7MjGngKWFWt5BDi5veykvZq3jgvQUteY0IwEBhv9c0o928RHc8nbuASfNWL55D7+fuJjM9i356zk9NZmLOCo8JJAxfdvyxZKtlFbWOB2nSc1cU8zlL/1ITHgwH948hBNaRzsdSfxQcGAA1wztyPTfj+bKQe15d179+D9fe5+owk/cZsqizewor/a6BduPxG0npRGn5R3kEKy13PfxMsKDA/njWd2cjiMe1iIsmBevyqSm1sVNb+awt/r/J3vZWV7NDW9kExMezDNXpBMSpD+14ryLMlPYW1PH54u3OB2lyXy5dCvXvjafdnERTJowWOunitu1jAzhr+f24ss7hjOmTxu6tvKtDxr010jcwlrLKzPX0611NIM7xzsd56i1CAvmt6edwLwNO/liqZZ3kF/6ZOFm5qzbwd1ndCMhKtTpOOKAzolRPHlpf5Zv2cPdHy7GWktNnYtb386lqKyKF67KICk6zOmYIgD0S42lS1IU909ZxvnPzOKBT5YyKaeA1dtKqfPBSV8mZudzy9s59ExuwXs3DiKphZ5r4jldWkXzj3F9fO6DvSCnA4h/mrN2Byu3lvKoD69XdfGAVN6Ys4G/T13BSd2SNAW7/KRkbw0Pfb6CvqmxXDpQY7eas9HdkvjdaSfwr69W0bNtC7aWVDJn3Q4ev6gvfVJinY4n8hNjDM9ekcH78zexuKCESTkFvD6nfoKi8OBAeiW3oHdyLH1SYuidEkPH+EgCArzz7/fLM9fzt8+WM7xLAs9fmUFEiN7OihwJPVPELV6ZtZ74yBDO6dfW6SjHLDCgfk22y178kZdnrufW0WlORxIv8fjXq9hZXsWrv6qf6Uuat1tGdWb5lj088uVKrIXrhnVkXHqK07FEfiEtKYp7z65fa9TlsqwrLmdJ4W4WF5SwpKCEd+Zt5JVZ9euVRYUG0Su5BX1SYumdHEOflBjaxUU4+mGutZYnvlnNk9/ncWav1vznkn6EBulDWZEjpcJPmtz64nK+W7mdX5/UxedbyYZ0TuD0nq14+oc8xmekqCuJsKSghDfnbuTKQe3pnRLjdBzxAsYY/nVhHzbv3ktcRAh/PFNjPsX7BQQY0pKiSEuK4vz+9R9U1Na5WFtUzuKC3SwpLGFxQQmvzd7w0+LVLcKC6gvBlBj6JNe3DCbHhnukGHS5LH/9dBmvz9nIRZkp/P383gRpJmWRo2L8ZeKKzMxMm52d7XQMAR74ZCnvzstn5j2j/WJ8y8Yd5Zz6+AzO6deWx8b3dTqOOKjOZRn3zCwKd1fy3W9HEhPu/WtTiue4XBZj8Nnu7SIHUlPnYvW2UpYUlLC4sITFBbtZtbWUmrr6949xkSE/tQjW/xtLqxahTfo8qKlz8fuJi/h44WZuGN6RP53VXc8zkYMwxuRYazMPdJ9a/KRJleytYWJOAWP7tvWLog+gfXwk1wzrwPPT13H14A5q5WnG3p23iUUFJfz3kn4q+uQXvHU8lMjxCA4MoGfbGHq2jeGShm2VNXWs2lrK4sISlhTUdxV9ZlrxT5PEJEaH/tQiWF8QxpIYfWyTYFXW1HHr27l8t3I7vz/9BG4Z1VlFn8gxUuEnTer9+ZuoqK7j2mEdnI7SpG4bncaHOQX89dNlTJwwWH90mqHisioe/XIlgzvFc05f3x27KiJyvMKCA+mbGkvf1FigPQB7q+tYvmVPfSFYWD9m8PtV29nXsaxNTNj/tww2jBuMiww55HVKK2u4/vVs5m3Yyd/O68WVg9q79wcT8XMq/KTJ1Na5eH32RgZ1iqNnW/9qFYsOC+Z3p53APZOX8NniLYzVG/9m5+9TV7C3po6/nddLhb+IyH7CQwLJaN+SjPYtf9pWXlXLss17fhozuKSghK+Xb/vp/pSW4T+1CPZJiaFXcsxPvSl2lFVx9avzWLmllP9c3I9z+yV7/GcS8Tcq/KTJfLVsG4W79/KXc3o6HcUtxmem8vqcjfzzi5Wc2qOVz09cI0du7rodTM4t5NbRnUlLinI6joiIT4gMDWJgxzgGdoz7adueyhqWNhSB+1oGpy75//VyO8RH0DsllmWbSyjctZcXrsrgpG6tnIgv4ndU+EmTeWXWetrHR3BStySno7hFYIDhgbE9uOSFubw4Yx2/PrmL05HEA2rqXNz38VKSY8O5bbQecxGR49EiLJghnRMY0jnhp227K6p/mkV0SUEJuRt3UVXr4s3rTvxZ0Sgix0eFnzSJhfm7ydm4iwfG9vDrdc0GdYrnzF6teWbaWi4akEorLe/g916ZuZ4128t46apMwkPUyisi0tRiI0IY3iWR4V0Sf9pmrVW3epEmpgVQpEm8MnM90aFBjM9MdTqK2/3xzO7UuSyPfLnS6SjiZoW79/Kfb9dwSvdWnNJDXY1ERDxFRZ9I01PhJ8dtS8lepi7ZwsUDUokK9f9G5HbxEVw3vCOTcwtZmL/b6TjiRg9+ugyL5S/n9HA6ioiIiMhxUeEnx+2NORtxWcvVQzo4HcVjbh2dRkJUaH1hsG+uavEr36/cxlfLtnH7yV1IaRnhdBwRERGR46LCT45LRXUt7/y4idN7tiY1rvm8OY4KDeLu008gd9Nupiza7HQcaWJ7q+t4YMoyOidGcv2wTk7HERERETluKvzkuEzOLaRkbw3XDevodBSPuzAjhZ5tW/DIFyvZW13ndBxpQs9MyyN/517+dl4vQoL0MikiIiK+T+9o5Ji5XJZXZ62nT0rMzxZsbS4CAgwPjO3J5pJKXpixzuk40kTWFpXx/PR1nN8/+WfTjYuIiIj4MhV+csymrylibVE51w3r2Gxn3xrYMY6ze7fhuelr2VKy1+k4cpystTzwyTJCgwP441ndnI4jIiIi0mRU+Mkxe2Xmelq1COXMXm2cjuKoe87sRp21PPrlKqejyHH6dPEWZuYV8/vTTyApWms0ioiIiP9Q4SfHZPW2UrLWFHPV4A7NfgxUalwENwzvyEcLCsndtMvpOHKM9lTW8LfPltM7OYbLT2zvdBwRERGRJtW837HLMXt11npCgwK4bGA7p6N4hVtGpZEYHcqDny7X8g4+6olvVlNcVsVD5/UiMKB5dl0WERER/6XCT47azvJqJucWMi49hZaRIU7H8QqRDcs7LMzfzScLfWt5BxWqsGxzCa/P3sDlJ7ajb2qs03FEREREmlyQ0wHE97zz40aqal1cO7SD01G8ygXpKbwxZyP//GIlp/VsRUSIM08vay1lVbXsKKumuKyK4oZ/d5RVs6O86qdtOxr+3VtTx29P7cqNIzo1y0l6XC7Lnz9eSlxkCL8/TRO6iIiIiH9y6ztTY8wZwH+BQOAla+0/97v/N8D1QC1QBFxrrd3YcF874CUgFbDAWdbaDe7MK4dXXevijTkbGdE1kS6top2O41UCAgz3j+3B+Ofm8Pz0ddx1atcmO3dtnYudFdX1xdtPBd3/F3A7yv+/uCsuq6Kq1nXA88SEB5MQFUJ8VCgntI5maFQom3ZW8I8vVrJ1TyX3nd2DgGbWzfH97HwWbNrN4xf1JSYi2Ok4IiIiIm7htsLPGBMIPA2cChQA840xU6y1yxvttgDItNZWGGNuBh4FLm647w3gYWvtN8aYKODA72TFoz5fspntpVX8a3zzW7D9SAzoEMeYPm14fsZaLh6QStvY8IPuW1Fd3ypXtK81rlEx9/MWump2VVRzoB6ZwYGG+MhQ4qNCSIgKJS0pisSo+u/jI0NJiA4lPjKExOhQWkaEHHAiHpfL8tDnK3hl1nq2l1bx+EV9CQ0KbMr/Fq+1o6yKf36xkoEd4zi/f7LTcURERETcxp0tfgOBPGvtOgBjzHvAucBPhZ+19odG+88FrmjYtwcQZK39pmG/MjfmlCNkreXlmetJS4piRBctbH0w95zZjW+Wb+OPk5cw6oTEX3a5LK8v6iqq6w54fHRo0E8FW6eEKAZ0qG+hS2xoqYuPDCEhOpSEyFBahAcdd/fMgADDfWO60zomlL9PXcmOsipeuCqTFmH+3/r1zy9WUl5Vy0Pn9WqW3VxFRESk+XBn4ZcM5Df6vgA48RD7Xwd80fB1V2C3MWYy0BH4FrjHWvuzd8rGmBuBGwHatdPsku42f8Mulhbu4e/n99ab5ENIaRnBTSM78+R3a5i+uojAAENcZH2LXEJUCB3iI0iICiW+4fuEqP9vsYuLDCEs2POtbcYYbhzRmaToMH43cREXPTeH164ZSOsY/13Lbv6GnUzMKeCmkZ3oqm7LIiIi4ue8YnIXY8wVQCYwsmFTEDAc6A9sAt4HfgW83Pg4a+0LwAsAmZmZmprQzV6ZuZ7YiGB1iTsCd5zchfP7JxMTHkxseLDPjJs7r38y8VEhTHgzhwuenc3r1w4gLcn/iqKaOhd//mgpbWPCuOPkLk7HEREREXE7dy7nUEj9xCz7pDRs+xljzCnAvcA51tqqhs0FwEJr7TprbS3wMZDuxqxyGPk7K/h6+VYuG9iO8JDmMf7reAQGGDomRBIXGeIzRd8+w7sk8v5Ng6mqdXHBs3PI2bjT6UhN7vXZG1i1rZQHzunp2OyrIiIiIp7kzsJvPtDFGNPRGBMCXAJMabyDMaY/8Dz1Rd/2/Y6NNcYkNnx/Eo3GBornvTZ7AwHGcNXgDk5HEQ/olRzD5JuHEBcZwmUv/sjXy7Y6HanJbCnZyxPfrOakbkmc1qOV03FEREREPMJthV9DS91twFfACuADa+0yY8yDxphzGnb7FxAFTDTGLDTGTGk4tg74HfCdMWYJYIAX3ZVVDq20sob35+dzdp82fj3mS36uXXwEkyYMplubFkx4K4e3f9zodKQm8bfPllPrsvxlbE+NVRUREZFmw619nKy1U4Gp+227v9HXpxzi2G+APu5L5341dS6y1hQxoksiQYHubFx1r4nZBZRV1XLdMC3h0NzER4Xy7g0ncuvbudz70VK2lVRy16ldfbZgmrZqO1OXbOV3p3WlXXyE03FEREREPMZ3qxEfMH1VEde+ls2Qf37PI1+uZH1xudORjlqdy/Lq7PVktm9Jn5RYp+OIAyJCgnjxqkwuykzhye/zuOfDJdTW+d6ympU1dTwwZRmdEiK5YUQnp+OIiIiIeJQKPzcaeUIiL1yZQZ+UGF6YsY7Rj03joufn8GFOAXsPsoabt/l2xTbyd+5Va18zFxQYwCMX9OHXJ6XxfnY+N76ZQ0V1rdOxjsqz09aycUcFD57bq9ksUC8iIiKyj7HWP1ZByMzMtNnZ2U7HOKhteyr5MLeAD+bns2FHBdGhQYzt15aLM1PpkxLjtV3nLn5+DgW79jL996N8uruqNJ235m7k/k+W0icllld+NYC4yBCnIx3W+uJyTv/PDE7v2Zr/Xdrf6TgiIiIibmGMybHWZh7wPhV+nmWtZd76nbyfnc/UJVuorHHRrXU0F2Wmcl7/ZK96E720sIQx/5vJn8/uzvXD1TVO/t9Xy7Zy+7sLSI4N5/VrB5Ia573j5ay1XP3qfHI37uL7344kqYUmKBIRERH/dKjCT004HmaM4cRO8Tx+UT/m3XsKD5/fi9CgAB78bDmD/v4dt76dy/TVRdS5nC/IX5m1nsiQQC4akHr4naVZOb1na96+/kR2lFcz7tnZLC0scTrSQX2xdCszVhfx29O6qugTERGRZkstfl5i5dY9vD8/n48WFLK7oobk2HAuyEhhfEaKI60p2/dUMvSR77n8xPb85ZyeHr+++IY120q5+pV57Kms5bkrMhjWJcHpSD9TVlXLKf+eTlxkCFNuG6ruyiIiIuLX1OLnA7q1bsEDY3vy459O5qnL+tMpMZL/fb+GEf/6gSte+pEpizZTWeO5CWHemruRWpflV0M6eOya4nu6tIpm8i1DSWkZzjWvzeOThYVOR/qZ/3yzmm2llTx0fi8VfSIiItKsuXUdPzl6oUGBjOnTljF92lKwq4JJOQVMzC7g9ncXEBMezPn9k7koM5UebVu4LUNlTR1v/biJk7u1okNCpNuuI/6hdUwY7980mBvfyOaO9xayfU+VVyyXsGLLHl6dvYFLBrQjvV1Lp+OIiIiIOEpdPX2Ay2WZtbaY9+fn8/WybVTXueidHMNFA1I5p29bYsKDm/R678/fxB8+XMK7NwxicOf4Jj23+K/Kmjp++8EiPl+yheuGdeTes7oTEODMbLUul2X883NYX1zOd78ZSUsvmjRJRERExF0O1dVTLX4+ICDAMLxLIsO7JLKrvJqPFxby/vx87vt4KQ99tpyzerfhosxUBnWKO+5lIay1vDxzPd3btGBQp7gm+gmkOQgLDuR/l/YnMTqUl2euZ3tpFY+N7+PImnmTcgrI2biLRy/so6JPREREBBV+PqdlZAjXDO3Ir4Z0YElhCe/Pz2fKws18tKCQ9vERXJSZygXpKbSOObbZC2fl7WD1tjIeG9/Xa9cWFO8VEGB4YGwPWseE8c8vVlJcWsXzV2XQIqxpW6UPZVd5Nf/4YgUDOrTkwvQUj11XRERExJupq6cf2FtdxxdLt/D+/Hx+XL+TAAOjTkjiosxUTu6eRPBRTGpx7WvzWVxQwqx7RjvSUiP+Y3JuAXdPWkxaUhSvXzuQVh5aSuGPkxfzQXYBn98+jG6t3TcWVkRERMTbqKunnwsPCWRcegrj0lPYUFzOB9n5TMop4PuV20mICuH8/slcPCCVtKToQ55nbVEZ36/czl2ndFXRJ8dtXHoKCVGhTHgrh3HPzOb1aweSlhTl1mvmbNzFu/PyuWF4RxV9IiIiIo2oxc9P1da5mL66iPfn5/P9yu3Uuizp7WK5eEAqY/q0JTL0lzX/fR8v5f35+cz+40kkRIU6kFr80ZKCEq55bR61LsvLVw8go717ZtisrXMx9qlZ7Cqv5tvfjiTqAL/jIiIiIv5M6/g1Q0GBAZzcvRUvXJXJnD+ezB/P7MbuvTX84cMlDHj4W+6etIicjTvZV/jvrqhmUk4B5/Zrq6JPmlTvlBg+vHkIseHBXP7SXL5Zvs0t13ljzkZWbNnDA2N7qOgTERER2Y9a/JoRay05G3fx/vx8Pl+yhYrqOtKSorgoM4VdFTU8O20tX9wxnO5t1EVOml5xWRXXvTafJYUlPHx+by4d2K7Jzr1tTyUn/3s6Ge1b8to1AzQxkYiIiDRLh2rxU+HXTJVV1fLZos28n53Pgk27ARjSOZ53bhjkbDDxaxXVtdz6di4/rCrijpO7cOcpXZqkSLvtnVy+Xr6Nr+8cQYeEyCZIKiIiIuJ7NLmL/EJUaBCXDGzHJQPbsWZbKZ8t3sJZvds4HUv8XERIEC9clcmfJi/hv9+tYXtpJX87txdBRzHz7P6y1hTx2eIt3HlKFxV9IiIiIgehwk/o0iqau0499IyfIk0lODCARy/sQ6sWYTz1Qx5FpVX879J0wkOOfibZqto67v9kGR3iI5gwsrMb0oqIiIj4B03uIiIeZ4zhd6efwN/O68V3K7dz2Utz2VlefdTneX76OtYXl/Pgub0IC9YSJCIiIiIHo8JPRBxz5aD2PHt5Bss27+HC52aTv7PiiI/duKOcp37I4+zebRjRNdGNKUVERER8nwo/EXHUGb1a8/b1J1JcWsW4Z2ezbHPJYY+x1vKXKcsIDjDcN6aHB1KKiIiI+DYVfiLiuAEd4vjw5iEEBxgufn4us/OKD7n/V8u28cOqIu46tSutY8I8lFJERETEd6nwExGv0KVVNB/eMoTk2HCufnUeUxZtPuB+5VW1/PXTZXRrHc2vhnTwbEgRERERH6XCT0S8RpuYcD6YMJj+7Vpy+7sLeClr3S/2efK7NWwpqeSh845vGQgRERGR5kTvmkTEq8SEB/PGtQM5q3drHvp8BQ99thyXywKwamspL89cz8WZqWR2iHM4qYiIiIjv0Dp+IuJ1woID+d+l6SRFL+elmevZXlrFv8b34b6PlxIVFsQfzuzmdEQRERERn6LCT0S8UmCA4YGxPWjVIoxHvlzJ4oLdbNhRwT/H9SYuMsTpeCIiIiI+RYWfiHgtYww3j+pMqxah3D1pMf3bxXJRZqrTsURERER8jgo/EfF649JT6JMSS2JUKAEBxuk4IiIiIj5HhZ+I+IS0pCinI4iIiIj4LM3qKSIiIiIi4udU+ImIiIiIiPg5FX4iIiIiIiJ+ToWfiIiIiIiIn1PhJyIiIiIi4ufcWvgZY84wxqwyxuQZY+45wP2/McYsN8YsNsZ8Z4xpv9/9LYwxBcaYp9yZU0RERERExJ+5rfAzxgQCTwNnAj2AS40xPfbbbQGQaa3tA0wCHt3v/r8BM9yVUUREREREpDlwZ4vfQCDPWrvOWlsNvAec23gHa+0P1tqKhm/nAin77jPGZACtgK/dmFFERERERMTvubPwSwbyG31f0LDtYK4DvgAwxgQA/wZ+d6gLGGNuNMZkG2Oyi4qKjjOuiIiIiIiIfwpyOgCAMeYKIBMY2bDpFmCqtbbAGHPQ46y1LwAvNJyjyBiz0d1Zj0ECUOx0CDksPU6+QY+Tb9Dj5P30GPkGPU6+QY+T92tOj1H7g93hzsKvEEht9H1Kw7afMcacAtwLjLTWVjVsHgwMN8bcAkQBIcaYMmvtLyaI2cdam9hkyZuQMSbbWpvpdA45ND1OvkGPk2/Q4+T99Bj5Bj1OvkGPk/fTY1TPnYXffKCLMaYj9QXfJcBljXcwxvQHngfOsNZu37fdWnt5o31+Rf0EMAct+kREREREROTg3DbGz1pbC9wGfAWsAD6w1i4zxjxojDmnYbd/Ud+iN9EYs9AYM8VdeURERERERJort47xs9ZOBabut+3+Rl+fcgTneA14ramzedALTgeQI6LHyTfocfINepy8nx4j36DHyTfocfJ+eowAY611OoOIiIiIiIi4kTuXcxAREREREREvoMJPRERERETEz6nwayLGmDOMMauMMXnGmF/MQGqMCTXGvN9w/4/GmA4OxGzWjDGpxpgfjDHLjTHLjDF3HGCfUcaYkobJhhYaY+4/0LnEvYwxG4wxSxoeg+wD3G+MMU82PJ8WG2PSncjZXBljTmj0HFlojNljjLlzv330XHKAMeYVY8x2Y8zSRtvijDHfGGPWNPzb8iDHXt2wzxpjzNWeS938HORx+pcxZmXDa9pHxpjYgxx7yNdHaToHeZz+YowpbPTadtZBjj3k+0JpGgd5jN5v9PhsMMYsPMixze65pDF+TcAYEwisBk4FCqhfyuJSa+3yRvvcAvSx1k4wxlwCnG+tvdiRwM2UMaYN0MZam2uMiQZygPP2e5xGAb+z1o5xJqVA/Ysx9cu4HHCx1YY/tL8GzgJOBP5rrT3Rcwlln4bXv0LgRGvtxkbbR6HnkscZY0YAZcAb1tpeDdseBXZaa//Z8Aa0pbX2D/sdFwdkA5mApf71McNau8ujP0AzcZDH6TTge2ttrTHmEYD9H6eG/TZwiNdHaToHeZz+ApRZax87xHGHfV8oTeNAj9F+9/8bKLHWPniA+zbQzJ5LavFrGgOBPGvtOmttNfAecO5++5wLvN7w9STgZGOM8WDGZs9au8Vam9vwdSn1y4wkO5tKjtG51L/IW2vtXCC2obAXzzsZWNu46BPnWGtnADv329z478/rwHkHOPR04Btr7c6GYu8b4Ax35WzuDvQ4WWu/blgKC2AukOLxYPIzB3k+HYkjeV8oTeBQj1HD++yLgHc9GsqLqfBrGslAfqPvC/hlQfHTPg0v7CVAvEfSyS80dLXtD/x4gLsHG2MWGWO+MMb09GwyaWCBr40xOcaYGw9w/5E858QzLuHgf1T1XPIOray1Wxq+3gq0OsA+ek55l2uBLw5y3+FeH8X9bmvokvvKQbpO6/nkHYYD26y1aw5yf7N7Lqnwk2bHGBMFfAjcaa3ds9/duUB7a21f4H/Axx6OJ/WGWWvTgTOBWxu6coiXMcaEAOcAEw9wt55LXsjWj+/QGA8vZoy5F6gF3j7ILnp9dNazQGegH7AF+LejaeRQLuXQrX3N7rmkwq9pFAKpjb5Padh2wH2MMUFADLDDI+nkJ8aYYOqLvrettZP3v99au8daW9bw9VQg2BiT4OGYzZ61trDh3+3AR9R3m2nsSJ5z4n5nArnW2m3736HnklfZtq8rdMO/2w+wj55TXsAY8ytgDHC5PcgkDEfw+ihuZK3dZq2ts9a6gBc58P+/nk8Oa3ivPQ54/2D7NMfnkgq/pjEf6GKM6djwCfglwJT99pkC7Jsl7ULqB3DrU1cPaujr/TKwwlr7+EH2ab1v7KUxZiD1zxEV6B5kjIlsmHwHY0wkcBqwdL/dpgBXmXqDqB+4vQXxtIN+mqrnkldp/PfnauCTA+zzFXCaMaZlQ9e10xq2iYcYY84A7gbOsdZWHGSfI3l9FDfabzz5+Rz4//9I3heKe50CrLTWFhzozub6XApyOoA/aJiB6zbq/0gGAq9Ya5cZYx4Esq21U6gvON40xuRRPwj1EucSN1tDgSuBJY2m9v0T0A7AWvsc9UX5zcaYWmAvcIkKdI9rBXzUUDMEAe9Ya780xkyAnx6nqdTP6JkHVADXOJS12Wr4Q3kqcFOjbY0fIz2XHGCMeRcYBSQYYwqAB4B/Ah8YY64DNlI/2QHGmExggrX2emvtTmPM36h/wwrwoLX2WCa1kCNwkMfpj0Ao8E3D69/chpnA2wIvWWvP4iCvjw78CM3CQR6nUcaYftR3md5Aw2tg48fpYO8LPf8T+L8DPUbW2pc5wPhzPZe0nIOIiIiIiIjfU1dPERERERERP6fCT0RERERExM+p8BMREREREfFzKvxERERERET8nAo/ERERERERP6fCT0REBDDG1BljFja63dOE5+5gjPH7NaJERMR7aR0/ERGRenuttf2cDiEiIuIOavETERE5BGPMBmPMo8aYJcaYecaYtIbtHYwx3xtjFhtjvjPGtGvY3soY85ExZlHDbUjDqQKNMS8aY5YZY742xoQ37H+7MWZ5w3nec+jHFBERP6fCT0REpF74fl09L250X4m1tjfwFPCfhm3/A1631vYB3gaebNj+JDDdWtsXSAeWNWzvAjxtre0J7AYuaNh+D9C/4TwT3POjiYhIc2estU5nEBERcZwxpsxaG3WA7RuAk6y164wxwcBWa228MaYYaGOtrWnYvsVam2CMKQJSrLVVjc7RAfjGWtul4fs/AMHW2oeMMV8CZcDHwMfW2jI3/6giItIMqcVPRETk8OxBvj4aVY2+ruP/x9mfDTxNfevgfGOMxt+LiEiTU+EnIiJyeBc3+ndOw9ezgUsavr4cyGr4+jvgZgBjTKAxJuZgJzXGBACp1tofgD8AMcAvWh1FRESOlz5VFBERqRdujFnY6PsvrbX7lnRoaYxZTH2r3aUN234NvGqM+T1QBFzTsP0O4AVjzHXUt+zdDGw5yDUDgbcaikMDPGmt3d1EP4+IiMhPNMZPRETkEBrG+GVaa4udziIiInKs1NVTRERERETEz6nFT0RERERExM+pxU9ERERERMTPqfATERERERHxcyr8RERERERE/JwKPxERERERET+nwk9ERERERMTP/R+OMySiwBlKfwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# .99 active learning .01 replay buffer\n",
    "# w/o  wasserstein loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "# load model weights\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "print(\"Loaded model weights\")\n",
    "\n",
    "# load replay buffer params\n",
    "means = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "means = np.nan_to_num(means)\n",
    "covs = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "print(\"Loaded means and covariances for replay buffer\")\n",
    "\n",
    "batch_size = 10\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "loss = []\n",
    "target_miou = []\n",
    "\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda2, label='active learning ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31m_FallbackException\u001b[0m                        Traceback (most recent call last)",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name)\u001b[0m\n\u001b[1;32m    925\u001b[0m         \u001b[0;34m\"explicit_paddings\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"data_format\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 926\u001b[0;31m         \"dilations\", dilations)\n\u001b[0m\u001b[1;32m    927\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0m_result\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31m_FallbackException\u001b[0m: Expecting int64_t value for attr strides, got numpy.int32",
      "\nDuring handling of the above exception, another exception occurred:\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-17-610970e6c53e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     45\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     46\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mactive_learning\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 47\u001b[0;31m         \u001b[0mpred_target_labels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdeeplabv3\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     48\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m         \u001b[0mX_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_labeled\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mX_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mY_eval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mselected_idx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_active_XY\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget_train_data\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_target_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_train_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m256\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_classes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrategy\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mstrategy\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    966\u001b[0m           with base_layer_utils.autocast_context_manager(\n\u001b[1;32m    967\u001b[0m               self._compute_dtype):\n\u001b[0;32m--> 968\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcast_inputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    969\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle_activity_regularization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    970\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_set_mask_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_masks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/network.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs, training, mask)\u001b[0m\n\u001b[1;32m    717\u001b[0m     return self._run_internal_graph(\n\u001b[1;32m    718\u001b[0m         \u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtraining\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtraining\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmask\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmask\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 719\u001b[0;31m         convert_kwargs_to_constants=base_layer_utils.call_context().saving)\n\u001b[0m\u001b[1;32m    720\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    721\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0mcompute_output_shape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/network.py\u001b[0m in \u001b[0;36m_run_internal_graph\u001b[0;34m(self, inputs, training, mask, convert_kwargs_to_constants)\u001b[0m\n\u001b[1;32m    886\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    887\u001b[0m           \u001b[0;31m# Compute outputs.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 888\u001b[0;31m           \u001b[0moutput_tensors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlayer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcomputed_tensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    889\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    890\u001b[0m           \u001b[0;31m# Update tensor_dict.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    966\u001b[0m           with base_layer_utils.autocast_context_manager(\n\u001b[1;32m    967\u001b[0m               self._compute_dtype):\n\u001b[0;32m--> 968\u001b[0;31m             \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcast_inputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    969\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle_activity_regularization\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    970\u001b[0m           \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_set_mask_metadata\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_masks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/keras/layers/convolutional.py\u001b[0m in \u001b[0;36mcall\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m    205\u001b[0m       \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marray_ops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpad\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_compute_causal_padding\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    206\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 207\u001b[0;31m     \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_convolution_op\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkernel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    208\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    209\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_bias\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m   1104\u001b[0m           call_from_convolution=False)\n\u001b[1;32m   1105\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1106\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconv_op\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1107\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1108\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m    636\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    637\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m__call__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m  \u001b[0;31m# pylint: disable=redefined-builtin\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 638\u001b[0;31m     \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    639\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    640\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inp, filter)\u001b[0m\n\u001b[1;32m    235\u001b[0m         \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpadding\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    236\u001b[0m         \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 237\u001b[0;31m         name=self.name)\n\u001b[0m\u001b[1;32m    238\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    239\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, data_format, dilations, name, filters)\u001b[0m\n\u001b[1;32m   2012\u001b[0m                            \u001b[0mdata_format\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2013\u001b[0m                            \u001b[0mdilations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdilations\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2014\u001b[0;31m                            name=name)\n\u001b[0m\u001b[1;32m   2015\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2016\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name)\u001b[0m\n\u001b[1;32m    931\u001b[0m             \u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstrides\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0muse_cudnn_on_gpu\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0muse_cudnn_on_gpu\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    932\u001b[0m             \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mpadding\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mexplicit_paddings\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 933\u001b[0;31m             data_format=data_format, dilations=dilations, name=name, ctx=_ctx)\n\u001b[0m\u001b[1;32m    934\u001b[0m       \u001b[0;32mexcept\u001b[0m \u001b[0m_core\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_SymbolicException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    935\u001b[0m         \u001b[0;32mpass\u001b[0m  \u001b[0;31m# Add nodes to the TensorFlow graph.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py\u001b[0m in \u001b[0;36mconv2d_eager_fallback\u001b[0;34m(input, filter, strides, padding, use_cudnn_on_gpu, explicit_paddings, data_format, dilations, name, ctx)\u001b[0m\n\u001b[1;32m   1020\u001b[0m   explicit_paddings, \"data_format\", data_format, \"dilations\", dilations)\n\u001b[1;32m   1021\u001b[0m   _result = _execute.execute(b\"Conv2D\", 1, inputs=_inputs_flat, attrs=_attrs,\n\u001b[0;32m-> 1022\u001b[0;31m                              ctx=ctx, name=name)\n\u001b[0m\u001b[1;32m   1023\u001b[0m   \u001b[0;32mif\u001b[0m \u001b[0m_execute\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmust_record_gradient\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1024\u001b[0m     _execute.record_gradient(\n",
      "\u001b[0;32m/global/cscratch1/sd/saksman/accda_conda/lib/python3.6/site-packages/tensorflow/python/eager/execute.py\u001b[0m in \u001b[0;36mquick_execute\u001b[0;34m(op_name, num_outputs, inputs, attrs, ctx, name)\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0mctx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mensure_initialized\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m     tensors = pywrap_tfe.TFE_Py_Execute(ctx._handle, device_name, op_name,\n\u001b[0;32m---> 60\u001b[0;31m                                         inputs, attrs, num_outputs)\n\u001b[0m\u001b[1;32m     61\u001b[0m   \u001b[0;32mexcept\u001b[0m \u001b[0mcore\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_NotOkStatusException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA34AAAJcCAYAAACmOnadAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAACpuUlEQVR4nOzdd3hUZd7G8e+TSS+EVDppNOkllFAEbGvvitgrlrW/W9zu7uqu6zZ1LYCIYsfV1dUVewGB0KXXkAQINQVCQnryvH+cSQEDBEgyyeT+XNdcZM555szvDC13nmastYiIiIiIiIj38vF0ASIiIiIiItK0FPxERERERES8nIKfiIiIiIiIl1PwExERERER8XIKfiIiIiIiIl5OwU9ERERERMTLKfiJiIiIiIh4OQU/ERGRozDGZBpjzvJ0HSIiIqdKwU9ERERERMTLKfiJiIicAGNMgDHmKWPMLvfjKWNMgPtctDHmf8aYA8aYPGPMd8YYH/e5nxtjdhpjCowxm4wxZ3r2TkREpC3x9XQBIiIircyvgFHAYMAC/wV+DfwG+D8gC4hxtx0FWGNMb+BeYLi1dpcxJh5wNW/ZIiLSlqnHT0RE5MRcB/zBWrvPWpsN/B64wX2uHOgExFlry62131lrLVAJBAB9jTF+1tpMa+1Wj1QvIiJtkoKfiIjIiekMbKvzfJv7GMBfgTTgc2NMujHmEQBrbRrwIPAosM8Y87YxpjMiIiLNRMFPRETkxOwC4uo87+4+hrW2wFr7f9baROBi4OHquXzW2jettWPdr7XAX5q3bBERacsU/ERERI7NzxgTWP0A3gJ+bYyJMcZEA78FXgcwxlxojOlhjDFAPs4QzypjTG9jzBnuRWBKgGKgyjO3IyIibZGCn4iIyLHNwQlq1Y9AYBmwGlgDrAAec7ftCXwJFAKpwPPW2m9w5vc9AeQAe4BY4BfNdwsiItLWGWfOuYiIiIiIiHgr9fiJiIiIiIh4OQU/ERERERERL6fgJyIiIiIi4uUU/ERERERERLycr6cLaCzR0dE2Pj7e02WIiIiIiIh4xPLly3OstTH1nfOa4BcfH8+yZcs8XYaIiIiIiIhHGGO2He2chnqKiIiIiIh4OQU/ERERERERL6fgJyIiIiIi4uW8Zo6fiIiIiEhLV15eTlZWFiUlJZ4uRVqxwMBAunbtip+fX4Nfo+AnIiIiItJMsrKyCAsLIz4+HmOMp8uRVshaS25uLllZWSQkJDT4dRrqKSIiIiLSTEpKSoiKilLok5NmjCEqKuqEe40V/EREREREmpFCn5yqk/kzpOAnIiIiIiLi5RT8RKRRlVVUsSg9l7R9hVhrPV2OiIiI1HHgwAGef/7547bLzMzkzTffbFC7/v37N/i4eI4WdxGRU1ZWUcWCtBw+XrObz9ft4WBJBQBRIf4kx0cwIiGKEfGRnNYpDF+Xft4kIiLiKdXB75577jlmu+rgd+211zZTZdLUFPxE5KTUF/bCAnw5u18HftSvIweKylickcfSzDw+W7cXgBB/F0PjIhgRH8nwhEgGd2tPoJ/Lw3ciIiLSdjzyyCNs3bqVwYMHc/bZZ/Pkk0/ys5/9jE8++QRjDL/+9a+ZNGkSjzzyCBs2bGDw4MHcdNNNXHbZZdxwww0cOnQIgGeffZbRo0c36D1LSkq4++67WbZsGb6+vvzjH/9g4sSJrFu3jltuuYWysjKqqqp477336Ny5M1dffTVZWVlUVlbym9/8hkmTJjXlR9JmKPiJSIOVVVSxYGsOH6/+Ydi7YEAnxvaMJsDlA7lpEBzFpOHdAdiTX8KSzDyWZuSxJCOPv3+xGQB/lw8Du4YzPCGSEfGRDIuPoF1gw/ejERERac1+/9E61u862KjX7Nu5Hb+7qN9Rzz/xxBOsXbuWlStXAvDee++xcuVKVq1aRU5ODsOHD+f000/niSee4G9/+xv/+9//ACgqKuKLL74gMDCQLVu2MHnyZJYtW9agmp577jmMMaxZs4aNGzdyzjnnsHnzZqZOncoDDzzAddddR1lZGZWVlcyZM4fOnTvz8ccfA5Cfn39qH4jUUPATkWOqDntzVu/ms7phr28HLhhYJ+ztWQ3fToV178P+DOfFsX0hbjQdu6dwccJoLh7kjPU/UFTGssz9LM3MY0lmHi/OS+eFb7diDJzWsR0jEiIZHh/J8IQIYsMCPXj3IiIi3m3+/PlMnjwZl8tFhw4dGD9+PEuXLqVdu3aHtSsvL+fee+9l5cqVuFwuNm/efELvcd999wHQp08f4uLi2Lx5MykpKTz++ONkZWVx+eWX07NnTwYMGMD//d//8fOf/5wLL7yQcePGNer9tmUKfiLyA3XD3ufr95JfXF4T9s4f0IlxvY4Ie+s/gLx0MC5IOB1G3wtF+2HbAlj5Fiyd4Vy4fRzEjaF9XApndR/NWaf1AWMoKqtg5fYDTq9gZh6zl+7glYWZAMRHBdcEwREJkXSPDNYy2CIi4hWO1TPX0vzzn/+kQ4cOrFq1iqqqKgIDT/0Hs9deey0jR47k448/5vzzz2fatGmcccYZrFixgjlz5vDrX/+aM888k9/+9reNcAei4CciAJRXuufsHTfsrYG505yevZqwNw7GPAB9LoKQqDpX/SlUVjgBcdtC2J4KWz6DVe5VwkJiIS6F4O6jGR2Xwugz+oNPL8orq1i7M9/pEczYz+fr9/LOsiwAYsMCaoaGjkiIpHeHMHx8FARFREQaIiwsjIKCgprn48aNY9q0adx0003k5eUxb948/vrXv7Jz587D2uXn59O1a1d8fHyYNWsWlZWVDX7PcePG8cYbb3DGGWewefNmtm/fTu/evUlPTycxMZH777+f7du3s3r1avr06UNkZCTXX3897du3Z8aMGY16/22Zgp9IG1Yd9uas2c1n62rD3ll9nTl7Pwx7H0De1tqwN/p+OO0iCIk++pu4fKHLUOcx+l6wFnI2O72B21KdMLj+v07bgHbQbSR+cSkM6T6aIaOHMuX0JKqqLGnZhc5iMe4FYz5evRuAdoG+JMdX9whGMKBLe/x9tXKoiIhIfaKiohgzZgz9+/fnvPPO48knnyQ1NZVBgwZhjOHJJ5+kY8eOREVF4XK5GDRoEDfffDP33HMPV1xxBa+++irnnnsuISEhDX7Pe+65h7vvvpsBAwbg6+vLK6+8QkBAAO+88w6vvfYafn5+dOzYkV/+8pcsXbqUn/70p/j4+ODn58cLL7zQhJ9G22K8ZZ+t5ORk29AJpiJtWd2w9/n6vRwoKie0bs9ez2gCfX1g71qnV68m7Pk4wzj7Xnr8sHeiDmx3h8CFzq85m5zjrgDomgzdUyAuBbqNhIAwrLVk7S929wg68wTTs51VxgL9fBjcrX3NyqFDu0cQEqCfcYmISMuwYcMGTjvtNE+XIV6gvj9Lxpjl1trk+tor+Im0AeWVVSzcmsvHq3c1IOx94B7G6Q578eOg32WNH/aO5VCO0xNYHQZ3rwZb6dTTcSDEjXaHwdE1NeUUlrLMPTR0SWYu63cdpMqCy8fQv3M792IxTs9gZIh/89yHiIjIERT8pLEo+IkIUBv25qzezWfr99SEvbNOi+WCgZ3rhL11TtBb/4GzDUNN2LvUmbMXGuPpW4HSAsha6swT3JYKO5dBRYlzLrpXbQiMGw3tnS0kCkrKWbH9gLOFRGYeK3ccoKyiCoAesaEMj49kZIITBru0D/LUnYmISBuj4CeN5USDn8Y/iXiRY4W98wd04vReMbVh77sXncBXE/bGQsqPW07YqysgDJLOcB4AFaWwa6UzT3B7qtNLuWKWc65dV4hLIax7CuPjxjD+nN5gDKUVlazOymeJe47g/1bt4q0l2wHo0j6I4fERNYvG9IgN1cqhIiIi4lUU/ERaufLKKlK35vJxnbAX4u+qGcZ5eNib4YSk3C21YW/UPXDaxS0v7B2LbwB0H+k8AKoqYd96d4/gQsiYB2v+7ZwLioTuKQTEpTA8bjTDTx8EE3tQWWXZuOdgTRCcn5bLByt3ARAZ4k9yXETNNhL9OrfD16UFY0RERKT1UvATaYWqw96cNbv5dF1t2KtejbMm7O1bD9+95O7Zqxv27m59Ye9YfFzQcYDzGHmns3JoXrp7nqA7DG762GnrFwLdhuPqPpp+caPpNyKZW8YkYK0lM7eIpRl5zuqhmXl8vn4vAMH+LobFRTjzBOMjGdK9PYF+Lg/esIiIiMiJ8UjwM8acCzwNuIAZ1ton6mlzNfAoYIFV1tprm7VIkRambtj7bN0e9tcJe+cP6MT4I8Pe+g+cbROMD8SNcYe9iyA01tO30vSMgagk5zHkeufYwd21QXB7Knz7Z8CCjx90HoKJSyEhbgwJ/Udy9fBuAOw9WFLTI7gkI49/frkZa8HPZRjQJZwRCVGMSIhgWFwk4UF+nrtfERERkeNo9sVdjDEuYDNwNpAFLAUmW2vX12nTE3gHOMNau98YE2ut3Xes62pxF/FGFZVVpKa7h3EeM+xtqF2gpW7Y63epu2evDYS9E1W8H3Ysqd1PcNf3UFUOGOjQ7/AFY8I6ApBfVM6ybc5iMUsz8lizM5/ySosx0LtDGCMSnE3lR8RHEtsu0LP3JyIiLZI3LO4yYcIE/va3v5GcXO8aIg0WHx/PsmXLiI5u+KrhP/3pT5kzZw7nn38+P/vZz7jwwgspKyvjmWeeYdy4cadUz4lqrM/hZLWGxV1GAGnW2nQAY8zbwCXA+jpt7gCes9buBzhe6BPxJtVhb86a3Xy6tjbsnXlaBy4Y6A57fi4n7M2f6QS+umFv5J0Kew0RFAG9fuQ8AMqKYOdyd4/gQlj5Jix90TkXkQBxowmPG82Z3VM4s08fMIbiskq+37GfpRn7WZqZx7vLs3g1dRsAcVHBzqby7m0k4qOCtWCMiIi0KNZarLX4+LSeeezTp08nLy8Pl8vF22+/zYABA5gxY0aDX19ZWYnL1Tana3gi+HUBdtR5ngWMPKJNLwBjzAKc4aCPWms/PfJCxpgpwBSA7t27N0mxIs2hbtj7bN1e8g6V1YS98wd0YkLvI8PeB+5Nzo0zZ2/EFCfshXXw9K20Xv7BkDDOeQBUlsOe1bVbSGz6BFa+4ZwL7QBxownqPprRcSmMPqMf+PSkvLKK9bsO1mwq/9WGvby7PAuAmLAAJwS6Vw/t07EdLh8FQRERaV6ZmZn86Ec/YuTIkSxfvpw5c+bwzjvv8M4771BaWspll13G73//ezIzMzn33HMZNmwYK1asoF+/frz66qsEBwcfdr27776bpUuXUlxczJVXXsnvf/97vv76a5555hk++OADAL744guef/553n///R/U8+STT/LJJ58QFBTEm2++SY8ePbj55pu58MILufLKKwEIDQ2lsLCQiy++mMLCQoYNG8bkyZN57rnnKC4uZtmyZaSmpvLdd9/xu9/9jtLSUpKSknj55ZcJDQ0lPj6eSZMm8cUXX/Czn/2Ma665pub99+7dy1133UV6ejoAL7zwAqNHj+b111/nmWeeoaysjJEjR/L8888fMzC+9dZb/OlPf8JaywUXXMBf/vIXKisrue2221i2bBnGGG699VYeeughnnnmGaZOnYqvry99+/bl7bffPtXf1gZpqYu7+AI9gQlAV2CeMWaAtfZA3UbW2unAdHCGejZzjSKnpKKyikXpeXy8ZlcDwt7L7p69umHvDoW9puTygy7DnMfo+6CqyulZrd5CYluq83sCEBAO3UfhF5fCoO6jGTR6CHecnkhVlWVrdiFL3HMEl2bk8fGa3QCEBfqSHFe7hcSAruEE+LbNn0CKiLRZnzwCe9Y07jU7DoDzfrB8xmG2bNnCrFmzGDVqFJ9//jlbtmxhyZIlWGu5+OKLmTdvHt27d2fTpk289NJLjBkzhltvvZXnn3+en/zkJ4dd6/HHHycyMpLKykrOPPNMVq9ezcSJE7nnnnvIzs4mJiaGl19+mVtvvbXeWsLDw1mzZg2vvvoqDz74IP/73/+OWveHH35IaGgoK1euBKBDhw4sW7aMZ599lpycHB577DG+/PJLQkJC+Mtf/sI//vEPfvvb3wIQFRXFihUrfnDN+++/n/Hjx/P+++9TWVlJYWEhGzZsYPbs2SxYsAA/Pz/uuece3njjDW688cZ669q1axc///nPWb58OREREZxzzjl88MEHdOvWjZ07d7J27VoADhw4AMATTzxBRkYGAQEBNceagyeC306gW53nXd3H6soCFltry4EMY8xmnCC4tHlKFGkatWHPmbOXd6iMYH8XZx0t7K3/ALI3AsYZxqmw5zk+PhDbx3kMv805dmB77aqh21Nhy2fOcd9A6Docn+4p9IxLoefgEVw3Mg6ArP1F7sVi9rMkI5dvNmUDEODrw6Bu7RkR78wTHBoXQWhAS/3ZnIiItGZxcXGMGjUKgM8//5zPP/+cIUOGAFBYWMiWLVvo3r073bp1Y8yYMQBcf/31PPPMMz8Ifu+88w7Tp0+noqKC3bt3s379egYOHMgNN9zA66+/zi233EJqaiqvvvpqvbVMnjy55teHHnropO9p0aJFrF+/vqbesrIyUlJSas5PmjSp3td9/fXXNbW5XC7Cw8N57bXXWL58OcOHDweguLiY2NijT6FZunQpEyZMICbGWS39uuuuY968efzmN78hPT2d++67jwsuuIBzzjkHgIEDB3Lddddx6aWXcumll570PZ8oT3xXsRToaYxJwAl81wBHrtj5ATAZeNkYE40z9DO9OYsUaSxHC3tnnuZsvVAb9jbCglecXqS6Ye/8vzmrcboXGJEWpH135zHIPWSkMNsJgNWrh373N5hXBcYFnQZB3Gi6dk+ha68ULhvSFYDcwlKWZjpzBJdm5vH8t2k8+w34GOjXOdyZJ5jgbCURFRrgwZsVEZFGd5yeuaYSEhJS87W1ll/84hfceeedh7XJzMz8wdz0I59nZGTwt7/9jaVLlxIREcHNN99MSUkJALfccgsXXXQRgYGBXHXVVfj61h876l6z+mtfX1+qqqoAqKqqoqys7Lj3ZK3l7LPP5q233jruPTfkWjfddBN//vOfG/ya+kRERLBq1So+++wzpk6dyjvvvMPMmTP5+OOPmTdvHh999BGPP/44a9asOern05iaPfhZayuMMfcCn+HM35tprV1njPkDsMxa+6H73DnGmPVAJfBTa21uc9cqcrIqKqtYnJHH/1bXF/Y6MqF37BFh7wPI3oAT9kYr7LVWoTHQ92LnAVByELKWOMNCt6fCkhch9VnnXEwf6J5CVNxozu2ewrn9+wJQWFrBim37a7aQeGPxNmYuyAAgKSakZlP54fGRdI0I0oIxIiJySn70ox/xm9/8huuuu47Q0FB27tyJn5+zRdH27dtJTU0lJSWFN998k7Fjxx722oMHDxISEkJ4eDh79+7lk08+YcKECQB07tyZzp071wy/PJrZs2fzyCOPMHv27Joeuvj4eJYvX87VV1/Nhx9+SHl5+XHvY9SoUfz4xz8mLS2NHj16cOjQIXbu3EmvXr2O+bozzzyTF154gQcffLBmqOeZZ57JJZdcwkMPPURsbCx5eXkUFBQQFxdX7zVGjBjB/fffT05ODhEREbz11lvcd9995OTk4O/vzxVXXEHv3r25/vrrqaqqYseOHUycOJGxY8fy9ttvU1hYSPv27Y97j6fKI+OIrLVzgDlHHPttna8t8LD7IdIqVIe9j9fs5rO1e8g9WtjL3gQLZv0w7J33VycwKOx5j8B20OMs5wFQUQo7Vzirhm5LhbXvwfKXnXPh3SEuhdDuKZweN4bTe/YCYyitqGTtznxnU3n3DxPeWuKsj9UpPLAmCI5IiKRHTCg+WjBGREROwDnnnMOGDRtqQldoaCivv/46LpeL3r1789xzz3HrrbfSt29f7r777sNeO2jQIIYMGUKfPn0OGxZa7brrriM7O/uY21fs37+fgQMHEhAQUNNbd8cdd3DJJZcwaNAgzj333Ab11sXExPDKK68wefJkSktLAXjssceOG/yefvpppkyZwksvvYTL5eKFF14gJSWFxx57jHPOOYeqqir8/Px47rnnjhr8OnXqxBNPPMHEiRNrFne55JJLWLVqFbfccktN7+Wf//xnKisruf7668nPz8day/33398soQ88sI9fU9E+fuIJRwt7Z/SJ5cKBnRjfK5Ygf3fYW/eBexinO+x1T4F+lynstWVVlbB3rbtH0B0GD7l3rwmOhu6javcS7DAAXL5UVlk27Smo6RFckplHdoHzH1z7YD+S4yJJSYpibI9oenUIVY+giEgL01r28cvMzOTCCy+sWZjkZNx7770MGTKE2267rRErk2qtYR8/kVatorKKJRl5/K+esOfM2asOe5th4avOAi371lMT9s570lmgpV0nT9+KeJqPe+5fp0Ew6i6wFnK31obA7Qtho3t1M/9Q6DYCV/fR9I0bTd/hw7hpdDzWWrblFtVsKr8kM48vN+wFIDo0gDE9ohjTI5oxPaLp0j7IgzcrIiJtybBhwwgJCeHvf/+7p0sRN/X4iTRAddj72L2peu6hMoL8XJx5Wj1hb/0HTs9e3bDX71KFPTk5B3fVrhq6LRX2rXOOu/yh81CIS3EWAeo2AgLDAWfl0IVpucxPy2Hh1hxyCp1J8QnRIYzp4fQGpiRGEx7s56m7EhFps1pLj5+0fCfa46fgJ3IUlVWWxem5Natx5hQ6Ye+M02K5sN6w94H7m3LjDNHrd5nCnjS+ojzYsbg2DO76HqoqwPhAh37QfbSzCX38WAiKwFrLpr0FzN+Sw4K0HBZn5FFUVokxMKBLOGN6RDO2RzTD4iKcOagiItKkFPyksSj4iZyCyirL4oxcPl79w7B3wYBOTKwOezlbaufs1Q17fS915uy16+zhO5E2o+wQZC2r3UIiaymUFzlBsPMQSJzgPLqOAL9AyiqqWJV1gPlbnN7A77cfoKLK4u/rw/D4iJog2K9zOC4tFCMi0ugU/KSxKPi1IJVVFmstvi4fT5cix1Ad9ua4h3EeN+yt/8BZkAOg26jaBVoU9qQlqCiDncsh/VvnkbUUbCX4Bjk/nKgOgh0Hgo8PhaUVLMnIZUFaLgvScti4pwCAdoG+jE6KZkxPJwjGRwVroRgRkUag4CeNRYu7tCBLMvKY/OIifH0MgX4u98On5tcg97EA38Of17ap89zXRZB/7dcBfi53e6dt7bV8tJx7Axwe9vaSU1jqhL0+sVww0NlUPdjfF3LSIPUfPwx75z7hDOMM7+LR+xD5AV9/97y/FJj4C2cvwW0La4Pgl79z2gVFQsLphCZO4IzECZzRx9lHcF9BCalbnRA4f0sOn67bA0Dn8ECnN7BnNClJUcSGBXrm/kREROSkqMevCe3IK+L973dSUl5JcXklJeVVlJZXUlJRSXGZ87ykwv1reWWdRxXF5ZUn/b7+vj6HhcJAXxeB/i4CfX8YOgP9XATUfe5uE+RfG0jrBsvDQ6kP/i6fVtMLUFll3Qu07PpB2Dt/QCcm9qkT9ta/7/TuHdazd6nCnrR+BXsgY54TArd+AwW7nOPt42p7AxPGQ0gU1loyc4tYkObMD1y4NZf8YmcT3d4dwtxBMIoRCVGEBujniCIiDdHaevy+/fZb/P39GT16NABTp04lODiYG2+88ZSu++ijjxIaGspPfvKTxijzqEaPHs3ChQub9D2Op6nuVT1+LUi3yGDuP7PnSb3WWktpRRWl7nBYXFZ5WEgsLq90QmSd5zUBsqKSksOCZSXF7nMHispqgmVN0Kyooqyi6qTqNIba3kh3aHR6Iw8PiDVf+7oI8nd6LY/Wu1k3ZAbU9HQ61z/RYbPVYW/Omt18snYPOYWlBPr5cGafDj8Me4teh3X/hb1rnBd3G6mePfE+YR1h4NXOw1rITavtDVz3AayY5bTrOACTOIGExAkkDB3N9aPiqKyyrNuVXzMs9PXF25i5IANfH8Pgbu1regQHd2uPn4a4i4h4hW+//ZbQ0NCa4HfXXXd5uKLDVVRU4Ot79Ejj6dDXkij4NaXdq+GzX4JvIPgFOnNsqn/1DQC/IPc596912hm/QAJ9gwj0CyS8+lxgdfsQcDXuMuyVVZbSisofhsLDeiPrhsxKSivcz48IpXXbHiwpr+ndrHv9yqqT62n29TEEucPl4b2RPjXDZqtDaJWFuZuza8Kes89e59qwl7sVFj3l7tmrE/Z+9Gfoe4nCnng/YyC6p/MYcQdUVsDuVZD+jRMEF0+Dhf9yto7oNhJX4ngGJk5k4LjB3D0hiZLySpZv21/TI/jM11t4+qstBPu7GJkQWbN/YJ+OYa1mZICISFtw6aWXsmPHDkpKSnjggQeYMmUKAJ9++im//OUvqaysJDo6mpdeeompU6ficrl4/fXX+de//sVXX31FaGgoF154ITfeeCNLliwBnA3fL7roItasWcPy5ct5+OGHKSwsJDo6mldeeYVOnY6+yvnWrVv58Y9/THZ2NsHBwbz44ov06dOHjz76iMcee4yysjKioqJ444036NChA48++ihbt24lPT2d7t2707t3b7Zv3056ejrbt2/nwQcf5P777wcgNDSUwsJCvv32Wx599FGio6NZu3Ytw4YN4/XXX8cYw5w5c3j44YcJCQlhzJgxpKen87///e+wGisrK/n5z3/Op59+io+PD3fccQf33XffCd/rypUrueuuuygqKiIpKYmZM2cSERHBM888w9SpU/H19aVv3768/fbbzJ07lwceeAAAYwzz5s0jLCzspH/fFfyakq10llk/lA0VpVBRDOUltb9Wlp78tY2rAYHyGCHTz33c/VqXbxDBfoEE1wTQIAgKhHZ1QqnLz/lGsRGUV1bVHyxrhsE6PZFHDoE91rDZwtIKcgrLatqXV1pGJERw/oBOnNEn9vCwt/4D2OMOe11HuMPexRDetVHuT6RVcvlC12HO4/SfQFmRs1podY/g1485j4B2ED+OwMQJjEmcwJgf9YZz+5BfVE5qem5NEPxm0wYAokP9GZ3kLBIzukcUXSOCPXqbIiItxV+W/IWNeRsb9Zp9Ivvw8xE/P2abmTNnEhkZSXFxMcOHD+eKK66gqqqKO+64g3nz5pGQkEBeXh6RkZHcddddhw1T/Oqrr5z36dOHsrIyMjIySEhIYPbs2UyaNIny8nLuu+8+/vvf/xITE8Ps2bP51a9+xcyZM49az5QpU5g6dSo9e/Zk8eLF3HPPPXz99deMHTuWRYsWYYxhxowZPPnkkzUbwq9fv5758+cTFBTEo48+ysaNG/nmm28oKCigd+/e3H333fj5Hd5R8v3337Nu3To6d+7MmDFjWLBgAcnJydx555019z158uR6a5w+fTqZmZmsXLkSX19f8vLyTupeb7zxRv71r38xfvx4fvvb3/L73/+ep556iieeeIKMjAwCAgI4cOAAAH/729947rnnGDNmDIWFhQQGntr8egW/ptR5CNz66dHPV1U54a+8GCpKan+tKDk8INYcK27AuWInZBblHNHOfbyi5OTvx/gcv9fSL7BO72X9IRPfIPz8AvHzDSSsbruAIAip+9qgxgmbuVth8dPO1guHhb0/uXv2FPZE6uUfDD3OdB4Ah3Ihc15tENz0sXM8rDMkTiA8cQLnJo7n3P79Adh1oLgmBM5Py+XDVc58wvio4JptI1KSomgf7N/89yYi0oY988wzvP/++wDs2LGDLVu2kJ2dzemnn05CQgIAkZGRx73O1VdfzezZs3nkkUeYPXs2s2fPZtOmTaxdu5azzz4bcHrKjtUDVlhYyMKFC7nqqqtqjpWWOp0jWVlZTJo0id27d1NWVlZTG8DFF19MUFBQzfMLLriAgIAAAgICiI2NZe/evXTtevj3eCNGjKg5NnjwYDIzMwkNDSUxMbHm2pMnT2b69Ok/qPPLL7/krrvuqhlWGhkZydq1a0/oXvPz8zlw4ADjx48H4Kabbqq574EDB3Lddddx6aWXcumllwIwZswYHn74Ya677jouv/zyH9zPiVLw8yQfH/AJckJOczkybP4gSLoD4pFh9Fghs/q1dcNmxRHvcbKqw+Yxey2PEjJtFWz5HPasdq6lsCdyakKinO1L+l3mPM/LgIy5Tgjc/CmsetM5HtMHEifQOXECV/Ufw1XJ3bDWsmVfYc1G8h98v5M3Fm/HGOjfOdw9LDSK4fGR2kheRNqM4/XMNYVvv/2WL7/8ktTUVIKDg5kwYQIlJSf3vdqkSZO46qqruPzyyzHG0LNnT9asWUO/fv1ITU1t0DWqqqpo3749K1eu/MG5++67j4cffpiLL764ZqhmtZCQkMPaBgQE1HztcrmoqKj4wfUa0uZEWGtP6F6P5eOPP2bevHl89NFHPP7446xZs4ZHHnmECy64gDlz5jBmzBg+++wz+vTpc9LvoeDX1ngibFp7giHziEBZ77HqsJlb/zWqw2bX4Qp7Ik0lMsF5DLvZ+aHS3jW1vYHLZ8Hiqc6w9K7JmMQJ9EqcQK9Rydw6NoHyyipWZx1g/hZnaOhL89OZOncr/r4+JMdF1MwPHNBFG8mLiDSm/Px8IiIiCA4OZuPGjSxatAiAUaNGcc8999QM3awe6hkWFsbBgwfrvVZSUhIul4s//vGPTJo0CYDevXuTnZ1NamoqKSkplJeXs3nzZvr161fvNdq1a0dCQgL//ve/ueqqq7DWsnr1agYNGkR+fj5dujhrLsyaNasJPg2n3vT0dDIzM4mPj2f27Nn1tjv77LOZNm0aEydOrBnqeaL3Gh4eTkREBN999x3jxo3jtddeY/z48VRVVbFjxw4mTpzI2LFjefvttyksLCQ3N5cBAwYwYMAAli5dysaNGxX8pIUzxgmafkHQXHnTWmd+ZSMvgiMiR+HjA50GOY8xDzg/kNmxpDYIzvsrzP2LszhV3Gj8EicwLHECw87oywNn9eRQaQVLMvNYsCWH+Wk5/PWzTfz1s020C/QlJSmqJggmRodooRgRkVNw7rnnMnXqVE477TR69+7NqFGjAIiJiWH69OlcfvnlVFVVERsbyxdffMFFF13ElVdeyX//+1/+9a9//eB6kyZN4qc//SkZGRkA+Pv78+6773L//feTn59PRUUFDz744FHDEMAbb7zB3XffzWOPPUZ5eTnXXHMNgwYN4tFHH+Wqq64iIiKCM844o+Y9GlNQUBDPP/885557LiEhIQwfPrzedrfffjubN29m4MCB+Pn5cccdd3Dvvfee8L3OmjWrZnGXxMREXn75ZSorK7n++uvJz8/HWsv9999P+/bt+c1vfsM333yDj48P/fr147zzzjule9U+fiIi0vSKD0Dm/NogmLvFOR4S4+wbWL2HYPtuAOQUlrJwa25NENx5oBiATu6N5Mf0iGJMUjSx7bSRvIi0Lq1tH7+2oLCwkNDQUKy1/PjHP6Znz5489NBDni7ruLSPn4iItDxB7eG0C50HQP7O2vmB6d/C2ned45FJkDiB6MTxXNxzHBcPGoi1lu15RcxPy2FhWi5fbtjLu8uzAOjVIbRmxdCRiZGEBaqXX0RETsyLL77IrFmzKCsrY8iQIdx5552eLqlJqMdPREQ8y1rI3lgbAjPnQ1khYKDz4NrewG6jwC+QqirL+t0Hme9eMXRJRh6lFVW4qjeSdw8NHdI9An9fbSQvIi2LevyksZxoj5+Cn4iItCyV5bBzuTsIzoWsJc6cXd9A6D6qNgh2HAg+LkrKK1mxvXoj+VxWZx2gykKQn4uRiZGMSardSN5HC8WIiIdt2LCBPn36aL6ynBJrLRs3blTwExERL1JaCNsW1vYI7lvnHA9sDwmn1wbByEQwhvzichal57IwzZkfuDX7EABRIf6kJEUx1r1QTLdIbSQvIs0vIyODsLAwoqKiFP7kpFhryc3NpaCg4LC9DUHBT0REvEnBXsios5H8QWe+H+HdIdG9UEzCeAiNAWB3fjEL0mqD4L4CZ2PguKjgmvmBKUlRRIZoI3kRaXrl5eVkZWWd9N55IgCBgYF07doVP7/D57Yr+ImIiHeyFvLSIf0bJwRmzIOSfOdch/61vYHdUyDAWbEtbV8hC9JymJ+Wy6L0XApLKzAG+nZqV9MbODw+kiB/bSQvIiKti4KfiIi0DVWVsHtlbW/g9kVQWQY+ftBtRG0Q7DwUXL5UVFaxKiu/pjdwxfb9lFda/F0+DI1rXxMEB3QJx9elhWJERKRlU/ATEZG2qbzYCX/VQXD3KsCCfxgkjKvdQzCmNxhDUVkFSzLyWLg1l/lbcli/+yAAYYG+jEqsnR+YFKON5EVEpOVR8BMREQEoyjt8fuD+DOd4aMfa3sDE8dCuMwC5haWkpue6h4bmsCPP2Ui+Q7sAxvSIrgmCHbSRvIiItAAKfiIiIvXZv63ORvJzoSjHOR7duzYExo+FwHAAtucWsWBrjnsz+Rz2F5UD0CM2lLE9ojmvf0dGJESqN1BERDxCwU9EROR4qqqcrSKqewO3LYTyIjAu6DK0tkew63DwDaCqyrJhz8GahWKWZORSUl7FoG7tuev0RM7p1xGX9g0UEZFm1OKCnzHmXOBpwAXMsNY+ccT5m4G/Ajvdh5611s441jUV/EREpFFVlELW0toguHMF2ErwC4a40bVBMLYf+PhQUl7Ju8uzePG7dLblFpEQHcKU0xO5bEgXAv20QqiIiDS9FhX8jDEuYDNwNpAFLAUmW2vX12lzM5Bsrb23oddV8BMRkSZVkg+ZC2qDYM4m53hwVO0iMb3OpTIklk/X7mHq3K2s2ZlPTFgAt4yJ57qRcYQH+R3jDURERE5NSwt+KcCj1tofuZ//AsBa++c6bW5GwU9ERFqyg7uceYHVQbBwD7j8YdA1MPoBbFQSC7fmMnXuVr7bkkNogC/XjuzOrWMS6BiuxWBERKTxtbTgdyVwrrX2dvfzG4CRdUOeO/j9GcjG6R18yFq7o55rTQGmAHTv3n3Ytm3bmv4GREREjmQt7NsAS1+E799w9g487UIY8yB0TWbtznymz0vnf6t34fIxXDq4C3eOT6RHbJinKxcRES/SGoNfFFBorS01xtwJTLLWnnGs66rHT0REWoTCfbB4mhMCS/IhbiyMfRB6nMWO/cXM+C6d2ct2UFJexVmndeDuCYkMi4v0dNUiIuIFWlrwO+5QzyPau4A8a234sa6r4CciIi1KaQEsnwWpz0HBLmcRmDEPQP/LyS2u4tXUbcxKzeRAUTnJcRHcNT6JM/rE4qOVQEVE5CS1tODnizN880ycVTuXAtdaa9fVadPJWrvb/fVlwM+ttaOOdV0FPxERaZEqymDtu7DgacjeCOHdIOVeGHoDRQTwztIdvPhdBjsPFNMzNpQppydyyeAu+Pv6eLpyERFpZVpU8AMwxpwPPIWzncNMa+3jxpg/AMustR8aY/4MXAxUAHnA3dbajce6poKfiIi0aFVVsOUzmP8U7FgEQZEwYgqMmEJ5YAQfr97N1Llb2bingI7tArltbALXjOhGWKBWAhURkYZpccGvKSj4iYhIq7F9kdMDuGkO+AbB0Bsg5V5s++7M3ZzNtLnppKbnEhboyw2j4rhlTAIxYQGerlpERFo4BT8REZGWaN9GWPgvWD0bbBX0v9yZB9hxACt3HGDa3K18um4Pfi4frhjalSmnJ5IQHeLpqkVEpIVS8BMREWnJ8nfCoudh+StQVghJZzorgcaPIyO3iOnz0nlvRRbllVWc268jd41PYlC39h4uWkREWhoFPxERkdageD8smwmLpsKhfdB5qNMDeNpF7DtUzisLMnlt0TYKSioYlRjJXeOTGN8rBmO0EqiIiCj4iYiItC7lJbDqTVjwDOzPgMgkGH0fDJpMYZUvby3ezkvzM9hzsIQ+HcO4a3wSFw7shK9LK4GKiLRlCn4iIiKtUVUlbPjQWQl090oIiYVRd0HybZT5teO/K3cybV46afsK6dI+iNvHJTBpeDeC/X09XbmIiHiAgp+IiEhrZi1kzIMFT8HWr8E/DJJvhlH3UBXaia837mPq3K0s27af9sF+3JgSz82j44kM8fd05SIi0owU/ERERLzF7lXOENB1/wHjgoGTYMz9ENObZZl5TJ2bzpcb9hLo58Ok5G7cPi6RbpHBnq5aRESagYKfiIiIt9mfCQufhe9fh4pi6H2BsxJotxGk7Stg2tx0Pli5kyoLFwzoxJ3jE+nXOdzTVYuISBNS8BMREfFWh3JgyXTnUbwfuo92VgLteQ57CsqYuSCDNxdvp7C0gnE9o7l7fBIpSVFaCVRExAsp+ImIiHi70kL4/jWnF/BgFsSc5gTAAVeSXwZvLN7GzPmZ5BSWMqBLOHeNT+Lc/h1x+SgAioh4CwU/ERGRtqKyHNa+Bwuehn3roV1XSLkHht5EiU8Q73+/k+nz0snIOURcVDB3jEvkymFdCfRzebpyERE5RQp+IiIibY21sOULZyXQbQsgsD2MuANG3EllcDRfrN/DC3PTWbXjANGh/tw8Op4bRsUTHuzn6cpFROQkKfiJiIi0ZTuWOgFw48fgGwCDr4PR92Ej4lmckcfUuVv5dlM2wf4uJo/ozm1jE+jcPsjTVYuIyAlS8BMRERHI2eIMAV09G6oqoO+lzkqgnQaxYfdBps9L58NVuzDAxYM7c9f4JHp1CPNw0SIi0lAKfiIiIlLr4G5Y/AIsexlKD0LiRGchmMQJZB0o5qX5Gby9ZAfF5ZWc2SeWO8cnMTw+QiuBioi0cAp+IiIi8kMl+bBsJix6AQr3QqfBTgDsewn7iyt5bdE2XlmYSd6hMoZ2b8+d45M4+7QO+GglUBGRFknBT0RERI6uohRWvQ0Ln4HcNIiIh9H3weDrKLb+vLt8B9O/S2dHXjGJMSHceXoilw7pQoCvVgIVEWlJFPxERETk+KoqnQVgFjwFO5dDSAyMvBOG306FfzifrN3D1LlbWbfrILFhAdw6NoFrR3anXaBWAhURaQkU/ERERKThrHW2gJj/FKR9Af6hMOxmGHU3tl0X5qflMG1uOvPTcggL8OW6UXHcOiae2HaBnq5cRKRNU/ATERGRk7NnrbMS6Nr3wBgYcLUzDzC2D2uy8pk2bytz1uzG18eHy4d24Y7TE0mKCfV01SIibZKCn4iIiJyaA9sh9TlY8SqUF0Gv85wAGJfCttxDzPgug3eW7aCssopz+nbgzvFJDO0e4emqRUTaFAU/ERERaRxFebDkRVg8FYrzoNtIGPMg9DqXnKJyXl2YyazUbeQXlzMiIZK7xicysXestoIQEWkGCn4iIiLSuMqK4PvXIfVfTm9gdG8Ycz8MuJpDlT7MXrqDGd+lsyu/hN4dwrhzfCIXDeqMn8vH05WLiHgtBT8RERFpGpUVsO59Zx7g3jUQ1hlS7oFhN1PuG8JHq3YxbW46m/YW0Dk8kNvGJXLN8G6EBPh6unIREa+j4CciIiJNy1rY+pWzEmjmdxAYDsNvh5F3YUNi+HZTNlPnbmVxRh7hQX7cmBLHTaPjiQ4N8HTlIiJeQ8FPREREms/O5U4P4PoPweUPg691NoSPSuL77fuZNjedz9bvwd/lw1XJXbljXCJxUSGerlpEpNVrccHPGHMu8DTgAmZYa584SrsrgHeB4dbaY6Y6BT8REZEWJncrLHwGVr4FlWXQ92JnIZguQ9maXciM79J5b/lOKqqqOG9AJ+46PYkBXcM9XbWISKvVooKfMcYFbAbOBrKApcBka+36I9qFAR8D/sC9Cn4iIiKtVMFeZxXQpS9BaT4knO4EwKQz2FdQyssLM3k9dRsFpRWM6RHFXeOTGNsjWiuBioicoJYW/FKAR621P3I//wWAtfbPR7R7CvgC+CnwEwU/ERGRVq7kICx/BRY9DwW7oeMAJwD2vZSCcsubi7fz0vwM9hWU0q9zO+4cn8T5/Tviq5VARUQa5FjBzxP/knYBdtR5nuU+VsMYMxToZq39+FgXMsZMMcYsM8Ysy87ObvxKRUREpPEEtnO2fHhgFVz8LFSUwnu3wb+GErb6Fe5M6cR3P5/Ik1cMpKS8kvvf+p6Jf/+WV1MzKS6r9HT1IiKtWov7EZoxxgf4B/B/x2trrZ1urU221ibHxMQ0fXEiIiJy6nwDYOgNcM9iuOZNCO0Ac34CT/UnYP7fuLpfCF88NJ7pNwwjJjSA3/53HWP+8jVPf7mF/YfKPF29iEir1OKGehpjwoGtQKH7JR2BPODiYw331FBPERGRVspa2J7qrAS6+VPwC4ahN0HKj6F9N5Zm5jFt7la+3LCPID8Xk4Z34/ZxCXSNCPZ05SIiLUpLm+Pni7O4y5nATpzFXa611q47Svtv0Rw/ERGRtmHvemcl0DX/dgLhgCthzAPQoR+b9xYwfV46H3y/EwtcNLATd45P4rRO7TxdtUizs9ZyqKyS0ABfT5ciLUiLCn4AxpjzgadwtnOYaa193BjzB2CZtfbDI9p+i4KfiIhI23JgByx6wVkMpvwQ9DzHWQgmbjS7D5bw0ncZvLVkO4fKKpnQO4YbRsUxKjGKEH0TLF5sd34x87fkMD8thwVpueQUlhITFkDP2FB6dQijZ4dQesaG0atDKO2D/T1drnhAiwt+TUHBT0RExAsV5TnbQCyeCkU50HW40wPY+wLySyp5ffE2Xl6QQU5hGb4+hkHd2jMmKYqUpGiGxrUnwNfl6TsQOWkHS8pZtDWXBWk5fJeWQ3r2IQCiQ/0Z0yOanrGhZOYWsWVvAVv2FVJUZxEkBcK2ScFPREREWrfyYlj5Biz8F+zPhKiezgqhAydRYn1Zvm0/C9JyWLg1l9VZB6iyEODrw/D4SFKSohjTI5r+ndtpawhp0coqqvh+u/NneX5aDquy8qmssgT5uRiREMm4ntGM6RFN7w5h+Pgcvs9lVZVlV34xW/YWsmVfAZv3FioQtkEKfiIiIuIdKitgw39h/lOwZzWEdoRRd8OwmyGoPeD0kixJz2Ph1lwWbs1h454CAMICfBmZGMVodxDs1SFUm8SLR1lr2by3kPlpOczfks3ijDyKyirxMTCwa3vG9ohmbM9ohnSv03udvxPSv3UeeekQlQQxfdyP3hARDz61Pd0KhG2Lgp+IiIh4F2sh/RtnJdD0b8G4oMswSDgdEsdD1xHgFwhATmEpqVtzWbg1l9StOWTmFgHOcLmUpGhGJzlhsHtksIKgNLk9+SXuOXpOr152QSkAidEhjOnh9OilJEURHuTnvKAkHzLn14a9nM3O8eBoJ+ztz4CDO2vfwDcQonvWCYPuR2RC/YFwnxMEN+8tZMu+QtL2FnBIgbDVUvATERER77VrJWz4CDLmwc7lYCvBFQDdRzpBMGECdB4CLmfhl6z9RTVBcEFaDvvc33h3aR9U0xuYkhRFh3aBHrsl8R4FJeUsSs+rCXpp+5wdy6JCnHl6Y3tEM6ZnNF3aBzkvqCiDrKXuoPcN7Fzh/Jn2C4a40ZA4ERInQGxf8HEPXS7Jh+zNkL2xzmMT5O+oLcQV4A6EvSHmNPevfSAysebvBigQtnYKfiIiItI2lBx09gTMmAfpc2HvGue4f5j7m+bxThiM7Qc+Plhr2Zp9iNStzvzA1PRcDhSVA5AU4/TAjE6KYlRilL6hlQYpr6xi5Y4DfLfF6dVbueMAlVWWQD8fRiREMc7dq9eno3uenrWwbz1s/cYJe9sWQHkRGB+nFztxgvPoOhx8A06smNKCOoFwgxMGszfCge21bXz86g+EUUng8qtppkDYOij4iYiISNt0KBcyv4OMuU4YzE1zjgdHQfw4d4/geOebXGOoqrKs332Qhe4guMQ958oY6Ne5HaPdQ0OHx0dq6wgBnHl6W/YV1myzsDg9l0PueXoDurZnbA+nF3lYXESdeXpZtUM30+fCoX3O8aietUEvfmzNvNVGV1roDBmtDoLVj/3bAHc28PGFqB61QbD6EdUDfGtDnAJhy6LgJyIiIgLOwhiZ3znfbGfMrZ0b1a6LOwS6H+FdAaf3ZtWOAzXDQr/ffoCyyip8fQxDurevmSN42OIb4vX2Hixh/pbaeXrVw4Xjo4IZ29MZvpmSGE14sLvHrPjA4fP0crc4x0Nia4Ne4viaP3ceU1ZUfyDMy6AmEBqXe0GZI3oIo3se1iNprWXnAQXC5qbgJyIiInIka51VEat7AzPmQVGucy4y6fAgGBINQHFZJcu2Va8Ymssa99YRgX7O1hHVPYL9u4Tj8tFCMd6isLSCxem5NcM3t7jn6UWG+DM6KYpxPaMZnRRNt8hg5wUVpc48verhm7tWgK0CvxCIH1Mb9mL7QmtYUKi8GHK21BMI0537AmdoamRi7eqi1aEwuif4BdVcSoGwaSn4iYiIiBxPVZUz16o6BG5bAKUHnXMd+teGwLgxENgOgPzicpZkOAt3pG7NZdNe99YRgb6MqrN1RM9YbR3RmlT39DrbLDjz9CqqLAG+PoxIiKzZZuG0ju2ceXpVVbBvXW2P3raF7nl6rnrm6XlRcCkvcYZPVy8mUz2PMHersyANOIEwIr6eQNgL/INrLmWtZVd+CZv3FjjbTewtZLMC4QlT8BMRERE5UZUVsHul0yOYPhd2LIaKEueb+c5DareO6Daypkcju6CU1HRn24gFablsz6veOiKgZtuI0UnRdI8KPsYbS3Oz1pK2r7Bmm4VF6XkUllZgDAzsEl6z+ubQuAgC/dxDeg/sqA16GXPhULZzPLqXO+hNdHr3AsM9dFceVFF2lECYBlUV7kYGIuLqD4QBoTWXUiA8MQp+IiIiIqeqvMQZvpcxz/lGf+dy55tYl78T/qoXiukytGY1xB15RaSm57IwzVkspnouWNeIOltHJEYRq60jmt2+gyUs2JpTM3xz70Hn9yYuKtjp0XNv61ETHIr3O/P0qodv5m11jod2qO3RSxgP4V08cTutQ2W50xt4ZCDM2QJV5bXt2nc/IhD2gZheEBBW00SBsH4KfiIiIiKNrbQAti9y9/jMgz1rAOvM46q7dUSHAXW2jiisWSgmdWsuB0uc3o+esaGMTooiJckJgjWLgkijKSytYElGLvO35DI/LZvNe515ehHBfozuEV2zzcJh8/R2LK7t1dv1fZ15emPrzNM7rXXM02vJKiucjej3bTh8HmHOZqgsq20X3u2Hq4zG9DqsV7WtB8ImDX7GmBCg2FpbZYzpBfQBPrHWlh/npY1KwU9EREQ8qijP6RGqXiwmZ7NzPCji8K0jonuCMVRWWdbvOnzriOJyZ+uI/p3DGd3DGRY6PD6CYH9tHXGiKiqrWJV1gPlbnKC9Yvv+w+bpVQ/f7Nupzjy9vWsPn6dXUewM7e2aXBv0uiR71zy9lqyyAg5scwfCOr2EOVucYdfV2nWpJxD2Pmw7jLYSCJs6+C0HxgERwAJgKVBmrb3ulC58ghT8REREpEU5uLt2oZiMuZC/wzke1unwFUPbdwegrMIJKgvcw0K/376f8kqLn8swpFsEKe6hoYO7tcff18eDN9YyOT2qh5i/JZv5abksTs+lwD1Pr3/n8JptFoYdNk9v++H76RXlOMejezshL2niYYv5SAtRVekOhBuPGDa62Qnr1UI7QmyfH84jDI6saXIqgXBg1/DaP0stRFMHvxXW2qHGmPuAIGvtk8aYldbawad04ROk4CciIiItlrXOULaaIDivdjGQiIQ6QXA8hMYAztYRSzOrt47IYe3OfKosBPm5SI6PYEwPZ+uIfp3b7tYR+wpKWJhWu83CnoNOL1D3yGDG9IhmXE9n6GxESJ15ehnfQXr1PL1053hox8P302vX2RO3I6eqqgrytx8RCN2/lh+qbRcSW08g7AMhUTVNGhIIP3/odHp1CDuyCo9q6uD3PXAP8E/gNmvtOmPMGmvtgFO68AlS8BMREZFWw1pn+Fp1CMycD6X5zrnYvodvHeEerpZfVM7ijNyaIFg9R62de+uI6iDYw4u3jjhUWsGSjLya1Tc37nG2z2gf7MeYpOia4Zs1q6aWlxw+T2/3Smeenn/o4fP0Yvponp43q6qCg1lOADxsHuEmKCuobRcc7czZPHLYaEh0zZ+PuoFwbI9o/Fwtq/e9qYPfeOD/gAXW2r8YYxKBB62195/ShU+Qgp+IiIi0WpUVsGdVnT0EU93zy3yg0+A6W0eMqtn7bF9BCalbc1mYlsvC9Bx25DlD3GLCDt86omaxklaoorKK1Tvzmb8lh/lpOTXDX/19fRgRXztPr1/nuvP01tSZp5daZ57ecGfoZuIEZ289lxbQafOshYM7nRB4ZC9h9R6eAEGR9QfC0NgW9wODZlvV0xjjA4Raaw8et3EjU/ATERERr1FRClnLaucHZi11to7w8YNuI5whoQmnOwHGvdDIjryimoViFm7NJdu9dUS3yCBGJ0YzukcUKUlRxIa13K0jrLWk5xxiQZqzzcKirbXz9Pp1bsfYHjGM7RFNcnydeXr7tx2+n15RrnM85rTaHr240ZqnJw1nLRTsricQboCS/Np2d34HnQZ6rs56NHWP35vAXUAlzsIu7YCnrbV/PaULnyAFPxEREfFapYWwY5GzAEnGPNi9CmfriGDonlK7dUTHgeDjqtmQvHrriEXptVtH9OoQyugkZ4+6UYlRhAd5tucru6CUhVtzmO+ep7cr35mn1zUiiHE9oxnbI4aUpCgiq+fpFeVB5ne1++ntz3COh3U6fD+9dp08cTvizayFwr21gXDYTeAX5OmqDtPUwW+ltXawMeY6YCjwCLDcWtus8VfBT0RERNqM6s3Eq4eGZm90jge2d+auVfcIxvSu2Tpi3a78miC4NDOPkvIqfAz07xLO6CRnfuDw+EiC/Jt2lcKiMvc8Pffwzep5euFBfozp4cxVHNcj5oh5eovq7Ke3ErDgH1Y7Ty9pIkT3anHD7kSaW1MHv3XAYOBN4Flr7VxjzCpr7aBTuvAJUvATERGRNqtgj7NaZfUegge2OcdDOxy+dUREPAClFZWs2pFfs5H89zvqbB3RPYIxSc7Q0EFdT33riIrKKtbszK8Zvrmiep6ey4fk+IiabRZqVietqoI9q2uD3vZUZ882H19nnl5i9Ty9oZqnJ3KEpg5+9wM/B1YBFwDdgdetteNO6cInSMFPRERExG1/5uFbRxTudY63j6vdNiLhdAjrADi9cEsz9ztzBNNyWbsrH2sh2N/F8PhIRrv3EDytU7vjbh1hrSXDPU9vvntPwgL3MFNnnl40Y3tGkxxXp3dxf+bh++kV5znHY/sePk8voGUtnS/S0jTb4i513tDXWlvR6Bc+BgU/ERERkXpY6yxMUb1QTOZ3tQtUxPSpDYLxYyAoAnC2jkhNzyXVvVjMln3O1hHhQX6MSoys2ToiKcbZOiKnsJQF7i0WFqTlsvOAs8Jol/bOPL3q9lGhAc77FuU5tVSHvf2ZzvGwTrU9eonjIaxjs31MIt6gqXv8woHfAae7D80F/mCtzT/6qxqfgp+IiIhIA1RVOkMpM+Y5vWvbU6G8CDDQaVDt1hHdU8A/BIB9B0tITc/9QbCLDQsgMsS/Zp5eu0BfxvSo3U8vLirY2VOwvBi215mnV704jX8YJIxzB72JEN1T8/RETkFTB7/3gLXALPehG4BB1trLT+nCJ0jBT0REROQkVJTBzuW1PYI7lkBVubN1RNfk2mGhXZPB1+mx257rbB2xYGsuB4rKGJUYxdge0fTvUj1Pr/KIeXqL6szTG1G7n17noeDy9eTdi3iVZlnV83jHmpqCn4iIiEgjKCs6YuuIlWCrwDcI4lJqF4rpNBh86qwAmpfhDnrfOK8r3u8cj+13xDy90Ga/JZG24ljBrzF+xFJsjBlrrZ3vfrMxQPFxCjoXeBpwATOstU8ccf4u4Mc4ewMWAlOstesboVYRERERORb/YEg6w3kAFB+AbQtqF4r58lHneEC4s51CcOThK4m26wK9z3eGbtZZQEZEPKsxevwGAa8C4e5D+4GbrLWrj9LeBWwGzgaycDZ9n1w32Blj2llrD7q/vhi4x1p77rHqUI+fiIiISDMo3FdnxdC5ULTfCYDVwzejemienoiHNGmPn7V2FTDIGNPO/fygMeZBoN7gB4wA0qy16e7i3gYuAWqCX3XocwsBGn/pURERERE5caGxMOBK5yEircap7chZh7X2YJ3A9vAxmnYBdtR5nuU+dhhjzI+NMVuBJ4H767uQMWaKMWaZMWZZdnb2SVYuIiIiIiLi3Rot+B3hlPv3rbXPWWuTcDaH//VR2ky31iZba5NjYmJO9S1FRERERES8UlMFv2MNzdwJdKvzvKv72NG8DVzaCDWJiIiIiIi0SSc9x88YU0D9Ac8AQcd46VKgpzEmASfwXQNce8S1e1prt7ifXgBsQURERERERE7KSQc/a23YSb6uwhhzL/AZznYOM62164wxfwCWWWs/BO41xpwFlONeJfRk6xQREREREWnrGmMfvxNmrZ0DzDni2G/rfP1AsxclIiIiIiLipZpqjp+IiIiIiIi0EAp+IiIiIiIiXk7BT0RERERExMsp+ImIiIiIiHg5BT8REREREREvp+AnIiIiIiLi5RT8REREREREvJyCn4iIiIiIiJdT8BMREREREfFyCn4iIiIiIiJeTsFPRERERETEyyn4iYiIiIiIeDkFPxERERERES+n4CciIiIiIuLlFPxERERERES8nIKfiIiIiIiIl1PwExERERER8XIKfiIiIiIiIl5OwU9ERERERMTLKfiJiIiIiIh4OQU/ERERERERL6fgJyIiIiIi4uUU/ERERERERLycgp+IiIiIiIiXU/ATERERERHxcgp+IiIiIiIiXk7BT0RERERExMt5JPgZY841xmwyxqQZYx6p5/zDxpj1xpjVxpivjDFxnqhTRERERETEGzR78DPGuIDngPOAvsBkY0zfI5p9DyRbawcC7wJPNm+VIiIiIiIi3sMTPX4jgDRrbbq1tgx4G7ikbgNr7TfW2iL300VA12auUURERERExGt4Ivh1AXbUeZ7lPnY0twGf1HfCGDPFGLPMGLMsOzu7EUsUERERERHxHi16cRdjzPVAMvDX+s5ba6dba5OttckxMTHNW5yIiIiIiEgr4euB99wJdKvzvKv72GGMMWcBvwLGW2tLm6k2ERERERERr+OJHr+lQE9jTIIxxh+4BviwbgNjzBBgGnCxtXafB2oUERERERHxGs0e/Ky1FcC9wGfABuAda+06Y8wfjDEXu5v9FQgF/m2MWWmM+fAolxMREREREZHj8MRQT6y1c4A5Rxz7bZ2vz2r2okRERERERLxUi17cRURERERERE6dgp+IiIiIiIiXU/ATERERERHxcgp+IiIiIiIiXk7BT0RERERExMsp+ImIiIiIiHg5BT8REREREREvp+AnIiIiIiLi5RT8REREREREvJyCn4iIiIiIiJdT8BMREREREfFyCn4iIiIiIiJeTsFPRERERETEyyn4iYiIiIiIeDkFPxERERERES+n4CciIiIiIuLlFPxERERERES8nIKfiIiIiIiIl1PwExERERER8XIKfiIiIiIiIl5OwU9ERERERMTLKfiJiIiIiIh4OQU/ERERERERL6fgJyIiIiIi4uUU/ERERERERLycgp+IiIiIiIiX80jwM8aca4zZZIxJM8Y8Us/5040xK4wxFcaYKz1Ro4iIiIiIiLdo9uBnjHEBzwHnAX2BycaYvkc02w7cDLzZvNWJiIiIiIh4H18PvOcIIM1amw5gjHkbuARYX93AWpvpPlflgfpERERERES8iieGenYBdtR5nuU+dsKMMVOMMcuMMcuys7MbpTgRERERERFv06oXd7HWTrfWJltrk2NiYjxdjoiIiIiISIvkieC3E+hW53lX9zERERERERFpAp4IfkuBnsaYBGOMP3AN8KEH6hAREREREWkTmj34WWsrgHuBz4ANwDvW2nXGmD8YYy4GMMYMN8ZkAVcB04wx65q7ThEREREREW/hiVU9sdbOAeYccey3db5eijMEVERERERERE5Rq17cRURERERERI5PwU9ERERERMTLKfiJiIiIiIh4OQU/ERERERERL6fgJyIiIiIi4uUU/ERERERERLycgp+IiIiIiIiXU/ATERERERHxcgp+IiIiIiIiXk7BT0RERERExMsp+ImIiIiIiHg5BT8REREREREvp+AnIiIiIiLi5RT8REREREREvJyCn4iIiIiIiJdT8BMREREREfFyCn4iIiIiIiJeTsFPRERERETEyyn4iYiIiIiIeDkFPxERERERES+n4CciIiIiIuLlFPxERERERES8nIKfiIiIiIiIl1PwExERERER8XIKfiIiIiIiIl7O19MFeLOV+1byyHePAGAwNceNMYcdO/J5XT9oW891TqTtke9R9xpHvq7e15gTaHuc+6231iOuf8z3OJG2rVRrr/9kWaxnX2899/6nWvuxXn60ax/tfo9Vy1HPHfXwiV/rpOo6wd+7k7nWUeut73i9h354sL73qrddA1/b4Pqa672b4T0ay5H/t9bb5jj/Njfk3+6GvE9jXMOTtTbmvy0n8+egUf4On+T7N+q/LSfw3vV9j1X3+NF+v4/6umN839nQ1x7t/Km8f33fmzbJ+x+nrj+O+SOdQjvVe82WyCPBzxhzLvA04AJmWGufOOJ8APAqMAzIBSZZazObu85TFeYfxtDYoYf9Ra7+uvovcc25On+na9ocrW0dR56rr+2R73Hk9eu9zrHqOpG29bym5usj/iE7su2x3qMhbZv7mwWpZbEeD62n+k3VqdZ/Kq9v0tqPcuqo/xke41pH/SbgJO79RK5lMEf/jMxJ3MsxP64Tu1Z9ddXbtt5DDXttQ9+joe/bJO/dwD/DjX5/x9CQH6o0qM1x/m855R/eNOA9Gvo+jVFrg+7H4tF/W07m38yTuVaz/NtyCv+u1vd93WHH6/k+rUGvO/J9jtLmmN+/neL7N+g6TfD+R/uetbVq9uBnjHEBzwFnA1nAUmPMh9ba9XWa3Qbst9b2MMZcA/wFmNTctZ6qpPZJ/GncnzxdhoiIiIiItHGemOM3Akiz1qZba8uAt4FLjmhzCTDL/fW7wJmmMcZEiIiIiIiItEGeCH5dgB11nme5j9XbxlpbAeQDUUdeyBgzxRizzBizLDs7u4nKFRERERERad1a9aqe1trp1tpka21yTEyMp8sRERERERFpkTwR/HYC3eo87+o+Vm8bY4wvEI6zyIuIiIiIiIicIE8Ev6VAT2NMgjHGH7gG+PCINh8CN7m/vhL42nrLcjoiIiIiIiLNrNlX9bTWVhhj7gU+w9nOYaa1dp0x5g/AMmvth8BLwGvGmDQgDyccioiIiIiIyEnwyD5+1to5wJwjjv22ztclwFXNXZeIiIiIiIg3atWLu4iIiIiIiMjxKfiJiIiIiIh4OeMta6YYY7KBbZ6uox7RQI6ni2jD9Pl7jj57z9Fn7zn67D1Hn73n6LP3HH32ntNSP/s4a229+9x5TfBrqYwxy6y1yZ6uo63S5+85+uw9R5+95+iz9xx99p6jz95z9Nl7Tmv87DXUU0RERERExMsp+ImIiIiIiHg5Bb+mN93TBbRx+vw9R5+95+iz9xx99p6jz95z9Nl7jj57z2l1n73m+ImIiIiIiHg59fiJiIiIiIh4OQU/ERERERERL6fg14SMMecaYzYZY9KMMY94up62xBgz0xizzxiz1tO1tCXGmG7GmG+MMeuNMeuMMQ94uqa2whgTaIxZYoxZ5f7sf+/pmtoaY4zLGPO9MeZ/nq6lrTHGZBpj1hhjVhpjlnm6nrbEGNPeGPOuMWajMWaDMSbF0zW1BcaY3u4/79WPg8aYBz1dV1thjHnI/X/tWmPMW8aYQE/X1BCa49dEjDEuYDNwNpAFLAUmW2vXe7SwNsIYczpQCLxqre3v6XraCmNMJ6CTtXaFMSYMWA5cqj/3Tc8YY4AQa22hMcYPmA88YK1d5OHS2gxjzMNAMtDOWnuhp+tpS4wxmUCytbYlbqbs1Ywxs4DvrLUzjDH+QLC19oCHy2pT3N9z7gRGWmu3eboeb2eM6YLzf2xfa22xMeYdYI619hXPVnZ86vFrOiOANGtturW2DHgbuMTDNbUZ1tp5QJ6n62hrrLW7rbUr3F8XABuALp6tqm2wjkL3Uz/3Qz/ZaybGmK7ABcAMT9ci0lyMMeHA6cBLANbaMoU+jzgT2KrQ16x8gSBjjC8QDOzycD0NouDXdLoAO+o8z0LfAEsbYoyJB4YAiz1cSpvhHmq4EtgHfGGt1WfffJ4CfgZUebiOtsoCnxtjlhtjpni6mDYkAcgGXnYPc55hjAnxdFFt0DXAW54uoq2w1u4E/gZsB3YD+dbazz1bVcMo+IlIozPGhALvAQ9aaw96up62wlpbaa0dDHQFRhhjNMy5GRhjLgT2WWuXe7qWNmystXYocB7wY/dwf2l6vsBQ4AVr7RDgEKA1DZqRe3jtxcC/PV1LW2GMicAZxZcAdAZCjDHXe7aqhlHwazo7gW51nnd1HxPxau75Ze8Bb1hr/+Ppetoi91Crb4BzPVxKWzEGuNg9z+xt4AxjzOueLaltcf8EHmvtPuB9nOkW0vSygKw6owvexQmC0nzOA1ZYa/d6upA25Cwgw1qbba0tB/4DjPZwTQ2i4Nd0lgI9jTEJ7p/GXAN86OGaRJqUe4GRl4AN1tp/eLqetsQYE2OMae/+OghnYamNHi2qjbDW/sJa29VaG4/zb/3X1tpW8dNfb2CMCXEvJoV7mOE5gFZ0bgbW2j3ADmNMb/ehMwEt5tW8JqNhns1tOzDKGBPs/r7nTJw1DVo8X08X4K2stRXGmHuBzwAXMNNau87DZbUZxpi3gAlAtDEmC/idtfYlz1bVJowBbgDWuOeaAfzSWjvHcyW1GZ2AWe7V3XyAd6y12lZA2oIOwPvO91/4Am9aaz/1bEltyn3AG+4fcqcDt3i4njbD/YOOs4E7PV1LW2KtXWyMeRdYAVQA3wPTPVtVw2g7BxERERERES+noZ4iIiIiIiJeTsFPRERERETEyyn4iYiIiIiIeDkFPxERERERES+n4CciIiIiIuLlFPxEREQAY0ylMWZlnccjjXjteGOM9pYTERGP0T5+IiIijmJr7WBPFyEiItIU1OMnIiJyDMaYTGPMk8aYNcaYJcaYHu7j8caYr40xq40xXxljuruPdzDGvG+MWeV+jHZfymWMedEYs84Y87kxJsjd/n5jzHr3dd720G2KiIiXU/ATERFxBB0x1HNSnXP51toBwLPAU+5j/wJmWWsHAm8Az7iPPwPMtdYOAoYC69zHewLPWWv7AQeAK9zHHwGGuK9zV9PcmoiItHXGWuvpGkRERDzOGFNorQ2t53gmcIa1Nt0Y4wfssdZGGWNygE7W2nL38d3W2mhjTDbQ1VpbWuca8cAX1tqe7uc/B/ystY8ZYz4FCoEPgA+stYVNfKsiItIGqcdPRETk+OxRvj4RpXW+rqR2nv0FwHM4vYNLjTGafy8iIo1OwU9EROT4JtX5NdX99ULgGvfX1wHfub/+CrgbwBjjMsaEH+2ixhgfoJu19hvg50A48INeRxERkVOlnyqKiIg4gowxK+s8/9RaW72lQ4QxZjVOr91k97H7gJeNMT8FsoFb3McfAKYbY27D6dm7G9h9lPd0Aa+7w6EBnrHWHmik+xEREamhOX4iIiLH4J7jl2ytzfF0LSIiIidLQz1FRERERES8nHr8REREREREvJx6/ERERERERLycgp+IiIiIiIiXU/ATERFp5YwxrxhjHvPA+35ijLmpud9XREROnIKfiIicFGNMpjGmzBgTfcTx740x1hgT735+WCgxxgQYY/5sjNlujCk2xmwxxvzUGGOOuPZZR1z3ZmPM/Ca+LYwxjxpjXm/q92mp738irLXnWWtneboOERE5PgU/ERE5FRnU7muHMWYAEHyc1/wbOBM4HwgDbgCmAE83UY3NyhijPXJFRKTFUfATEZFT8RpwY53nNwGvHq2xMeZM4BzgCmvtWmtthbV2EXA98GNjTI+TLcQYc4cxJs0Yk2eM+dAY07nOOWuMucvdu3jAGPNc3R7GOu3OBX4JTDLGFBpjVrmP32KM2WCMKTDGpBtj7qzzmgnGmCxjzM+NMXtwNnUPMsbMMsbsd7/uZ8aYrDqv6WyMec8Yk22MyTDG3H+s96+nziHGmBXuemYDgSf4Wdzj/iwKjDF/NMYkGWMWGmMOGmPeMcb4u9tGGGP+565zv/vrrnWu9a0x5nb31zcbY+YbY/7mbpthjDnvRH4PRUSk6Sj4iYjIqVgEtDPGnGaMcQHXAMcapng2sNhau6PuQWvtYiALpyfwhBljzgD+DFwNdAK2AW8f0exCYDgw0N3uR0dex1r7KfAnYLa1NtRaO8h9ap/79e2AW4B/GmOG1nlpRyASiMPpvfwdEA8kuu/5+jq1+gAfAauALu57ftAY86NjvH/de/UHPsAJ3ZE4PahXnOBn8SNgGDAK+Bkw3V1jN6A/tb24PsDL7vvqDhQDzx5ZUx0jgU1ANPAk8FJ9AVtERJqfgp+IiJyq6l6/s4ENwM5jtI0Gdh/l3G73+ZNxHTDTWrvCWlsK/AJIqZ5n6PaEtfaAtXY78A0wuKEXt9Z+bK3dah1zgc+BcXWaVAG/s9aWWmuLcULXn6y1+621WcAzddoOB2KstX+w1pZZa9OBF3FCc0OMAvyAp6y15dbad4Gldc435LN40lp70Fq7DlgLfG6tTbfW5gOfAEPc951rrX3PWltkrS0AHgfGH6O2bdbaF621lcAsnODZoYH3JSIiTUjzEERE5FS9BswDEjjGME+3HKDnUc51cp8HqMAJN3X5AeVHeW1nYEX1E2ttoTEmF6dHLdN9eE+d9kVA6HFqreEesvg7oBfOD02DgTV1mmRba0uOqKdur2bdr+OAzsaYA3WOuYDvGlhOZ2CntdbWObbtiPPH+yz21mlfXM/zjgDGmGDgn8C5QIT7fJgxxuUOd0eq+YyttUXuzr4Gf84iItJ01OMnIiKnxFq7DWeRl/OB/xyn+ZfASGNMt7oHjTEjcYYZfu0+tB1nqGRdCRwecOrahROoqq8XAkRx7N7Ho6kbqDDGBADvAX8DOlhr2wNzAHO01+D0Xnat87zu/e4AMqy17es8wqy15x/lWkfaDXQ5Yghl9zpfN+Zn8X9Ab2CktbYdcHr1ZU/iWiIi4kEKfiIi0hhuA86w1h46ViNr7ZfAV8B7xph+xhiXMWYUzrzAF6y1W9xNZ+PMe+tjHMnArfxwrlq1t4BbjDGD3UHtTzhzCTNP4l72AvHuuXgA/kAAkA1UuHv/zjnONd4BfuFeHKULcG+dc0uAAvdiMEHuz6C/MWb4Ud7/SKk4PaL3G2P8jDGXAyPqnG/MzyIMpwfwgDEmEqfXU0REWiEFPxEROWXu+W/LGtj8Cpw5dp8ChTih7yXgvjptXsRZVOQjIB9nCOmv3Iuf1Pf+XwK/wemZ2w0k0fA5c0f6t/vXXGPMCvfctvtxwtx+4Frgw+Nc4w84i9Vk4PRyvguUumutxFkoZrD7fA4wAwiv7/2PvLC1tgy4HLgZyAMmUaentZE/i6eAIHeNi3B+z0REpBUyh08REBERkcZmjLkbuMZae6yFUURERJqMevxEREQamTGmkzFmjDHGxxjTG2eu3PuerktERNoureopIiLS+PyBaTgL0hzAmZv4vCcLEhGRtk1DPUVERERERLychnqKiIiIiIh4Oa8Z6hkdHW3j4+M9XYaIiIiIiIhHLF++PMdaG1PfOa8JfvHx8Sxb1tCVxEVERERERLyLMWbb0c5pqKeIiIiIiIiXU/ATERERERHxcgp+IiIiIiIiXk7BT0RERERExMsp+ImIiIiIiHg5BT8REREREREvp+AnIiIiIiLi5RT8REREREREvJyCn4iIiIhIK3KwpJx/frGZ22ctZd/BEk+XI62Er6cLEBERERGR4ztUWsErCzOZPi+d/OJy/F0+XDUtlddvG0m3yGBPlyctnIKfiIiIiEgLVlJeyeuLtvHCt1vJPVTGmX1ieejsXpRXVnHzy0u5eloqr902kh6xoZ4uVVowBT8RERERkRaotKKS2Ut38OzXaewrKGVcz2geOrsXQ7tH1LR5e8oobnhpCZOmpTLr1hH07xLuwYqlJTPWWk/X0CiSk5PtsmXLPF2GiIiIiMgpKa+s4r3lWfzr6zR2HihmRHwk/3dOL0YmRtXbPiPnENfPWMzB4nJm3jKc4fGRzVyxtBTGmOXW2uR6zyn4iYiIiIh4XmWV5b8rd/LUl1vYnlfE4G7t+b9zejG2RzTGmGO+dteBYq6fsZhd+cVMuyGZ8b1imqlqaUmOFfy0qqeIiIiIiAdVVVn+t3oX5/xzLg+/s4qwQF9euimZ9+8ZzbieMccNfQCd2wfxzl0pJEaHcvuspXyyZnczVC6tiYKfiIiIiIgHWGv5fN0ezn/mO+5983tcPoap1w/lo3vHcuZpHRoU+OqKDg3grSmjGNi1PT9+cwXvLs9qosqlNdLiLiIiIiIizchay9zN2fzji82szsonITqEp68ZzIUDO+PyObGwd6TwID9eu20Ed762nJ/8exWFJeXcPCahkSqX1kzBT0RERESkmSzcmsPfP9/M8m376RoRxF+vHMhlQ7rg62q8gXjB/r7MuCmZ+9/6nkc/Wk9BSQX3ntHjhHsQxbso+ImIiIiINLFlmXn8/fPNpKbn0rFdII9f1p+rhnXD37dpZl4F+Lp47tqh/Oy91fz9i80UlFbwi/P6KPy1YQp+IiIiIiJNZHXWAf7++Wbmbs4mOjSA313Ul8kjuhPo52ry9/Z1+fC3KwcRFuDL9HnpFJRU8Nil/U95OKm0Tgp+IiIiIiKNbMPug/zji818sX4vEcF+/OK8PtyQEkewf/N+++3jY3j04n6EBfrx7DdpFJZW8I+rB+HXiENLpXVQ8BMRERERaSRp+wr455db+Hj1bsICffm/s3txy9gEQgM89223MYaf/Kg3YYG+/PmTjRwqreD564Y2S6+jtBwKfiIiIiIip2hb7iGe/nILH6zcSZCfi/vO6MHtYxMJD/bzdGk17hyfRGigL7/+YC03zVzCjJuSCQtsOfVJ01LwExERERE5SVn7i3j26zT+vTwLP5fhjnGJ3Dk+icgQf0+XVq/rRsYRGuDLw++s4voZi3nllhFEtNBapXEp+ImIiIiInKC9B0t47ps03lqyHYPhhlFx3DMxidiwQE+XdlyXDO5CiL8v97y5gknTU3n9tpHEtmv5dcupMdZaT9fQKJKTk+2yZcs8XYaIiIiIeLGcwlKmfruV1xZto7LKcvXwbtw7sQed2wd5urQTtnBrDnfMWkZUaABv3D6SbpHBni5JTpExZrm1Nrnecwp+IiIiIiLHdqCojOnz0nllYSYl5ZVcPrQr95/Rk+5RrTssrdxxgJtmLiHIz8Xrt4+gR2yYp0uSU6DgJyIiIiJyEg6WlPPSdxnMnJ9BYVkFFw3szANn9SQpJtTTpTWaTXsKuP6lxVRWWV69dQT9u4R7uiQ5SQp+IiIiIiIn4FBpBbNSM5k2N5384nLO7deRh87uRe+O3tkjlplziOtmLOZgcTkzbxnO8PhIT5ckJ0HBT0RERESkAUrKK3l90TZe+HYruYfKOKNPLA+f3atN9ILtzi/m+hmL2XmgmGk3JDO+V4ynS5ITpOAnIiIiInIMpRWVzF66g+e+SWPvwVLG9ojm4XN6MbR7hKdLa1a5haXcOHMJm/cW8Mw1QzhvQCdPlyQnQMFPRERERKQe5ZVV/GdFFs98lcbOA8WMiI/k4XN6MSoxytOleUx+cTm3vbKUFdv388QVA7k6uZunS5IGOlbw0z5+IiIiItLmVFZZ/rtyJ09/tYVtuUUM6taeJ64YwNge0RhjPF2eR4UH+fHqbSO487Xl/Ozd1RSWVHDr2ARPlyWnSMFPRERERNqMqirLnLW7eerLLaTtK6Rvp3a8dFMyZ/SJbfOBr65gf19m3JTMA2+t5A//W09BSQX3n9lDn1ErpuAnIiIiIl7PWssX6/fyjy82s3FPAT1jQ3nhuqH8qF9HfHwUZuoT4Ovi2WuH8PP31vDPLzdTUFLOry44TeGvlVLwExERERGvZa1l7uZs/vHFZlZn5ZMQHcLT1wzmwoGdcSnwHZevy4e/XjmQsEBfZszPoLC0gscvG6DPrhVq0uBnjDkXeBpwATOstU8ccf5h4HagAsgGbrXWbnOfexK4APABvgAesN6yEo2IiIiINLmFW3P4x+ebWbZtP10jgnjyyoFcPqQLvi4fT5fWqvj4GH53UV/aBfryzNdpFJZW8I+rB+Pvq8+xNWmy4GeMcQHPAWcDWcBSY8yH1tr1dZp9DyRba4uMMXcDTwKTjDGjgTHAQHe7+cB44NumqldEREREvMPybXn8/fPNLNyaS8d2gTx2aX+uTu6moHIKjDE8fE5vwgL9eHzOBg6VVvDC9cMI9HN5ujRpoKbs8RsBpFlr0wGMMW8DlwA1wc9a+02d9ouA66tPAYGAP2AAP2BvE9YqIiIiIq3c6qwD/OOLzXy7KZvo0AB+e2Ffrh3ZXeGkEd1xeiKhgb788v013DRzCTNuSiYs0M/TZUkDNGXw6wLsqPM8Cxh5jPa3AZ8AWGtTjTHfALtxgt+z1toNR77AGDMFmALQvXv3RipbRERERFqTDbsP8s8vNvP5+r1EBPvxyHl9uDEljmB/LWfRFCaP6E5IgC8Pz17JdTMWM+uWEUSE+Hu6LDmOFvG3wRhzPZCMM5wTY0wP4DSgq7vJF8aYcdba7+q+zlo7HZgOzgbuzVexiIiIiHha2r4C/vnlFj5evZuwQF8ePrsXt4yJVw9UM7h4UGdCA1zc/foKrp6Wyuu3j6RDu0BPlyXH0JTBbyfQrc7zru5jhzHGnAX8ChhvrS11H74MWGStLXS3+QRIAb478vUiIiIi0rZsyz3E019u4YOVOwnyc3HvxB7cMS6R8GAFvuZ0Rp8OvHLLCG6ftZSrpqbyxu0j6RYZ7Omy5CiacobrUqCnMSbBGOMPXAN8WLeBMWYIMA242Fq7r86p7cB4Y4yvMcYPpyfwB0M9RURERKTt2HmgmEfeW80Zf5/LnLW7uX1cIvN+NpGf/Ki3Qp+HpCRF8cYdozhYUs6VUxeyZW+Bp0uSozBNuUOCMeZ84Cmc7RxmWmsfN8b8AVhmrf3QGPMlMABnLh/Admvtxe4VQZ8HTsdZ6OVTa+3Dx3qv5ORku2zZsqa6FRERERHxkL0HS3jumzTeXuIsH3HtyO7cMyGJWA0tbDE27Sng+pcWU1FZxau3jmRA13BPl9QmGWOWW2uT6z3nLVvjKfiJiIiIeJecwlKmfruV1xZto7LKclVyN+47owed2wd5ujSpx7bcQ1w3YzEHisp56aZkRiZGebqkNkfBT0RERERajQNFZUyfl84rCzMpKa/ksiFdeeDMnnSP0vyxlm53fjHXz1hM1v5ipt4wjIm9Yz1dUpui4CciIiIiLV5BSTkvzc/gpe8yKCyr4MKBnXnwrJ4kxYR6ujQ5AbmFpdz08hI27SngqUlDuGBgJ0+X1GYcK/i1iO0cRERERKTtKiqr4JWFmUyfl86BonLO7deRh87uRe+OYZ4uTU5CVGgAb94xitteWcp9b63gUOlArh7e7fgvlCal4CciIiIiHlFSXsnri7Yxde5WcgrLOKNPLA+f3Yv+XbQwSGvXLtCPV28dyV2vL+dn762moLSC28YmeLqsNk3BT0RERESaVWlFJe8s3cGz36Sx92ApY3tE89DZvRgWF+Hp0qQRBfm7ePHGZB6c/T1//N96CkrKeeDMnhhjPF1am6TgJyIiIiLNoryyiv+syOKZr9LYeaCYEfGRPH3NEEZp9Uev5e/rwzPXDCHEfw1PfbmFgpIKfn3BaQp/HqDgJyIiIiJNqrLK8uGqnTz95RYyc4sY1K09f758AON6RisAtAG+Lh/+csVAQgN9eWl+BoUlFfzp8gG4fPR735wU/ERERESkSVRVWT5Zu4d/frmZtH2F9O3Ujhk3JnPmabEKfG2Mj4/htxf2JSzQj2e+2kJhaQX/nDQYf18fT5fWZij4iYiIiEijstby5YZ9/OOLzWzYfZCesaG8cN1QftSvIz7q5WmzjDE8fHYv2gX68tjHGzhUVsEL1w0jyN/l6dLaBAU/EREREWkU1lrmbcnhH59vYlVWPvFRwTw1aTAXDeqsYX1S4/ZxiYQG+PKL99dw08wlvHRzMmGBfp4uy+sp+ImIiIjIKUvdmsvfP9/Esm376dI+iCevHMjlQ7rg69JQPvmha0Z0JyTAl4dmr+TaFxcz69YRRIb4e7osr6bgJyIiIiInbfm2PP7++WYWbs2lQ7sA/nhpfyYld9PcLTmuiwZ1JjTAl7teX86kaam8dttIOoYHerosr6W/kSIiIiJywqqqLL/5YC1XvJDK5r0F/ObCvsz96URuGBWn0CcNNrFPLLNuHcHu/BKumraQ7blFni7Ja+lvpYiIiIickKoqyy/+s4bXFm3j1jEJzPvZRG4bm0CgnxbpkBM3KjGKN+8YSUFJBVdOXcjmvQWeLskrKfiJiIiISINVVll+/t5qZi/bwf1n9OA3F55GsL9mD8mpGdi1Pe/cmQLApGmprM464NmCvJCCn4iIiIg0SGWV5afvruLfy7N48KyePHxOb+3HJ42mV4cw3r1rNKGBvlz74mIWp+d6uiSvouAnIiIiIsdVWWX5yb9X8Z8VO3norF48eFYvT5ckXqh7VDD/vnM0HcMDuXHmEr7ZuM/TJXkNBT8REREROaaKyioefmcl73+/k5+c04sHzurp6ZLEi3UMD+SdO1Po1SGMO15dxkerdnm6JK+g4CciIiIiR1VRWcVD76zivyt38bNze3PvGQp90vQiQ/x5446RDO0ewf1vf8/bS7Z7uqRWT8FPREREROpVXlnFA2+v5KNVu/jFeX24Z0IPT5ckbUi7QD9m3TqC03vG8Mh/1jDju3RPl9SqKfiJiIiIyA+UV1Zx/1vf8/Ga3fzq/NO4c3ySp0uSNijI38WLNyZzwYBOPPbxBv7xxWastZ4uq1XS2rsiIiIicpiyiirue2sFn63by68vOI3bxyV6uiRpw/x9fXhm8hBCAlw889UWCkrK+c0FffHx0YqyJ0LBT0RERERqlFVU8eM3V/DF+r387qK+3DImwdMlieDyMfzlioGEBfrx0vwMCksqeOKKgbgU/hpMwU9EREREACitqOSe11fw1cZ9/OGSftyYEu/pkkRqGGP49QWnERboy1NfbuFQWQVPTRqCv69mrzWEgp+IiIiIUFJeyd2vL+ebTdn88dL+3DAqztMlifyAMYYHz+pFWKAff/zfeg6VLmPq9cMI8nd5urQWT/FYREREpI0rKa/kztec0PenywYo9EmLd9vYBJ68YiDfbcnmpplLOFhS7umSWjwFPxEREZE2rKS8kjteXca8Ldk8cfkArh3Z3dMliTTI1cO78a/JQ/l+x36ufXERuYWlni6pRVPwExEREWmjissquX3WMuan5fCXywdyzQiFPmldLhjYiek3JrNlbyFXT0tlT36Jp0tqsRT8RERERNqg4rJKbpu1lAVbc/jrlYO4eng3T5ckclIm9o7l1VtHsPdgKVdOXci23EOeLqlFUvATERERaWOKyiq45ZUlLErP5e9XDeLKYV09XZLIKRmZGMWbd4zkUGkFV01NZdOeAk+X1OIo+ImIiIi0IYdKK7j55aUsycjjn5MGc/lQhT7xDgO7tuedO1MAmDQ9lVU7Dni2oBZGwU9ERESkjSgsreDml5ewLDOPp64ZwiWDu3i6JJFG1bNDGO/eNZp2gX5c++IiUrfmerqkFkPBT0RERKQNKCgp5+aZS1ix/QDPTB7CxYM6e7okkSbRPSqYf9+VQuf2Qdz88hK+3rjX0yW1CAp+IiIiIl7uYEk5N81cwsodB/jX5CFcOFChT7xbh3aBzL4zhd4dw5jy6nI+WrXL0yV5nIKfiIiIiBc7WFLOjS8tYXVWPs9eO5TzB3TydEkizSIyxJ83bh/J0LgI7n/7e95ast3TJXmUgp+IiIiIl8ovLueGGYtZtyuf568byrn9O3q6JJFmFRbox6u3jmBCrxh+8Z81vDgv3dMleYyCn4iIiIgXyi8q54aXFrN+90FeuG4Y5/RT6JO2KdDPxbQbkrlgYCcen7OBf3y+CWutp8tqdk0a/Iwx5xpjNhlj0owxj9Rz/mFjzHpjzGpjzFfGmDj38YnGmJV1HiXGmEubslYRERERb3GgqIzrXlrExt0FTLthGGf17eDpkkQ8yt/Xh2euGcI1w7vxzNdp/P6j9VRVta3w59tUFzbGuIDngLOBLGCpMeZDa+36Os2+B5KttUXGmLuBJ4FJ1tpvgMHu60QCacDnTVWriIiIiLfYf6iM62YsJi27kGk3DmNi71hPlyTSIrh8DH++fAChAb7MmJ9BYWkFT1w+AF9X2xgE2WTBDxgBpFlr0wGMMW8DlwA1wc8d8KotAq6v5zpXAp9Ya4uasFYRERGRVi/vUBnXvriI9JxDvHhjMuN7xXi6JJEWxRjDry44jbBAP/755WYOlVbw1DWDCfB1ebq0JteU8bYLsKPO8yz3saO5DfiknuPXAG/V9wJjzBRjzDJjzLLs7OyTLlRERESktcstLOXaFxeRkXOIl25S6BM5GmMMD5zVk99e2JdP1u7h9lnLKCqr8HRZTa5F9GsaY64HkoG/HnG8EzAA+Ky+11lrp1trk621yTEx+sdNRERE2qbsglImv7iIzNxDzLx5OON66vsikeO5dWwCT145kAVpOdz40hIOlpR7uqQm1ZTBbyfQrc7zru5jhzHGnAX8CrjYWlt6xOmrgfettd79uyAiIiJykvYVlDD5xUVszyti5s3DGdMj2tMlibQaVyd349lrh7Iq6wCTpy8it/DIOOI9mjL4LQV6GmMSjDH+OEM2P6zbwBgzBJiGE/r21XONyRxlmKeIiIhIW7fvYAmTpy9i5/5iXrllBKOTFPpETtT5Azrx4o3JbM0u5OppqezOL/Z0SU2iyYKftbYCuBdnmOYG4B1r7TpjzB+MMRe7m/0VCAX+7d62oSYYGmPicXoM5zZVjSIiIiKt1d6DJVwzfRG780uYdesIRiVGebokkVZrQu9YXr11JPsOlnLlC6lk5hzydEmNznjL5oXJycl22bJlni5DREREpMntyXeGd+476IS+5PhIT5ck4hXW7sznxplLcPkYXr9tJL07hnm6pBNijFlurU2u71yLWNxFRERERBpm14FiJk1PJbuglFdvU+gTaUz9u4Tzzp2j8DFw9bRUVu444OmSGo2Cn4iIiEgrsfNAMddMX0ReYRmv3jaCYXEKfSKNrUdsGO/eNZrwID+ue3ERC7fmeLqkRqHgJyIiItIKZO0v4prpqewvKuO120cytHuEp0sS8VrdIoP5910pdIkI4uaXl/Ll+r2eLumUKfiJiIiItHA78oqYNG0R+UXlvHH7SAZ3a+/pkkS8Xod2gcyekkKfjmHc9fpy/rvyBzvTtSoKfiIiIiIt2PbcIq6ZvojC0grevGMUA7u293RJIm1GRIg/b9w+kmFxETw4eyVvLt7u6ZJOmoKfiIiISAu1LfcQ10xP5VBZBW/cPpL+XcI9XZJImxMW6MesW0cwsXcsv3x/DdPmbvV0SSdFwU9ERESkBcrIOcSkaYsoLq/kzdtHKfSJeFCgn4tpNwzjokGd+fMnG/nbZ5tobdvi+Xq6ABERERE53NbsQq59cRHllZY37xjFaZ3aebokkTbPz+XDU5MGExrg4tlv0jjztFiGtKJFlo4a/IwxDx9xyAI5wHxrbUaTViUiIiLSRqXtc0JfZZXlrTtGtboNpEW8mcvH8KfLBnD50K6tKvTBsYd6hh3xaAckA58YY65phtpERERE2pS0fQVcM30RVdby1hSFPpGWyBjD8PjWt4fmUXv8rLW/r++4MSYS+BJ4u6mKEhEREWlrNu8t4NoXF2GM4a07RtEjVqFPRBrPCS/uYq3NA0wT1CIiIiLSJm3cc5DJ0xfhYwxvT1HoE5HGd8LBzxgzEdjfBLWIiIiItDkbdh/k2hcX4+tyQl9STKinSxIRL3SsxV3W4CzoUlcksAu4sSmLEhEREWkL1u3K5/oZiwnwdfHWlFEkRId4uiQR8VLH2s7hwiOeWyDXWnuoCesRERERaRPW7szn+pcWE+znhL64KIU+EWk6Rx3qaa3dZq3dBiQCFwEXAyOaqzCRU3GotIL3lmdxsKTc06WIiIj8wJqsfK59cREh/r68PSVFoU9Emtyxhnp2Af4DlADL3YevMsb8BbjMWruzGeoTOSFVVZb/rtrJE59sZO/BUs5cE8uMm5IxRusRiYhIy7BqxwFueGkxYYF+vD1lFN0igz1dkoi0Acca6vks8IK19pW6B40xNwLPA5c0YV0iJ+z77fv5/UfrWbnjAAO7hnPhwM68ND+DF79LZ8rpSZ4uT0REhO+37+fGmUtoH+zHW3eMomuEQp+INI9jBb++1trLjjxorX3VGPOrJqxJ5ITsyS/hL59u5P3vdxIbFsDfrhrE5UO6YAzszi/mL59uYlhcBMPiWt9GmyIi4j2Wb9vPzTOXEBHiz1tTRtGlfZCnSxKRNuRYwa/e+X/GGB/A1TTliDRcSXklL85L5/lvt1JpLT+emMQ9E3oQElD7x/qJKwaybtd87n3zez6+fxyRIf4erFhERNqq5dvyuGnmUqJC/Xl7yig6hSv0iUjzOtY+fv8zxrxojKmZbez+eiowp8krEzkKay0fr97NmX+fy9+/2MyE3jF89fB4fvqjPoeFPoB2gX48d+1QcgvLePidlVRVHblDiYiISNNampnHjS8tISYsgNlTUhT6RMQjjhX8fgbkA9uMMcuNMSuATOAg8JNmqE3kB9buzGfStEX8+M0VhAX68tYdo3jh+mHHnBjfv0s4v7moL99uymbqvK3NWK2IiLR1i9NzuWnmEjqEB/L2lFF0DA/0dEki0kYddaintbYc+Ikx5jdAD/fhrdbaomapTKSO7IJS/vbZJt5ZvoOIYH/+dNkAJg3vhsunYat1Xj+yO4vTc/n755tJjotkRILm+4mISNNK3ZrLra8spXP7QN66YxSx7RT6RMRzjrWdw+X1HO5ZvSy+tfY/TVWUSLXSikpeWZDJv75Oo6S8ktvGJHDfmT0JD/I7oesYY/jz5QNYt+sg9721gjn3jyMqNKCJqhYRkbZuYVoOt85aSreIYN64YySxYQp9IuJZx1rc5aJjnLM4e/yJNAlrLV9u2MfjH68nM7eIM/vE8qsLTiMxJvSkrxnmnu936fMLeOidVbxy83B8GthjKCIi0lAL0nK4bdZSukcG8+Ydo4jWDxpFpAU41lDPW5qzEJFqm/YU8Mf/rWd+Wg49YkOZdesIxveKaZRr9+3cjkcv6scv31/D89+mce8ZPRvluiIiIgDzNmdzx6vLSIgO4Y3bR2p0iYi0GMfq8RNpVvsPlfHPLzfzxuLthPi7+N1Ffbl+VBx+rmOtQXTiJo/oxuKMXP7xxWaGxUWSkhTVqNcXEZG26dtN+5jy2nISo0N4845R2kJIRFoUBT/xuPLKKl5ftI2nvtxCQUk514+K46GzehHRRP9hGmP402UDWLMzn/vf/p45948jJkw/kRURkZP3zcZ93PnacnrEhvLG7SOb7P8wEZGT1bhdKSInaO7mbM57+jt+/9F6BnQJ55MHTucPl/Rv8v8wQwJ8ef66oRSUlPPQ7JVUan8/ERE5SV9t2Mudry2nV8dQ3rxDoU9EWqYG9fgZY0YD8XXbW2tfbaKapA1Izy7k8Y838NXGfcRHBfPijcmcdVos1avGNoc+Hdvxh4v787P3VvPs12k8cJbm+4mIyIn5Yv1e7nljOad1asdrt44kPPjEVp0WEWkuxw1+xpjXgCRgJVDpPmwBBT85YfnF5fzrqy3MSs0kwNfFL87rw81j4gnwdXmknquSu7IoI5envtrM8PgIRveI9kgdIiLS+ny2bg/3vrmCvp3a8eptI094qyERkebUkB6/ZKCvtVZj4eSkVVZZZi/dwd8/30ReURmTkrvxf+f09vjcOmMMj13an9VZ+dz/9krmPDBWey2JiMhxfbp2N/e++T39u4Tz6m0jaBeo0CciLVtD5vitBTo2dSHivVK35nLhv+bzy/fXkBQTykf3juWJKwZ6PPRVC/Z35vsdKq3ggbc0309ERI5tzprd/PjN7xnYNZzXFPpEpJVoSI9fNLDeGLMEKK0+aK29uMmqEq+wI6+IP83ZwCdr99ClfRDPXjuECwZ0atZ5fA3Vq0MYf7y0Pz/59yqe/nIzD5/T29MliYhIC/TRql08OHslQ/6/vfuOrrq+/zj+fGex9yY3QBiyR5ILCoh7I6BAUBwtjmrdrV12/Gy1tlptneDerYoJoqIoDsSBAnITwl5h5oa9AgEy7+f3R64ttYJRk3yTe1+Pczjkfu/Nva98Dyfc1/1+RlJznr9yCI3raYF0EakbKvPb6k/VHUIiS2FxGY/OyeXpuRuINeMXZx7HT07qSv14b+bxVdb4NB8L1u/mkTm5DE5uyYgeVbNpvIiIRIY3c/L5+as5+Du35NkrBqv0iUid8q2/sZxzn9REEKn7QiHH9EX53DtrFTsOFHNhSiK/OacX7ZvVnTlzd47px+LgPn42NYd3bhlBu6Z1J7uIiFSfNxblc2tGDv4uLXlu0mAaqfSJSB3zrXP8zOwEM1toZoVmVmJm5Wa2vzJPbmbnmNlqM8s1s9u+4f5bzWyFmS0xs9lm1vmI+zqZ2ftmtjL8mC7f6SeTGpW1aS8XPvo5v8xcTMfmDZh+/TAeuGhQnSp9AA0SYnn00lQOl5Zz0yuLKCsPeR1JREQ8Nj07yK0ZORyf3Irnr1DpE5G6qTKLu0wGJgJrgQbA1cCUb/smM4sNP+5coA8w0cz6fO1hiwC/c24AMA2494j7XgTuc871BoYAOyqRVWrY1oLD3DJ1EeMe+4Jt+4t44KKBTL9uGKmdWngd7Xvr3rYJf7mwH19u2MMDH67xOo6IiHgoM5DHLzIXc0LXVjw7aTANE1T6RKRuqtRvL+dcrpnFOufKgefMbBHw22/5tiFArnNuPYCZTQXGACuOeN45Rzx+PnBZ+LF9gDjn3AfhxxVW8ueRGnK4pJwnP13PY5/kEnJw02nd+enJ3SLmU9ALU3wsWL+HKXPWMbhLS07p2dbrSCIiUsMyFubxm+lLGN6tNU/9yE+DhNo9V11E5Fgq8y79kJklADlmdi+wlcpdKUwE8o64HQSOP8bjrwLeDX99HLDPzKYDycCHwG3h4vlvZnYNcA1Ap06dKhFJfijnHG8t2co976xkS0ERI/t34LZze5HUsqHX0arcn0b3JSdvHz9/tWK+X4dmDbyOJCIiNWTql5u5bfpSRvSoKH21fYEyEZFvU5kCd3n4cTcCB4EkYFxVhjCzy6jYKP6+8KE4YATwS2Aw0BWY9PXvc8496ZzzO+f8bdpoBcbqtjRYQPrj87j5lUU0b5jAq9ecwJRLUyOy9AHUj6+Y71dSFuKmlzXfT0QkWry0YBO3TV/KKT3bqPSJSMT41uLnnNsEGNDBOXeHc+5W51xuJZ47n4qS+BVf+Nh/MbMzgN8Do51zX+0TGARynHPrnXNlwBtAaiVeU6rBjgNF/CpzMaOnzGXj7oPcM7Y/b910Isd3beV1tGrXtU1j/jq2P4FNe/n7+5rvJyIS6f45byO/f30Zp/VqyxOXp6n0iUjE+NahnmY2Cvg7kAAkm9kg4M5KbOC+EOhhZslUFL6LgUu+9twpwBPAOc65HV/73uZm1sY5txM4DQhU7keSqlJcVs6zczcy+aO1lJSH+MmIrtx4Wnea1o/3OlqNGjMokQUb9vD4J+sYktyC03q18zqSiIhUgxe+2MgfZyznjN5tmXJpKvXiVPpEJHJUdgP3IcDHAM65nHCZOybnXJmZ3Qi8B8QCzzrnlpvZnUDAOTeDiqGdjYFMMwPY7Jwb7ZwrN7NfArOt4o4s4Knv/NPJ9+Kc4/0V2/nLzJVs3nOIM3q34/cje5PcupHX0Txz+/l9yNm8j1szFjPz5hEkNtd8PxGRSPLc5xu4460VnNmnHVMuSSUhrjKzYURE6g5zzh37AWbznXMnmNki51xK+NiS8BYMtYbf73eBgC4K/lCrtu3nzrdW8MW63fRo25jbR/VhRA/NnwTYuOsg5z8yl+PaNebVa4cSH6s3BSIikeDpz9Zz18yVnN23HY9MVOkTkbrLzLKcc/5vuq8yv9mWm9klQKyZ9TCzR4AvqjSheG7PwRL+8MZSznvoM1Zs3c+dY/ry7i0jVPqO0KV1I/42bgDZm/dx76xVXscREZEq8OSn67hr5krO7deeybrSJyIRrDJDPW+iYvGVYuAVKoZu/rk6Q0nNKS0P8eK8TTz04RoOlpTzo6Fd+NkZPWjeMMHraLXSyAEdWLChM099toEhya04s4/m+4mI1FWPf7KOe95dxcj+HXjw4kEaySEiEe1bh3rWFRrq+d3NWb2DP7+9gvU7DzKiR2tuP78PPdo18TpWrVdcVs64x74gb89hZt58Ir4WkbmdhYhIJJsyJ5f73lvNqIEdeWDCQOJU+kQkAhxrqOdRr/iZ2YxjPWklVvWUWip3RyF3zVzBx6t3kty6Ec/82M9pvdoSXmBHvkW9uFimXJLK+Q/P5YaXF5F57VANDRIRqUMemb2Wf3ywhjGDOvKPdJU+EYkOxxrqORTIo2J45wIq9vKTOqzgUCkPzV7Li/M20iA+lj+M7M2PhnZRafkeOrdqxL3jB3DdS9nc8+4qbh/Vx+tIIiJSCQ99uJYHPlzD2JRE7ksfSGyM3t6ISHQ4VvFrD5wJTKRi/72ZwCvOueU1EUyqTll5iKkL8/jH+6vZd7iUiwd34hdnHUfrxvW8jlanndu/A5OGdeHZzzcwJLkl5/Rr73UkERE5CuccD364lodmr2Vcqo97xw9Q6RORqHLU4uecKwdmAbPMrB4VBfBjM7vDOTe5pgLKD/NF7i7ufHsFq7Yd4Pjkltw+qg99OzbzOlbE+N15vVm0eS+/mraYPh2a0qmV5vuJiNQ2zjnu/2ANj3yUS3qaj3vGqfSJSPQ55hg/M6tnZmOBfwE3AA8Dr9dEMPlhNu0+yDUvBrjk6QUUFpfx2KWpTL3mBJW+KpYQF8PkS1Ix4MZXsikuK/c6koiIHME5x33vreaRj3K5eHASf1PpE5EodazFXV4E+gHvAHc455bVWCr53gqLy5j8US7Pzt1AXKzxq7N7ctWJydSPj/U6WsRKatmQ+9IHcu0/s7j7nVX8aXRfryOJiAgVpe9vs1bz+CfrmDikE3+5oB8xKn0iEqWONcfvMuAgcAtw8xErPhrgnHNNqzmbfAehkGNadpB7Z61mV2Ex41J9/PqcnrRrWt/raFHh7L7tuerEZJ6ZWzHf77z+HbyOJCIS1Zxz3P3uKp78dD2XndCJO0er9IlIdDvWHD8t9VhHBDbu4Y63VrA0v4DUTs15+sd+BiU19zpW1PnNOb3I2rSX30xbQt+OTencqpHXkUREopJzjrtmruSZuRv40dDO3DG6r7YsEpGop3JXh+XvO8xNryxi/OPz2FVYzEMXD+K164ap9HmkYr5fCjExxvUvZVNUqvl+IiI1zTnHnW+v4Jm5G5g0rItKn4hImIpfHXSopIz7P1jD6f/4mPeXb+Pm03sw+xcnM2ZQov5z85ivRUPunzCQ5Vv2c9fMFV7HERGJKs45/jRjOc99vpErhyfzx1F99P+iiEjYseb4SS3jnGPG4i3c8+4qthYUcf6ADtx2bi98LbSFQG1yeu92XHtSV574dD3HJ7di1MCOXkcSEYl4oZDjjzOW88/5m/jJiGR+d15vlT4RkSOo+NURi/P2ccdby8nevI9+iU15eGIKg7u09DqWHMUvz+5JYNNefjt9Kf0Sm5HcWvP9pGas3LqfuBijR7smXkcRqTGhkOMPby7j5QWbufbkrtx2Ti+VPhGRr1Hxq+W27y/i3lmreS07SOvG9bh33ADGp/m0MlktFx8bwyMTUxj58Gdc/1I2r18/TFtqSLVyzvHkp+u5+91VAAxMas5F/iTOH9iBpvXjPU4nUj1KykJ8tGo7/5q/mbm5u7julG78+uyeKn0iIt/AnHNeZ6gSfr/fBQIBr2NUmaLScp6Zu4Epc3IpK3dceWIyN5zajSZ6A1enzFm9gyueW8jEIZ24e2x/r+NIhCorD/HHGct5acFmRvbvQEqn5mQE8lizvZD68TGc168D6f4kjk9uqQ+NJCIs31JAZiDImzn57D1UStsm9bjmpK5cdWKySp+IRDUzy3LO+b/pPl3xq2Wcc8xato2/vruSvD2HOatPO34/sre2BqijTu3ZlutO6cZjH6/jhK4tGTMo0etIEmEKi8u48eVsPl69k5+eXHG1IybGuOrEZJYEC8gI5DEjZwvTF+XTqWVD0tN8jEvz0bF5A6+ji3wnew6W8GZOPpmBICu27ichNoYz+7RjvN/HiO6tiYvVenUiIseiK361yIot+7nz7eXMX7+Hnu2acPuoPgzv3trrWPIDlZWHuOSpBSzbUsCMG0+ke9vGXkeSCLG14DBXPh9gzfYD3HVBPyYO6fSNjztcUs57y7eREcjji3W7MYMRPdowwe/jzD7tqBenYchSO5WVh/hkzU4yA0Fmr9pOabmjf2Iz0v0+Rg3oSItGCV5HFBGpVY51xU/FrxbYXVjM399fw6sLN9OsQTy3ntWTiYOT9OllBNlWUMR5D39Gm8b1eOOG4TRI0Btt+WGWbyngyucXcrC4nCmXpnLycW0q9X15ew6RmRVkWiCPLQVFNG8YzwWDEkn3++jbsVk1pxapnLXbD5CZFWR6dj67Cotp1SiBC1Iq/p32at/U63giIrWWil8tVVIW4sV5G3lo9loOl5Rz+dDO/Oz042jWUPP4ItEna3Yy6bkvmZCWxN/GD/A6jtRhc1bt4MaXs2naIJ5nJw2md4fv/ka4POT4PHcXGYE83l++nZLyEH07NuWiwUmMHtiR5g11JUVqVsGhUmYs2cK0rCCL8/YRF2Oc2qst6Wk+Tu3Vlnh9GCoi8q1U/GoZ5xwfrdrBX2auZP2ug5zSsw1/GNlHQwCjwN/fW83kObncP2EgY1N9XseROuif8zfxxzeX0btDU56dNJh2Tev/4Ofcd6iEN3O2kBHIY/mW/STExXB23/ZM8PsY3q21FoSRalMecszN3cW0rCDvLd9GSVmIXu2bMD7NxwUpibRuXM/riCIidYqKXy2Su+MAd769kk/X7KRrm0b838g+nNqrrdexpIaUlYe49OkFLAkWMOPG4dprTSotFHLcM2sVT366ntN6teWRiSk0qlf163Mtyy9gWlaQ1xflU3C4lMTmDRiX5iM9zUdSy4ZV/noSnTbsOsi0rDymZ+eztaCIZg3iuWBQR8anJdEvsalW5hQR+Z5U/GqBfYdKePDDtfxz/iYaJsTyszOO40dDO2voShTasb9ivl+Lhgm8eeNwGiZocV05tqLScn7+ag7vLtvGj4Z25vbz+1T7HOCi0nI+WLGdjEAec3N34RwM69aKCf4kzunXXvtSyndWWFzGzCVbyAwECWzaS4zByce1YXxaEmf0aatFhkREqoCKn4fKykO8/OVm7v9gDfsPlzJxSCduPfM4Wmn4SlT7PHcXlz2zgLEpPv4xYaDXcaQW21VYzNUvBFgc3Mfvz+vtyT5l+fsO81pWkMysPPL2HKZJ/ThGD+zIBH8SA3zNdHVGjioUcszfsJtpgSDvLtvG4dJyurZpRHpaEmNTE6tkqLKIiPyHip9HluUXcGtGDmu2FzK0aytuH9Xney3CIJHpgQ/W8NDstdw3fgDp/iSv40gtlLujkCue/5KdB4p58KIUzunX3tM8X72JzwwEeWfpVorD87HS/UlcMKijPtCSf8vbc4hpWUFeyw4S3HuYJvXiOH9gR9L9PlKSmuvDAhGRaqLi55HNuw9x5QsL+eVZPTm7bzv9Ryf/pTzkuPyZBWRv3subN5xIz/aa7yf/MX/9bq55MUBCXAxP/3gwg5Kaex3pv+wvKuWtxVvICFSswBgfa5zRux0T/EmM6KHNtKPRoZIyZi3bRmYgyLz1FftFDu/WmnS/j7P6tNc2NiIiNUDFz0POORU+OaodB4o476G5NGsQx4wbT6yWxTqk7nl9UZBfT1tC51aNeG7S4Fq/qMrqbQfIDOQxfVE+ew6W0K5pPcal+kj3J5HcupHX8aQaOefI2rSXzECQmUu3UlhcRudWDRmf6mNsmo/E5g28jigiElVU/ERqsS/W7eKypxcwZlAi908YqA8Kophzjodmr+XBD9cytGsrHr8srU7t61lSFuKjVdvJCAT5ePUOQg6GdGlJut/HyAEdtJBRBNlacJjp2flMywqyYddBGibEcl7/DqSn+RiS3FK/x0REPKLiJ1LLPTx7Lfd/sIZ7xvbn4iGdvI4jHigpC3Hb9CVMz85nbGoi94wdQEJc3R0uuX1/Ea9lB8kMVBSDRgmxjBrYkXR/EqmdNMerLioqLef9FduZlhVk7tqdFcU+uSXpaT7O699BIxZERGoBFT+RWq485Jj03Jd8uWEPb9wwXIsARZmCQ6X89F9ZzFu/m5+fcRw3n949YoqRc47Apr1kLMxj5tKtHCopp1ubRkzwJ3FhaiJtm2hVx9rMOcfiYAHTsvKYkbOF/UVlFXs7piYyLs1H51YayisiUpuo+InUAbsKiznvoc9oXC+OGTedSGN9eh4V8vYcYtJzX7J5zyHuHT+AC1N8XkeqNoXFZbyzZCsZgTwCm/YSG2Oc2rMtE/w+Tu3VVvua1iI7DhTxxqJ8MgNB1u4opF5cDOf2a8/4tCSGdWtFTExkfDAhIhJpVPxE6oj563dzyVPzOX9ARx66eFDEXPWRb7Zo815+8mKA0nLHE5encULXVl5HqjHrdhaSGahY7n/ngWJaN05gbKqPCX4f3dtqhVsvfDVHMzMQ5OM1OykPOVI7NWd8WhLnD+xA0/p1Z76piEi0UvETqUOmzMnlvvdW85cL+3Hp8Z29jiPVZNayrdwyNYd2Tevz7KTBdG/b2OtInigrD/Hx6p1kBPL4aNUOykKOlE7NmeBP4vwBHWiislHtlm8pIDMQ5M2cfPYeKqVtk3qMTfUxPs0Xtf8uRUTqKhU/kTokFHJc8fxC5q3fzfTrhtEvsZnXkaQKOed4Zu4G/vLOSgYlNefpH/m18XnYzgPFvLEon1cDeeTuKKRBfMVKkRP8Wimyqu05WMIbiypW5VyxdT8JsTGc2acd4/0+RnTXPowiInWVip9IHbPnYAnnPfQZ9eNjeOumE3XVI0KUlYe4460V/HP+Js7t154HLhpE/Xhtav11zjly8vaREQjy1uItFBaX0aVVQ9L9SYxL9dG+mRaE+T7KykN8smYnmYEgs1dtp7Tc0T+xGel+H6MGdKRFowSvI4qIyA+k4idSBy3cuIeLn5zPOX3bM/mSFF3tqOMOFpdx48vZzFm9k2tP6spvzumlBTIq4VBJGbOWbSMjkMf89XuIMTjpuDZM8Cdxeu+21ItTcf42a7cfIDMryPTsfHYVFtOqUQIXpiQy3u+jV3utICwiEkk8K35mdg7wEBALPO2cu+dr998KXA2UATuBK51zm8L3lQNLww/d7JwbfazXUvGTSPTYx+v426xV/HlMXy4f2sXrOPI9bSso4srnF7J6+wHuGN2Xy07Q3M3vY9Pug0zLCjItK8jWgiJaNIzngpREJviTtAXK1xQcKmXGki1MywqyOG8fcTHGab3aMj5NK6iKiEQyT4qfmcUCa4AzgSCwEJjonFtxxGNOBRY45w6Z2XXAKc65i8L3FTrnKj2rXMVPIlEo5LjqhYV8nrub164bRn+f5vvVNSu27OeqFxay/3Apky9N5dSebb2OVOeVhxyfra0Ysvj+im3/HrI4we9j9KBEmjWIzqHR5SHH3NxdTMsK8t7ybZSUhejVvgnj03xckJJIa80lFRGJeF4Vv6HAn5xzZ4dv/xbAOXf3UR6fAkx2zg0P31bxEwH2Hixh5MOfERcbw9s3n6gl1euQj1fv4IaXsmlSP55nJw2mT0ddlapqew6W8GZOPq8uzGPVtgPUi4vhnH7tmeBPYmjX6NhvbsOug0zLymN6dj5bC4po3jCeMQM7ku5Pom/HphomLiISRbwqfuOBc5xzV4dvXw4c75y78SiPnwxsc87dFb5dBuRQMQz0HufcG9/wPdcA1wB06tQpbdOmTdXwk4h4L2vTXi56Yh5n9G7HY5el6o1cHfDygs3835vLOK5dE56d5KdDswZeR4pozjmWb9lPRiCPNxbls7+ojMTmDUj3V2xL4GvR0OuIVaqwuIyZS7aQGQgS2LSXGIOTj2tDuuY+iohEtVpf/MzsMuBG4GTnXHH4WKJzLt/MugIfAac759Yd7fV0xU8i3VOfrucv76zkj6P6cMXwZK/jyFGEQo6/vbeKJz5Zzyk92zD5klQa14vzOlZUKSot5/0V28lYmMfn63YBMLxba9L9Ps7u277OrqQaCjnmb9jNtECQd5dt43BpOV3bNCI9LYmxqYm0a6rVTkVEot2xil91vhvJB5KOuO0LH/svZnYG8HuOKH0Azrn88N/rzexjIAU4avETiXRXj0hmwYbd/PWdlaR0asGgpOZeR5KvKSot5xcZi5m5dCuXHt+JO0b31X5oHqgfH8vogR0ZPbAjeXsO8Vp2kMxAkFum5tC0fhxjBlUsCNMvsW4Mg8zbc4hpWUFeyw4S3HuYJvXiuCAlkXS/j5Sk5nXiZxAREe9V5xW/OCoWdzmdisK3ELjEObf8iMekANOouDK49ojjLYBDzrliM2sNzAPGHLkwzNfpip9Eg32HShj58FzMYOZNI2jWUPP9aovdhcX85MUA2Zv38bvzevGTEV31hrwWCYUc89bvJiOQx7vL/rPwyQR/EhemJNa6Pey+2sYiMxBk3vrdmEXGVUsREaleXm7ncB7wIBXbOTzrnPuLmd0JBJxzM8zsQ6A/sDX8LZudc6PNbBjwBBACYoAHnXPPHOu1VPwkWizavJcJT8zjlJ5tefLyNJWLWmDdzkKueG4h2/cX8eBFgzi3fwevI8kxfLXVQWYgjyXBAhJiYzizTzvS/T5G9GhDrEcLwjjnyNq0l8xAkJlLt1JYXEbnVg0Zn+pjbJqPxOaaJyoiIsemDdxFIswzczfw57dX8IeRvbl6RFev40S1Bet3c80/s4iLMZ76sZ/UTi28jiTfwcqt+8kMBHl9UZC9h0pp37Q+49N8pPt9dG7VqEYybC04zPTsfKZlBdmw6yANE2IZ2b8D49N8DEluqQ93RESk0lT8RCKMc46f/iuL2St3kPHToSobHnljUT6/nrYEX8sGPD9pCJ1aRdbKkdGkuKycj1bu4NVAHp+u2UnIwfHJLZngT+Lc/u1pmFC1U+K/WoBmWlaQuWv/83rj03yc178DjbQgkIiIfA8qfiIRqOBwKSMf/gznYObNJ9K8Ye2aoxTJnHNM/iiXf3ywhuOTW/LE5Wk6/xHkqytwGYE8Nu0+RON6cYwa2JEJfh+DfsBiKs45FgcLmJaVx4ycLf/ecmJcaiLj0mruCqOIiEQuFT+RCLU4bx/jH/+Ck3q04akf+aNis2qvlZSF+N3rS5mWFWRsSiJ3j+uvPdMilHOOLzfsISMQ5J2lWzlcWk6Pto2Z4E/igpRE2jSpV6nn2XGgiDcW5ZMZCLJ2RyH14mI4t1970qNok3kREakZKn4iEez5zzfwp7dW8LvzenHNSd28jhPRCg6Xct2/svhi3W5uOb0HPzujh+ZfRYkDRaXMXLKVjEAe2Zv3ERdjnNarLRP8SZzSs83/bNtRUhbio1XbyQwE+XjNTspDjtROzUn3JzFyQAea1teKvCIiUvVU/EQimHOOG17O5r3l28m49gTSOrf0OlJEyttziCufX8jG3Qe5e+wAxqf5vI4kHsndcYDMQMW+ersKS2jTpB5jUxNJT0uiuKyczECQN3Py2XuolLZN6jE21cf4NB/d2zb2OrqIiEQ4FT+RCLe/qJRRj8ylpCzEzJtH0LKW7UlW1y3O28dVLwQoLivnicvTGNattdeRpBYoLQ8xZ9UOMgJB5qzeQXmo4v/Tr7aHGO/3MaJ76/+5GigiIlJdVPxEosCy/ALGPvoFw7q34tkfD9a8oSry/vJt3Dx1Ea0b1+P5KwbTvW0TryNJLbTjQBEzcrZQLz6WUQM6aLEfERHxxLGKnz6GFIkQ/RKb8X+j+vDx6p08/uk6r+PUec45npm7gWv/lUXP9k15/frhKn1yVG2b1OfqEV25/ITOKn0iIlIraaMgkQhy2fGdWLB+N/94fw3+zi0Zkqz5ft9Hechx51vLeWHeJs7u244HL0qhQYJW7hQREZG6S1f8RCKImXH32P50atmQm17JZndhsdeR6pyDxWVc82KAF+Zt4icjknn00jSVPhEREanzVPxEIkyT+vFMviSFvYdK+XnGYkKhyJjHWxO27y/ioifnMWf1Dv48pi+/H9mHWM2VFBERkQig4icSgfp2bMafRvXl0zU7efTjXK/j1Amrtu3nwimfs37nQZ7+sZ/Lh3bxOpKIiIhIlVHxE4lQE4ckMWZQR+7/YA3z1u32Ok6t9umanYx/bB5lIUfGtUM5rVc7ryOJiIiIVCkVP5EIZWb89cL+dGndiJunLmLnAc33+yZTv9zMFc8vxNeiAW/cMJx+ic28jiQiIiJS5VT8RCJYo3pxTLkklf2HS/n5qzn/3mBaIBRy3DtrFbdNX8rw7q3J/OlQOjZv4HUsERERkWqh4icS4Xp3aMqdY/oyN3cXkz/SfD+AotJybp66iEc/XsfEIZ145sd+mtSP9zqWiIiISLXRPn4iUWCCP4kF6/fw4Ow1DO7SgmHdW3sdyTN7DpZwzYsBApv2ctu5vbj2pK6YaeVOERERiWy64icSBcyMuy7sR7c2jbl5ag47DhR5HckTG3YdZOyjn7Mkv4Apl6Ty05O7qfSJiIhIVFDxE4kSDRPiePTSVA4Wl3HLK9E332/hxj1c+Ojn7C8q45WfHM/IAR28jiQiIiJSY1T8RKLIce2a8OcL+jFv/W4e+nCN13FqzJs5+Vz61AJaNkzg9euHkda5pdeRRERERGqUip9IlBmf5mN8mo9H5uTy2dqdXsepVs45pszJ5ZapOQxKas5r1w2jc6tGXscSERERqXEqfiJR6M9j+tGjbWN+NjWH7fsjc75faXmI215byn3vrWbMoI788+ohtGiU4HUsEREREU+o+IlEoQYJsTx6aSqHS8u56ZVFlJWHvI5UpfYXlXLFcwt5NZDHzad158GLBlEvLtbrWCIiIiKeUfETiVLd2zbhLxf248sNe3gggub7BfceYvxjXzB//W7uHT+AW8/qqZU7RUREJOppHz+RKHZhio8F6/cwZc46BndpySk923od6QdZGizgyhcWUlRazgtXDmF4FO9XKCIiInIkXfETiXJ/Gt2XXu2b8PNXc9hacNjrON/bByu2M+GJeSTExjD9umEqfSIiIiJHUPETiXL142OZcmkqJWUhbnq5bs73e+7zDVzzzwA92jXm9RuG0aNdE68jiYiIiNQqKn4iQrc2jfnr2P4ENu3l7+/Xnfl+5SHHHW8t5463VnBG73ZMveYE2jap73UsERERkVpHc/xEBIAxgxJZsGEPj3+yjiHJLTitVzuvIx3ToZIybpmawwcrtnPl8GR+P7I3sTFaxEVERETkm+iKn4j82+3n96FPh6bcmrGY/H21d77fjgNFXPzkfGav3M4do/ty+6g+Kn0iIiIix6DiJyL/9tV8v7Jyx00vZ1NaC+f7rdl+gAunfMHa7YU8ebmfHw/r4nUkERERkVpPxU9E/kty60bcM64/2Zv3ce+sVV7H+S9z1+5i3KNfUFIeIuPaoZzRp3YPRxURERGpLVT8ROR/nD+gI5ef0JmnPtvAByu2ex0HgIyFeUx67ks6Nm/AGzcMp7+vmdeRREREROoMFT8R+UZ/OL83/RKb8svMxQT3HvIsh3OOv7+3ml+/toSh3VqRed1QEps38CyPiIiISF2k4ici36heXCxTLkklFHLc8PIiSspqfr5fcVk5t0zNYfKcXC4enMSzkwbTtH58jecQERERqetU/ETkqDq3asS94wewOG8f97xbs/P99h4s4fKnv2TG4i386uye3D22P/Gx+pUlIiIi8n1U67soMzvHzFabWa6Z3fYN999qZivMbImZzTazzl+7v6mZBc1scnXmFJGjO7d/ByYN68Kzn29g1rJtNfKaG3cdZOxjX5AT3McjE1O44dTumGm7BhEREZHvq9qKn5nFAlOAc4E+wEQz6/O1hy0C/M65AcA04N6v3f9n4NPqyigilfPb83ox0NeMX01bzObd1TvfL2vTHi589HP2HSrh5auPZ9TAjtX6eiIiIiLRoDqv+A0Bcp1z651zJcBUYMyRD3DOzXHOffUucj7g++o+M0sD2gHvV2NGEamEenGxTL4kFQNufCWb4rLyanmdt5dsYeJTC2jWIJ7p1w/H36VltbyOiIiISLSpzuKXCOQdcTsYPnY0VwHvAphZDPAP4JfHegEzu8bMAmYW2Llz5w+MKyLHktSyIfelD2RJsIC736na+X7OOR79OJcbX17EgMRmTL9+OMmtG1Xpa4iIiIhEs1qxUoKZXQb4gfvCh64H3nHOBY/1fc65J51zfuecv02bNtUdUyTqnd23PVedmMzzX2zknaVbq+Q5S8tD/Hb6Uu6dtZpRAzvyr6uPp2WjhCp5bhERERGpEFeNz50PJB1x2xc+9l/M7Azg98DJzrni8OGhwAgzux5oDCSYWaFz7n8WiBGRmvWbc3qRtWkvv5m2hL4dm9K51fe/MnegqJTrX8rms7W7uOHUbvzizJ7ExGgRFxEREZGqVp1X/BYCPcws2cwSgIuBGUc+wMxSgCeA0c65HV8dd85d6pzr5JzrQsVwzxdV+kRqh4S4GCZfkkJMjHH9S9kUlX6/+X5b9h0m/fF5zFu3m7+N68+vzu6l0iciIiJSTaqt+DnnyoAbgfeAlUCGc265md1pZqPDD7uPiit6mWaWY2YzjvJ0IlKL+Fo05B/pA1m+ZT93zVzxnb9/WX4BF0z5nPy9h3n+iiFcNLhTNaQUERERka+Yc87rDFXC7/e7QCDgdQyRqPLXd1by5KfreWRiSqW3XZi9cjs3vbKI5g3iee6KIfRs36SaU4qIiIhEBzPLcs75v+m+WrG4i4jUTb86uydpnVvw2+lL2bDr4Lc+/sV5G/nJiwG6tmnEGzcMV+kTERERqSEqfiLyvcXHxvDIxBTiY48936885Pjz2yu4/c3lnNarLRnXDqVt0/o1nFZEREQkeqn4icgP0rF5A+6fMIiVW/dzx1v/O9/vcEk517+UxTNzNzBpWBeeuNxPw4TqXFBYRERERL5OxU9EfrBTe7Xlpyd345UvN/Nmzn92bdl5oJiLn5zH+yu2c/v5ffjT6L7EauVOERERkRqnj91FpEr88qzjyNq0h99OX0rfjs1wznHF8wvZVVjME5elcVbf9l5HFBEREYlaWtVTRKrMtoIiznv4M5o1iGdXYTH14mJ5dpKfAb7mXkcTERERiXha1VNEakT7ZvV54KJBbNx9kA7N6vPGDcNU+kRERERqAQ31FJEqdfJxbXjn5hF0atmQRvX0K0ZERESkNtC7MhGpcr07NPU6goiIiIgcQUM9RUREREREIpyKn4iIiIiISIRT8RMREREREYlwKn4iIiIiIiIRTsVPREREREQkwqn4iYiIiIiIRDgVPxERERERkQin4iciIiIiIhLhVPxEREREREQinDnnvM5QJcxsJ7DJ6xzfoDWwy+sQUUzn3zs6997RufeOzr13dO69o3PvHZ1779TWc9/ZOdfmm+6ImOJXW5lZwDnn9zpHtNL5947OvXd07r2jc+8dnXvv6Nx7R+feO3Xx3Guop4iIiIiISIRT8RMREREREYlwKn7V70mvA0Q5nX/v6Nx7R+feOzr33tG5947OvXd07r1T58695viJiIiIiIhEOF3xExERERERiXAqfiIiIiIiIhFOxa8amdk5ZrbazHLN7Dav80QTM3vWzHaY2TKvs0QTM0syszlmtsLMlpvZLV5nihZmVt/MvjSzxeFzf4fXmaKNmcWa2SIze9vrLNHGzDaa2VIzyzGzgNd5oomZNTezaWa2ysxWmtlQrzNFAzPrGf73/tWf/Wb2M69zRQsz+3n4/9plZvaKmdX3OlNlaI5fNTGzWGANcCYQBBYCE51zKzwNFiXM7CSgEHjROdfP6zzRwsw6AB2cc9lm1gTIAi7Qv/vqZ2YGNHLOFZpZPDAXuMU5N9/jaFHDzG4F/EBT59z5XueJJma2EfA752rjZsoRzcxeAD5zzj1tZglAQ+fcPo9jRZXwe8584Hjn3Cav80Q6M0uk4v/YPs65w2aWAbzjnHve22TfTlf8qs8QINc5t945VwJMBcZ4nClqOOc+BfZ4nSPaOOe2Oueyw18fAFYCid6mig6uQmH4Znz4jz7ZqyFm5gNGAk97nUWkpphZM+Ak4BkA51yJSp8nTgfWqfTVqDiggZnFAQ2BLR7nqRQVv+qTCOQdcTuI3gBLFDGzLkAKsMDjKFEjPNQwB9gBfOCc07mvOQ8CvwZCHueIVg5438yyzOwar8NEkWRgJ/BceJjz02bWyOtQUehi4BWvQ0QL51w+8HdgM7AVKHDOve9tqspR8RORKmdmjYHXgJ855/Z7nSdaOOfKnXODAB8wxMw0zLkGmNn5wA7nXJbXWaLYic65VOBc4IbwcH+pfnFAKvCYcy4FOAhoTYMaFB5eOxrI9DpLtDCzFlSM4ksGOgKNzOwyb1NVjopf9ckHko647QsfE4lo4fllrwEvOeeme50nGoWHWs0BzvE4SrQYDowOzzObCpxmZv/yNlJ0CX8Cj3NuB/A6FdMtpPoFgeARowumUVEEpeacC2Q757Z7HSSKnAFscM7tdM6VAtOBYR5nqhQVv+qzEOhhZsnhT2MuBmZ4nEmkWoUXGHkGWOmcu9/rPNHEzNqYWfPw1w2oWFhqlaehooRz7rfOOZ9zrgsVv+s/cs7ViU9/I4GZNQovJkV4mOFZgFZ0rgHOuW1Anpn1DB86HdBiXjVrIhrmWdM2AyeYWcPw+57TqVjToNaL8zpApHLOlZnZjcB7QCzwrHNuucexooaZvQKcArQ2syDwR+fcM96migrDgcuBpeG5ZgC/c869412kqNEBeCG8ulsMkOGc07YCEg3aAa9XvP8iDnjZOTfL20hR5SbgpfCH3OuBKzzOEzXCH3ScCVzrdZZo4pxbYGbTgGygDFgEPOltqsrRdg4iIiIiIiIRTkM9RUREREREIpyKn4iIiIiISIRT8RMREREREYlwKn4iIiIiIiIRTsVPREREREQkwqn4iYiIAGZWbmY5R/y5rQqfu4uZaW85ERHxjPbxExERqXDYOTfI6xAiIiLVQVf8REREjsHMNprZvWa21My+NLPu4eNdzOwjM1tiZrPNrFP4eDsze93MFof/DAs/VayZPWVmy83sfTNrEH78zWa2Ivw8Uz36MUVEJMKp+ImIiFRo8LWhnhcdcV+Bc64/MBl4MHzsEeAF59wA4CXg4fDxh4FPnHMDgVRgefh4D2CKc64vsA8YFz5+G5ASfp6fVs+PJiIi0c6cc15nEBER8ZyZFTrnGn/D8Y3Aac659WYWD2xzzrUys11AB+dcafj4VudcazPbCficc8VHPEcX4APnXI/w7d8A8c65u8xsFlAIvAG84ZwrrOYfVUREopCu+ImIiHw7d5Svv4viI74u5z/z7EcCU6i4OrjQzDT/XkREqpyKn4iIyLe76Ii/54W//gK4OPz1pcBn4a9nA9cBmFmsmTU72pOaWQyQ5JybA/wGaAb8z1VHERGRH0qfKoqIiFRoYGY5R9ye5Zz7akuHFma2hIqrdhPDx24CnjOzXwE7gSvCx28BnjSzq6i4sncdsPUorxkL/CtcDg142Dm3r4p+HhERkX/THD8REZFjCM/x8zvndnmdRURE5PvSUE8REREREZEIpyt+IiIiIiIiEU5X/ERERERERCKcip+IiIiIiEiEU/ETERERERGJcCp+IiIiIiIiEU7FT0REREREJML9P9EAAB5lu2+xAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# active learning\n",
    "# w/o wasserstein loss\n",
    "# source clusters\n",
    "import math \n",
    "fig, ax=plt.subplots(2, figsize=(15,10))\n",
    "\n",
    "# load model weights\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "print(\"Loaded model weights\")\n",
    "\n",
    "# load replay buffer params\n",
    "means = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/means_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "means = np.nan_to_num(means)\n",
    "covs = np.load(\"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/extras/covs_1024x512_vgg16deeplabv3_SYNTHIA_VIPER_DENSE_13_1.npy\")\n",
    "print(\"Loaded means and covariances for replay buffer\")\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs2 = 20\n",
    "epochstep = 2\n",
    "\n",
    "\n",
    "# active learning set up \n",
    "ncentroids = 10\n",
    "with open('/global/cscratch1/sd/saksman/accda/anchors/cluster_centroids_full_%d.pkl' % ncentroids, 'rb') as f: \n",
    "    cluster_centroids = pickle.load(f)\n",
    "cluster_centroids = cluster_centroids.reshape(ncentroids, num_classes, 256)\n",
    "print(\"Loaded cluster centroids for active learning\")\n",
    "print(cluster_centroids.shape)\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "loss = []\n",
    "target_miou = []\n",
    "\n",
    "for itr in range(epochs2):\n",
    "    \n",
    "    if itr%epochstep == 0:\n",
    "        print('Iteration: {}'.format(itr))\n",
    "    \n",
    "    target_train_data, target_train_labels = utils.sample_batch(cityscapes_data_dir, batch_size=batch_size, seed=itr)\n",
    "    # target_train_labels_cat = tf.keras.utils.to_categorical(target_train_labels, num_classes=num_classes)\n",
    "     \n",
    "    if active_learning and strategy is not None:\n",
    "        pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "        X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.1, num_classes, strategy = strategy)\n",
    "        Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "        all_selected_idx.append(selected_idx)\n",
    "    else:\n",
    "        pass # TODO: run unsupervised method here\n",
    "\n",
    "    # make sure the #samples from gaussians match the distribution of the labels\n",
    "    n_samples = np.zeros(num_classes, dtype=int)\n",
    "    cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "    for i in range(len(cls)):\n",
    "        if cls[i] == 0:\n",
    "            continue\n",
    "        n_samples[cls[i]] = ns[i]\n",
    "\n",
    "    if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "        remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "        aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "        aux *= remaining\n",
    "        aux = np.floor(aux).astype('int')\n",
    "        \n",
    "        n_samples += aux\n",
    "\n",
    "        # in case there are extra samples left, dump them on the highest represented class\n",
    "        n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "    Yembed,Yembedlabels = utils.sample_from_gaussians(means, covs, n_samples = n_samples)\n",
    "    Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "    Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "    Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "    \n",
    "    theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "    \n",
    "    loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "    miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "    target_miou.append(miou)\n",
    "    \n",
    "    if miou > best_miou:\n",
    "        best_miou = miou\n",
    "        best_combined = combined_\n",
    "    \n",
    "    print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "    if itr%epochstep==0:\n",
    "        # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "        # the ignore class\n",
    "        if itr != 0:\n",
    "            ax[0].clear()\n",
    "            \n",
    "            ll = np.asarray(loss)\n",
    "            print(ll)\n",
    "            ax[0].plot(ll[:,0], label='total loss')\n",
    "            ax[0].plot(ll[:,1] * lambda1, label='replay buffer ce loss')\n",
    "            ax[0].plot(ll[:,2] * lambda2, label='active learning ce loss')\n",
    "            ax[0].legend()\n",
    "            \n",
    "        ax[0].set_title(\"Loss\")\n",
    "        ax[0].set_xlabel(\"Epochs\")\n",
    "        ax[0].set_ylabel(\"Loss\")\n",
    "        \n",
    "        if itr != 0:\n",
    "            ax[1].clear()\n",
    "            ax[1].plot(np.asarray(target_miou))\n",
    "        \n",
    "        ax[1].set_title(\"MIOU on target domain\")\n",
    "        ax[1].set_xlabel(\"Epochs\")\n",
    "        ax[1].set_ylabel(\"Mean IOU\")\n",
    "        \n",
    "        display.display(plt.gcf())\n",
    "        display.clear_output(wait=True)\n",
    "        \n",
    "        deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "        classifier.save_weights(fn_w_adapted_cls)\n",
    "        \n",
    "        time.sleep(1e-3) \n",
    "        \n",
    "    if itr == epochs2 - 1:\n",
    "        # run for best epoch\n",
    "        best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "        for k in best_cat_iou:\n",
    "            print(k, best_cat_iou[k])\n",
    "        print(best_miou2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Performance after adaptation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "source_cat_iou,source_miou = utils.compute_miou(source_data_dir, combined_, data_utils.label_ids_13, 0)\n",
    "\n",
    "for k in source_cat_iou:\n",
    "    print(k, source_cat_iou[k])\n",
    "print(source_miou)\n",
    "\n",
    "print('Computed ' + dataset + ' mIoU in', time.time() - start_time)\n",
    "\n",
    "# VIPER 1\n",
    "# road 0.45548384569118733\n",
    "# sidewalk 0.22908076309616884\n",
    "# building 0.7231414831527551\n",
    "# traffic light 0.13056876105257134\n",
    "# traffic sign 0.10746477632561094\n",
    "# vegetation 0.29910880116931143\n",
    "# sky 0.7702276075574622\n",
    "# person 0.3968236052300994\n",
    "# rider 0.0012631338117709702\n",
    "# car 0.3250214488701445\n",
    "# bus 0.011364789604663108\n",
    "# motorcycle 0.0014075893390935838\n",
    "# bicycle 0.0013295668128586176\n",
    "# 0.26556047474720745\n",
    "# Computed SYNTHIA_VIPER_DENSE_13_1 mIoU in 701.9752759933472\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "cit_cat_iou,cit_miou = utils.compute_miou(cityscapes_data_dir, combined_, data_utils.label_ids_13, 0)\n",
    "\n",
    "for k in cit_cat_iou:\n",
    "    print(k, cit_cat_iou[k])\n",
    "print(cit_miou)\n",
    "\n",
    "print('Computed VIPER mIoU in', time.time() - start_time)\n",
    "\n",
    "# VIPER-1\n",
    "# road 0.8170088043832361\n",
    "# sidewalk 0.33026286655816334\n",
    "# building 0.632127644160952\n",
    "# traffic light 0.4009491872069492\n",
    "# traffic sign 0.006254310466994813\n",
    "# vegetation 0.1306418583589563\n",
    "# sky 0.8650156523569378\n",
    "# person 0.19010653822990375\n",
    "# rider 0.0\n",
    "# car 0.8404013262102543\n",
    "# bus 0.0\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.32405909137941136\n",
    "# Computed VIPER mIoU in 18.68591570854187\n",
    "\n",
    "# VIPER-2\n",
    "# road 0.6166949852470773\n",
    "# sidewalk 0.3787735571503562\n",
    "# building 0.6049256780597033\n",
    "# traffic light 0.186964255325101\n",
    "# traffic sign 0.0008168001712463807\n",
    "# vegetation 0.09151754024216284\n",
    "# sky 0.8323521128681671\n",
    "# person 0.47658649376653717\n",
    "# rider 0.0\n",
    "# car 0.8210425378196808\n",
    "# bus 0.3683831960382732\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.33677362743756195\n",
    "# Computed VIPER 2 mIoU in 12.673284530639648\n",
    "\n",
    "# VIPER-3\n",
    "# road 0.7241212757775349\n",
    "# sidewalk 0.20394133957802926\n",
    "# building 0.4374052246162175\n",
    "# traffic light 0.1886698588396244\n",
    "# traffic sign 0.0013552072158326775\n",
    "# vegetation 0.23934770268178143\n",
    "# sky 0.8418698288550998\n",
    "# person 1.9359460000129065e-05\n",
    "# rider 0.0\n",
    "# car 0.8522794354467441\n",
    "# bus 0.0\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.26838532557468187\n",
    "# Computed VIPER 3 mIoU in 9.569716215133667\n",
    "\n",
    "# VIPER-4\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_4/vgg16_deeplabv3_adapted.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_4/vgg16_deeplabv3_classifier_adapted.h5')\n",
    "\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "\n",
    "start_time = time.time()\n",
    "source_cat_iou,source_miou = utils.compute_miou('./processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence010/', combined_, data_utils.label_ids_13, 0)\n",
    "for k in source_cat_iou:\n",
    "    print(k, source_cat_iou[k])\n",
    "print(source_miou)\n",
    "\n",
    "print('Computed ' + dataset + ' mIoU in', time.time() - start_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 \t : \t 0.3912087963836125\n",
      "2 \t : \t 0.3399949880875378\n",
      "3 \t : \t 0.2981430084764967\n",
      "4 \t : \t 0.32685762021034603\n"
     ]
    }
   ],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_adapted.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_classifier_adapted.h5')\n",
    "\n",
    "\n",
    "# deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "# classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "\n",
    "for target_i in range(1,11):\n",
    "    target_data_dir = '/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence' + str(target_i).rjust(3, '0') + '/'\n",
    "    source_cat_iou,source_miou = utils.compute_miou(target_data_dir, combined_, data_utils.label_ids_13, [])\n",
    "    # for k in source_cat_iou:\n",
    "    #     print(k, source_cat_iou[k])\n",
    "    print(target_i, '\\t : \\t', source_miou)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 \t : \t 0.3304527332435952\n",
      "2 \t : \t 0.33019344797055583\n",
      "3 \t : \t 0.2819066622949565\n",
      "4 \t : \t 0.31654615749708054\n",
      "5 \t : \t 0.3073701739865\n",
      "6 \t : \t 0.3069781458418397\n",
      "7 \t : \t 0.2667288525194096\n",
      "8 \t : \t 0.24169105277904077\n",
      "9 \t : \t 0.27019105419790657\n",
      "10 \t : \t 0.24983465619512438\n"
     ]
    }
   ],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_adapted.h5')\n",
    "classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/SYNTHIA_VIPER_DENSE_13_1/vgg16_deeplabv3_classifier_adapted.h5')\n",
    "\n",
    "\n",
    "# deeplabv3.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3.h5')\n",
    "# classifier.load_weights('/global/cfs/cdirs/m3691/segmentation_datasets/fangjun/SYNTHIA_VIPER_COMBINED_13/vgg16_deeplabv3_classifier.h5')\n",
    "\n",
    "for target_i in range(1,11):\n",
    "    target_data_dir = './processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence' + str(target_i).rjust(3, '0') + '/'\n",
    "    source_cat_iou,source_miou = utils.compute_miou(target_data_dir, combined_, data_utils.label_ids_13, 0)\n",
    "    # for k in source_cat_iou:\n",
    "    #     print(k, source_cat_iou[k])\n",
    "    print(target_i, '\\t : \\t', source_miou)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss = []\n",
    "target_miou = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "adapting to target domain 2 \t /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/ \t /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence002/\n",
      "0\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/427_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/428_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/429_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/430_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/431_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/432_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/433_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/434_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/435_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/436_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/437_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/438_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/439_image.npy could not be processed\n",
      "computed means in for target 2 778.3026814460754\n",
      "0\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/427_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/428_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/429_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/430_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/431_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/432_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/433_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/434_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/435_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/436_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/437_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/438_image.npy could not be processed\n",
      "Image /global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/439_image.npy could not be processed\n",
      "finished training gaussians in 831.0172436237335\n",
      "number of epochs for current iter :  20\n"
     ]
    }
   ],
   "source": [
    "prev_target_data_dir = '/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence' + str(target_i-1).rjust(3, '0') + '/'\n",
    "target_data_dir = '/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence' + str(target_i).rjust(3, '0') + '/'\n",
    "\n",
    "print('adapting to target domain {}'.format(target_i), '\\t', prev_target_data_dir, '\\t', target_data_dir)\n",
    "\n",
    "# learn gaussians for the previous target domain   \n",
    "start_time = time.time()\n",
    "cur_means, _, cur_ct = utils.learn_gaussians_seq(prev_target_data_dir, deeplabv3, combined, batch_size, data_utils_new.label_ids_13)\n",
    "print(\"computed means in for target {}\".format(target_i), time.time() - start_time)\n",
    "\n",
    "start_time = time.time()\n",
    "cur_means, cur_covs, cur_ct = utils.learn_gaussians_seq(prev_target_data_dir, deeplabv3, combined, batch_size, data_utils_new.label_ids_13, initial_means=cur_means)\n",
    "print(\"finished training gaussians in\", time.time() - start_time)\n",
    "\n",
    "cur_means = np.nan_to_num(cur_means)\n",
    "\n",
    "target_means.append(cur_means)\n",
    "target_covs.append(cur_covs)\n",
    "\n",
    "cur_epochs = max(epochs_target, epochs_target // (target_i - 1))\n",
    "print('number of epochs for current iter : ', cur_epochs) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration: 0\n"
     ]
    }
   ],
   "source": [
    "import data_utils_new\n",
    "importlib.reload(utils)\n",
    "\n",
    "dataset = \"SYNTHIA_VIPER_DENSE_13_\"\n",
    "\n",
    "fig,ax=plt.subplots(2,figsize=(15,10))\n",
    "\n",
    "batch_size = 20\n",
    "\n",
    "thres = .99\n",
    "epochs_target = 20\n",
    "epochstep = 2\n",
    "\n",
    "log_f = open(\"log.txt\", \"a\")\n",
    "\n",
    "starting_target = 2\n",
    "\n",
    "best_miou = -math.inf\n",
    "best_ious = []\n",
    "active_learning = True\n",
    "strategy = \"ours\"\n",
    "all_selected_idx = []\n",
    "\n",
    "for target_i in range(starting_target, 10):   \n",
    "\n",
    "    for itr in range(epochs_target):\n",
    "    \n",
    "        if itr%epochstep == 0:\n",
    "            print('Iteration: {}'.format(itr))\n",
    "            \n",
    "        for prev_target in range(0, target_i-1):            \n",
    "           \n",
    "            # target_train_data, target_train_labels = utils.sample_batch(target_data_dir, batch_size=batch_size, seed=itr)\n",
    "            target_train_data, target_train_labels = utils.sample_batch(target_data_dir, batch_size=batch_size, seed=itr)\n",
    "            \n",
    "            if active_learning and strategy is not None:\n",
    "                pred_target_labels = deeplabv3(target_train_data)\n",
    "\n",
    "                X_labeled, Y_labeled, X_eval, Y_eval, selected_idx = get_active_XY(target_train_data, pred_target_labels, target_train_labels, 256, batch_size, 0.2, num_classes, strategy = strategy)\n",
    "                Y_labeled = tf.keras.utils.to_categorical(Y_labeled, num_classes=num_classes)\n",
    "                all_selected_idx.append(selected_idx)\n",
    "\n",
    "            # make sure the #samples from gaussians match the distribution of the labels\n",
    "            n_samples = np.zeros(num_classes, dtype=int)\n",
    "            cls, ns = np.unique(target_train_labels, return_counts=True)\n",
    "            for i in range(len(cls)):\n",
    "                if cls[i] == 0:\n",
    "                    continue\n",
    "                n_samples[cls[i]] = ns[i]\n",
    "\n",
    "            if np.sum(n_samples) % np.prod(img_shape) != 0:\n",
    "                remaining = np.prod(img_shape[:-1]) - np.sum(n_samples) % np.prod(img_shape[:-1])\n",
    "\n",
    "                aux = np.copy(n_samples) / np.sum(n_samples)\n",
    "                aux *= remaining\n",
    "                aux = np.floor(aux).astype('int')\n",
    "\n",
    "                n_samples += aux\n",
    "\n",
    "                # in case there are extra samples left, dump them on the highest represented class\n",
    "                n_samples[np.argmax(n_samples)] += remaining - np.sum(aux)\n",
    "\n",
    "            Yembed,Yembedlabels = utils.sample_from_gaussians(target_means[prev_target], target_covs[prev_target], n_samples = n_samples)\n",
    "            Yembed = Yembed.reshape(-1, img_shape[0], img_shape[1], num_classes)\n",
    "            Yembedlabels = Yembedlabels.reshape(-1, img_shape[0], img_shape[1])\n",
    "            Yembedlabels = tf.keras.utils.to_categorical(Yembedlabels, num_classes=num_classes)\n",
    "            \n",
    "            # theta_instance = tf.keras.backend.variable(wasserstein_utils.generateTheta(num_projections,num_classes))\n",
    "            loss.append(train(inputs=[target_train_data, Yembed, Yembedlabels, X_labeled, Y_labeled]))\n",
    "            miou = combined_.evaluate(X_eval, Y_eval, verbose=False)[-1] * 14/13\n",
    "            target_miou.append(miou)\n",
    "            \n",
    "            if miou > best_miou:\n",
    "                best_miou = miou\n",
    "                best_combined = combined_\n",
    "    \n",
    "            print(\"mIoU at {0} epochs: {1}, best mIoU: {2}\".format(itr, miou, best_miou)) \n",
    "\n",
    "        if itr%epochstep==0:\n",
    "            # Debug info. First, the mIoU. Second, the categorical CE loss (ignoring class weights and containing) \n",
    "            # the ignore class\n",
    "            if itr != 0:\n",
    "                ax[0].clear()\n",
    "\n",
    "                ll = np.asarray(loss)\n",
    "                ax[0].plot(ll[:,0], label='total loss')\n",
    "                ax[0].plot(ll[:,1] * lambda2, label='active learning ce loss')\n",
    "                ax[0].plot(ll[:,2] * lambda1, label='replay buffer ce loss')\n",
    "                ax[0].legend()\n",
    "\n",
    "            ax[0].set_title(\"Loss\")\n",
    "            ax[0].set_xlabel(\"Epochs\")\n",
    "            ax[0].set_ylabel(\"Loss\")\n",
    "\n",
    "            if itr != 0:\n",
    "                ax[1].clear()\n",
    "                ax[1].plot(np.asarray(target_miou))\n",
    "\n",
    "            ax[1].set_title(\"MIOU on target domain\")\n",
    "            ax[1].set_xlabel(\"Epochs\")\n",
    "            ax[1].set_ylabel(\"Mean IOU\")\n",
    "\n",
    "            display.display(plt.gcf())\n",
    "            display.clear_output(wait=True)\n",
    "            \n",
    "            \n",
    "            fn_w_adapted_dlv3 = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + str(target_i) + \"/\" + backbone +\"_deeplabv3_adapted.h5\"\n",
    "            fn_w_adapted_cls = \"/global/cfs/cdirs/m3691/segmentation_datasets/tilek/mas3-continual_latest/weights/\" + dataset + str(target_i) + \"/\" + backbone + \"_deeplabv3_classifier_adapted.h5\"\n",
    "            \n",
    "            # print(fn_w_adapted_dlv3, fn_w_adapted_cls)\n",
    "\n",
    "            deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "            classifier.save_weights(fn_w_adapted_cls)\n",
    "\n",
    "            time.sleep(1e-3)    \n",
    "        \n",
    "        if itr == epochs2 - 1:\n",
    "            # run for best epoch\n",
    "            best_cat_iou, best_miou2 = utils.compute_miou(cityscapes_data_dir, best_combined, data_utils.label_ids_13, ids_to_ignore = all_selected_idx)\n",
    "            for k in best_cat_iou:\n",
    "                print(k, best_cat_iou[k])\n",
    "            print(best_miou2)\n",
    "            \n",
    "    print('completed adapting to target domain {}'.format(target_i), '\\t', prev_target_data_dir, '\\t', target_data_dir)\n",
    "\n",
    "    # calculate mIoU\n",
    "    start_time = time.time()\n",
    "    cit_cat_iou,cit_miou = utils.compute_miou(target_data_dir, combined_, data_utils.label_ids_13, 0)\n",
    "    for k in cit_cat_iou:\n",
    "        print(k, cit_cat_iou[k])\n",
    "    \n",
    "    print(cit_miou)\n",
    "    log_f.write(str(target_i) + ' : ' + str(cit_miou))    \n",
    "\n",
    "    print('Computed VIPER {} mIoU in'.format(target_i), time.time() - start_time)\n",
    "    \n",
    "log_f.close()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x1008 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "17.204325437545776\n"
     ]
    }
   ],
   "source": [
    "import data_utils_new\n",
    "import matplotlib.pyplot as plt \n",
    "import umap.umap_ as umap\n",
    "# Generate data from the gmm model and plot it\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "n_samples = np.ones(num_classes, dtype=int)\n",
    "n_samples[0] = 0\n",
    "n_samples *= 2000\n",
    "\n",
    "xx, yy = utils.sample_from_gaussians(means, covs, n_samples=n_samples)\n",
    "\n",
    "NUM_COLORS = num_classes\n",
    "\n",
    "reducer = umap.UMAP()\n",
    "\n",
    "umap_embedding = reducer.fit_transform(xx)\n",
    "\n",
    "plt.figure(figsize=(16,14))\n",
    "cm = plt.get_cmap('gist_rainbow')\n",
    "\n",
    "shift = 1 / len(data_utils_new.label_ids_13.keys())\n",
    "idx = 0\n",
    "for label in data_utils_new.label_ids_13:\n",
    "    ind = yy == data_utils_new.label_ids_13[label]\n",
    "    \n",
    "    plt.scatter(umap_embedding[:,0][ind], umap_embedding[:,1][ind], label=label, \\\n",
    "                color=cm(1.*idx/NUM_COLORS))\n",
    "    idx += 1\n",
    "\n",
    "plt.title(\"Embedding scatter-plot\")\n",
    "plt.legend()\n",
    "    \n",
    "plt.show()\n",
    "\n",
    "print(time.time() - start_time)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded model weights\n"
     ]
    }
   ],
   "source": [
    "deeplabv3.load_weights(fn_w_adapted_dlv3)\n",
    "classifier.load_weights(fn_w_adapted_cls)\n",
    "print(\"Loaded model weights\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "road 0.8308608968675483\n",
      "sidewalk 0.27348768244553456\n",
      "building 0.6513443882341332\n",
      "traffic light 0.37291672961621636\n",
      "traffic sign 0.0038099773782593166\n",
      "vegetation 0.16323353477649255\n",
      "sky 0.888175399577542\n",
      "person 0.19304997344212002\n",
      "rider 0.0\n",
      "car 0.8840926390088035\n",
      "bus 0.0\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.32776701702666533\n",
      "Computed SYNTHIA_VIPER_DENSE_13_1 mIoU in 159.04997277259827\n"
     ]
    }
   ],
   "source": [
    "import utils\n",
    "importlib.reload(utils)\n",
    "\n",
    "start_time = time.time()\n",
    "\n",
    "source_cat_iou,source_miou = utils.compute_miou('./processed-data/1024x512/13_classes/VIPER_DENSE_SEQ/Sequence001/', combined_, data_utils.label_ids_13, 0)\n",
    "\n",
    "for k in source_cat_iou:\n",
    "    print(k, source_cat_iou[k])\n",
    "print(source_miou)\n",
    "\n",
    "print('Computed ' + dataset + ' mIoU in', time.time() - start_time)\n",
    "\n",
    "road 0.8308608968675483\n",
    "sidewalk 0.27348768244553456\n",
    "building 0.6513443882341332\n",
    "traffic light 0.37291672961621636\n",
    "traffic sign 0.0038099773782593166\n",
    "vegetation 0.16323353477649255\n",
    "sky 0.888175399577542\n",
    "person 0.19304997344212002\n",
    "rider 0.0\n",
    "car 0.8840926390088035\n",
    "bus 0.0\n",
    "motorcycle 0.0\n",
    "bicycle 0.0\n",
    "0.32776701702666533\n",
    "Computed SYNTHIA_VIPER_DENSE_13_1 mIoU in 159.04997277259827"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# road 0.8615069121641106\n",
    "# sidewalk 0.9000754476234025\n",
    "# building 0.9430840666603088\n",
    "# traffic light 0.5906668306793044\n",
    "# traffic sign 0.4751323753940848\n",
    "# vegetation 0.8432769737226227\n",
    "# sky 0.9643724556369807\n",
    "# person 0.7238095621515034\n",
    "# rider 0.6191529188583779\n",
    "# car 0.8826804768268254\n",
    "# bus 0.9113284943230776\n",
    "# motorcycle 0.6885963930665517\n",
    "# bicycle 0.3477612301913242\n",
    "# 0.7501110874844981\n",
    "# Computed SYNTHIA_VIPER_DENSE_13 mIoU in 209.03599977493286\n",
    "\n",
    "# VIPER1\n",
    "# road 0.7598092688219289\n",
    "# sidewalk 0.3371948453270736\n",
    "# building 0.5888429412251485\n",
    "# traffic light 0.2412746294793218\n",
    "# traffic sign 0.007871538466283783\n",
    "# vegetation 0.07906209988581193\n",
    "# sky 0.8681167733611675\n",
    "# person 0.15913067911219309\n",
    "# rider 0.0\n",
    "# car 0.8381470531819963\n",
    "# bus 0.0\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.2984192176046866\n",
    "# Computed SYNTHIA_VIPER_DENSE_13_2 mIoU in 12.690653085708618\n",
    "\n",
    "# VIPER2\n",
    "# road 0.12746921970993438\n",
    "# sidewalk 0.01654514397536747\n",
    "# building 0.08359767761356621\n",
    "# traffic light 0.0014628747427203985\n",
    "# traffic sign 0.0004797761683849531\n",
    "# vegetation 0.0010155479171658137\n",
    "# sky 0.06442882520719335\n",
    "# person 1.3905304873809358e-05\n",
    "# rider 0.0\n",
    "# car 0.013538577154181806\n",
    "# bus 0.0040766048179061275\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.024048319431638026\n",
    "# Computed SYNTHIA_VIPER_DENSE_13_3 mIoU in 16.93789768218994\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "100\n",
      "200\n",
      "300\n",
      "400\n",
      "road 0.16361121759652467\n",
      "sidewalk 5.5880594346001465e-06\n",
      "building 0.0537824356711514\n",
      "traffic light 0.004082077289533864\n",
      "traffic sign 0.0\n",
      "vegetation 0.004795258655464463\n",
      "sky 0.03790733322264466\n",
      "person 0.0\n",
      "rider 0.0\n",
      "car 0.02297664564513909\n",
      "bus 0.0\n",
      "motorcycle 0.0\n",
      "bicycle 0.0\n",
      "0.022089273549222518\n",
      "Computed VIPER mIoU in 8.972235918045044\n"
     ]
    }
   ],
   "source": [
    "start_time = time.time()\n",
    "\n",
    "cit_cat_iou,cit_miou = utils.compute_miou(cityscapes_data_dir, combined_, data_utils.label_ids_13, 0)\n",
    "\n",
    "for k in cit_cat_iou:\n",
    "    print(k, cit_cat_iou[k])\n",
    "print(cit_miou)\n",
    "\n",
    "print('Computed VIPER mIoU in', time.time() - start_time)\n",
    "\n",
    "# road 0.7883116582422506\n",
    "# sidewalk 0.6215464735885322\n",
    "# building 0.751691007583865\n",
    "# wall 0.3993120030432491\n",
    "# fence 0.39173486089325144\n",
    "# pole 0.302509334397975\n",
    "# traffic light 0.3019573318279201\n",
    "# traffic sign 0.38603017233885695\n",
    "# vegetation 0.7792104073113252\n",
    "# terrain 0.484321694588526\n",
    "# sky 0.8323752669785053\n",
    "# person 0.5206867614087272\n",
    "# rider 0.2377222662760571\n",
    "# car 0.8020324193022697\n",
    "# truck 0.4754062370048565\n",
    "# bus 0.593789975462107\n",
    "# train 0.5966522990567177\n",
    "# motorcycle 0.3119814031424603\n",
    "# bicycle 0.47793033593638945\n",
    "# 0.5292211530728338\n",
    "# Computed CITYSCAPES mIoU in 232.43373942375183"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# NO ADAPTATION\n",
    "\n",
    "# road 0.4343089698314718\n",
    "# sidewalk 0.11156745414108413\n",
    "# building 0.3455360124643591\n",
    "# traffic light 0.13244824374732939\n",
    "# traffic sign 0.020891181218774212\n",
    "# vegetation 0.16245698445827797\n",
    "# sky 0.7351399630612382\n",
    "# person 0.18702045190416963\n",
    "# rider 0.0\n",
    "# car 0.8301180914913947\n",
    "# bus 0.06320265566955835\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.23251461599905057\n",
    "# Computed VIPER mIoU in 94.76635837554932\n",
    "\n",
    "# ALL IN ONE\n",
    "\n",
    "# road 0.6900469076232928\n",
    "# sidewalk 0.3999559532819935\n",
    "# building 0.4477707606588546\n",
    "# traffic light 0.14471677916664455\n",
    "# traffic sign 0.005208014396949388\n",
    "# vegetation 0.22113673614692125\n",
    "# sky 0.8054719866917823\n",
    "# person 0.20232922242992005\n",
    "# rider 0.0\n",
    "# car 0.7608467046022307\n",
    "# bus 0.5757047777388474\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.32716829559518745\n",
    "# Computed VIPER mIoU in 133.14108419418335\n",
    "\n",
    "# SYNTHIA-VIPER1\n",
    "# road 0.8153079075418371\n",
    "# sidewalk 0.3638901684131552\n",
    "# building 0.6422570606708572\n",
    "# traffic light 0.3448658440152766\n",
    "# traffic sign 0.030519858531548108\n",
    "# vegetation 0.11682109473012166\n",
    "# sky 0.86401077761864\n",
    "# person 0.18156178781292293\n",
    "# rider 0.0\n",
    "# car 0.8640675468674487\n",
    "# bus 0.0\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.32486938816936983\n",
    "# Computed VIPER mIoU in 12.139243841171265\n",
    "\n",
    "\n",
    "# VIPER1-VIPER2\n",
    "# road 0.588499111128223\n",
    "# sidewalk 0.369365725539779\n",
    "# building 0.6204748941380721\n",
    "# traffic light 0.1665936557512192\n",
    "# traffic sign 0.01265457278785601\n",
    "# vegetation 0.043019551365977134\n",
    "# sky 0.8271666403331733\n",
    "# person 0.3732262212752325\n",
    "# rider 0.0\n",
    "# car 0.816343619490096\n",
    "# bus 0.02369959332673774\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.29546489116433583\n",
    "# Computed VIPER mIoU in 11.642083406448364\n",
    "\n",
    "# VIPER2-VIPER3\n",
    "# road 0.14499194020624828\n",
    "# sidewalk 0.010885014754372853\n",
    "# building 0.052795147047877924\n",
    "# traffic light 0.00418806705772967\n",
    "# traffic sign 0.0018961732294710642\n",
    "# vegetation 0.006274219959573628\n",
    "# sky 0.06364095194179961\n",
    "# person 0.0\n",
    "# rider 0.0\n",
    "# car 0.013660761584344825\n",
    "# bus 0.0\n",
    "# motorcycle 0.0\n",
    "# bicycle 0.0\n",
    "# 0.022948636598570604\n",
    "# Computed VIPER mIoU in 14.073742628097534\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "deeplabv3.save_weights(fn_w_adapted_dlv3)\n",
    "classifier.save_weights(fn_w_adapted_cls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "deeplabv3.load_weights(fn_w_adapted_dlv3)\n",
    "classifier.load_weights(fn_w_adapted_cls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "accda",
   "language": "python",
   "name": "accda"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}