Train_Segmentation_Model_Unet+Resnet.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3aa08b4f",
   "metadata": {},
   "source": [
    "# IMPORTS\n",
    "\n",
    "First we started with some python imports. For model building we used the <mark>**segmentation_models_pytorch library**</mark>. This has library has a lot segmentation model architechtures. Before you can import the library it needs to be ensured that the library is installed in the conda environment. For this in the anaconda terminal just type the below command to install the library:-\n",
    "\n",
    "<div class=\"alert alert-block alert-success\">\n",
    "<b>Conda Command:</b> conda install -q segmentation-models-pytorch\n",
    "</div>\n",
    "\n",
    "Once, the libray is installed it can be imported in the script safely. We also imported the utility script Trainer for model training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0b51a092",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import segmentation_models_pytorch as smp\n",
    "from Utilities.Trainer import Trainer\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torchvision import models\n",
    "import torch.multiprocessing as mp\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "35fefb16",
   "metadata": {},
   "source": [
    "The below statement helps in better error tracking when training is done in cuda. More details about this can be found in the below url:-\n",
    "\n",
    "https://pytorch.org/docs/stable/notes/cuda.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "4d9921ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "CUDA_LAUNCH_BLOCKING=\"1\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c27b412e",
   "metadata": {},
   "source": [
    "# Model Building\n",
    "\n",
    "Next a Unet model with Resnet34 encoder was created with imagenet weights from the smp library."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "7cd848ce",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unet(\n",
      "  (encoder): ResNetEncoder(\n",
      "    (conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)\n",
      "    (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "    (relu): ReLU(inplace=True)\n",
      "    (maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)\n",
      "    (layer1): Sequential(\n",
      "      (0): BasicBlock(\n",
      "        (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (1): BasicBlock(\n",
      "        (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (2): BasicBlock(\n",
      "        (conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (layer2): Sequential(\n",
      "      (0): BasicBlock(\n",
      "        (conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (downsample): Sequential(\n",
      "          (0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "          (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        )\n",
      "      )\n",
      "      (1): BasicBlock(\n",
      "        (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (2): BasicBlock(\n",
      "        (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (3): BasicBlock(\n",
      "        (conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (layer3): Sequential(\n",
      "      (0): BasicBlock(\n",
      "        (conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (downsample): Sequential(\n",
      "          (0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "          (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        )\n",
      "      )\n",
      "      (1): BasicBlock(\n",
      "        (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (2): BasicBlock(\n",
      "        (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (3): BasicBlock(\n",
      "        (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (4): BasicBlock(\n",
      "        (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (5): BasicBlock(\n",
      "        (conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "    (layer4): Sequential(\n",
      "      (0): BasicBlock(\n",
      "        (conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (downsample): Sequential(\n",
      "          (0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)\n",
      "          (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        )\n",
      "      )\n",
      "      (1): BasicBlock(\n",
      "        (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "      (2): BasicBlock(\n",
      "        (conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "        (relu): ReLU(inplace=True)\n",
      "        (conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "        (bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "      )\n",
      "    )\n",
      "  )\n",
      "  (decoder): UnetDecoder(\n",
      "    (center): Identity()\n",
      "    (blocks): ModuleList(\n",
      "      (0): DecoderBlock(\n",
      "        (conv1): Conv2dReLU(\n",
      "          (0): Conv2d(768, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention1): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "        (conv2): Conv2dReLU(\n",
      "          (0): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention2): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "      )\n",
      "      (1): DecoderBlock(\n",
      "        (conv1): Conv2dReLU(\n",
      "          (0): Conv2d(384, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention1): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "        (conv2): Conv2dReLU(\n",
      "          (0): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention2): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "      )\n",
      "      (2): DecoderBlock(\n",
      "        (conv1): Conv2dReLU(\n",
      "          (0): Conv2d(192, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention1): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "        (conv2): Conv2dReLU(\n",
      "          (0): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention2): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "      )\n",
      "      (3): DecoderBlock(\n",
      "        (conv1): Conv2dReLU(\n",
      "          (0): Conv2d(128, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention1): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "        (conv2): Conv2dReLU(\n",
      "          (0): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention2): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "      )\n",
      "      (4): DecoderBlock(\n",
      "        (conv1): Conv2dReLU(\n",
      "          (0): Conv2d(32, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention1): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "        (conv2): Conv2dReLU(\n",
      "          (0): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "          (1): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "          (2): ReLU(inplace=True)\n",
      "        )\n",
      "        (attention2): Attention(\n",
      "          (attention): Identity()\n",
      "        )\n",
      "      )\n",
      "    )\n",
      "  )\n",
      "  (segmentation_head): SegmentationHead(\n",
      "    (0): Conv2d(16, 6, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
      "    (1): Identity()\n",
      "    (2): Activation(\n",
      "      (activation): Identity()\n",
      "    )\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "ENCODER = 'resnet34'\n",
    "ENCODER_WEIGHTS = 'imagenet'\n",
    "CLASSES = 6\n",
    "ACTIVATION = None # could be None for logits or 'softmax2d' for multiclass segmentation\n",
    "\n",
    "# create segmentation model with pretrained encoder\n",
    "model = smp.Unet(\n",
    "    encoder_name=ENCODER, \n",
    "    encoder_weights=ENCODER_WEIGHTS, \n",
    "    in_channels=3,\n",
    "    classes=CLASSES, \n",
    "    activation=ACTIVATION,\n",
    ")\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bd38019",
   "metadata": {},
   "source": [
    "# Model Training\n",
    "\n",
    "Next the model training was started with 20 epochs and initial learning rate as 5e-04. The model checkpoint was saved whenever the validation dice was reported the greater than the previous best validation dice."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "b1b7964e",
   "metadata": {},
   "outputs": [],
   "source": [
    "lr = 5e-04\n",
    "epochs = 20\n",
    "path = \"./Models/model_unet.pth\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ad04d58b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current Device is GPU: 0\n",
      "Total memory allocated in GB: 0.3\n",
      "Total memory cached in GB: 0.6\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 0 | phase: train | ⏰: 10:57:58\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [04:06<00:00,  1.26it/s, loss=0.0494]\n",
      "C:\\Users\\SIDDHARTA\\OneDrive\\Documents\\NEU-DET\\Utilities\\Meter.py:40: RuntimeWarning: Mean of empty slice\n",
      "  dice_neg = np.nanmean(self.dice_neg_scores)\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.3955 | IoU: 0.0200 | dice: 0.0358 | dice_neg: nan | dice_pos: 0.0358\n",
      "Starting epoch: 0 | phase: val | ⏰: 11:02:05\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.74it/s, loss=0.0399]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.3189 | IoU: 0.0072 | dice: 0.0135 | dice_neg: nan | dice_pos: 0.0135\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 1 | phase: train | ⏰: 11:02:55\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0354]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.2834 | IoU: 0.0295 | dice: 0.0455 | dice_neg: nan | dice_pos: 0.0455\n",
      "Starting epoch: 1 | phase: val | ⏰: 11:08:15\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:49<00:00,  2.66it/s, loss=0.0292]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.2336 | IoU: 0.0726 | dice: 0.0965 | dice_neg: nan | dice_pos: 0.0965\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 2 | phase: train | ⏰: 11:09:06\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [06:09<00:00,  1.19s/it, loss=0.0285]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.2284 | IoU: 0.0735 | dice: 0.0985 | dice_neg: nan | dice_pos: 0.0985\n",
      "Starting epoch: 2 | phase: val | ⏰: 11:15:16\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:47<00:00,  2.77it/s, loss=0.0255]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.2039 | IoU: 0.1361 | dice: 0.1815 | dice_neg: nan | dice_pos: 0.1815\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 3 | phase: train | ⏰: 11:16:04\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0253]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.2026 | IoU: 0.1084 | dice: 0.1433 | dice_neg: nan | dice_pos: 0.1433\n",
      "Starting epoch: 3 | phase: val | ⏰: 11:21:25\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 133/133 [00:49<00:00,  2.69it/s, loss=0.023]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1836 | IoU: 0.1747 | dice: 0.2245 | dice_neg: nan | dice_pos: 0.2245\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 4 | phase: train | ⏰: 11:22:15\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0231]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1851 | IoU: 0.1546 | dice: 0.1988 | dice_neg: nan | dice_pos: 0.1988\n",
      "Starting epoch: 4 | phase: val | ⏰: 11:27:35\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.73it/s, loss=0.0198]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1585 | IoU: 0.2154 | dice: 0.2611 | dice_neg: nan | dice_pos: 0.2611\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 5 | phase: train | ⏰: 11:28:25\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0205]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1637 | IoU: 0.2071 | dice: 0.2625 | dice_neg: nan | dice_pos: 0.2625\n",
      "Starting epoch: 5 | phase: val | ⏰: 11:33:45\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.75it/s, loss=0.0159]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1275 | IoU: 0.2336 | dice: 0.2785 | dice_neg: nan | dice_pos: 0.2785\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 6 | phase: train | ⏰: 11:34:34\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0184]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1472 | IoU: 0.2604 | dice: 0.3294 | dice_neg: nan | dice_pos: 0.3294\n",
      "Starting epoch: 6 | phase: val | ⏰: 11:39:54\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [08:44<00:00,  3.95s/it, loss=0.0137]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1097 | IoU: 0.4190 | dice: 0.5181 | dice_neg: nan | dice_pos: 0.5181\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 7 | phase: train | ⏰: 11:48:40\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 310/310 [01:50<00:00,  2.81it/s, loss=0.017]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1362 | IoU: 0.3676 | dice: 0.4680 | dice_neg: nan | dice_pos: 0.4680\n",
      "Starting epoch: 7 | phase: val | ⏰: 11:50:31\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.73it/s, loss=0.0137]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1096 | IoU: 0.4965 | dice: 0.6251 | dice_neg: nan | dice_pos: 0.6251\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 8 | phase: train | ⏰: 11:51:20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 310/310 [05:19<00:00,  1.03s/it, loss=0.015]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1201 | IoU: 0.4766 | dice: 0.6015 | dice_neg: nan | dice_pos: 0.6015\n",
      "Starting epoch: 8 | phase: val | ⏰: 11:56:39\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.73it/s, loss=0.013]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1043 | IoU: 0.5390 | dice: 0.6578 | dice_neg: nan | dice_pos: 0.6578\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 9 | phase: train | ⏰: 11:57:29\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.0145]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1156 | IoU: 0.5165 | dice: 0.6446 | dice_neg: nan | dice_pos: 0.6446\n",
      "Starting epoch: 9 | phase: val | ⏰: 12:02:49\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.72it/s, loss=0.0132]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1052 | IoU: 0.5064 | dice: 0.6154 | dice_neg: nan | dice_pos: 0.6154\n",
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 10 | phase: train | ⏰: 12:03:38\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [05:19<00:00,  1.03s/it, loss=0.0128]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.1026 | IoU: 0.5697 | dice: 0.6985 | dice_neg: nan | dice_pos: 0.6985\n",
      "Starting epoch: 10 | phase: val | ⏰: 12:08:58\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.74it/s, loss=0.00992]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0794 | IoU: 0.6126 | dice: 0.7393 | dice_neg: nan | dice_pos: 0.7393\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 11 | phase: train | ⏰: 12:09:47\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 310/310 [05:20<00:00,  1.03s/it, loss=0.012]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0957 | IoU: 0.5833 | dice: 0.7115 | dice_neg: nan | dice_pos: 0.7115\n",
      "Starting epoch: 11 | phase: val | ⏰: 12:15:08\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 133/133 [03:22<00:00,  1.53s/it, loss=0.0106]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0848 | IoU: 0.6059 | dice: 0.7340 | dice_neg: nan | dice_pos: 0.7340\n",
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 12 | phase: train | ⏰: 12:18:31\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [04:41<00:00,  1.10it/s, loss=0.0118]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0944 | IoU: 0.5851 | dice: 0.7137 | dice_neg: nan | dice_pos: 0.7137\n",
      "Starting epoch: 12 | phase: val | ⏰: 12:23:12\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:49<00:00,  2.71it/s, loss=0.00933]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0747 | IoU: 0.6439 | dice: 0.7646 | dice_neg: nan | dice_pos: 0.7646\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 13 | phase: train | ⏰: 12:24:02\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|████████████████████████████████████████████████████████████████████| 310/310 [20:44<00:00,  4.02s/it, loss=0.011]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0881 | IoU: 0.6038 | dice: 0.7311 | dice_neg: nan | dice_pos: 0.7311\n",
      "Starting epoch: 13 | phase: val | ⏰: 12:44:47\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:21<00:00,  6.30it/s, loss=0.00956]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0765 | IoU: 0.6492 | dice: 0.7701 | dice_neg: nan | dice_pos: 0.7701\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 14 | phase: train | ⏰: 12:45:09\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|███████████████████████████████████████████████████████████████████| 310/310 [01:34<00:00,  3.27it/s, loss=0.0103]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0823 | IoU: 0.6288 | dice: 0.7540 | dice_neg: nan | dice_pos: 0.7540\n",
      "Starting epoch: 14 | phase: val | ⏰: 12:46:43\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:38<00:00,  3.44it/s, loss=0.00817]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0653 | IoU: 0.6730 | dice: 0.7893 | dice_neg: nan | dice_pos: 0.7893\n",
      "******** New optimal found, saving state ********\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 15 | phase: train | ⏰: 12:47:23\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|█████████████████████████████████████████████████████████████████████| 310/310 [04:38<00:00,  1.11it/s, loss=0.01]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0802 | IoU: 0.6336 | dice: 0.7562 | dice_neg: nan | dice_pos: 0.7562\n",
      "Starting epoch: 15 | phase: val | ⏰: 12:52:02\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:49<00:00,  2.71it/s, loss=0.00847]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0678 | IoU: 0.6217 | dice: 0.7422 | dice_neg: nan | dice_pos: 0.7422\n",
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 16 | phase: train | ⏰: 12:52:51\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 310/310 [13:35<00:00,  2.63s/it, loss=0.00977]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0782 | IoU: 0.6337 | dice: 0.7559 | dice_neg: nan | dice_pos: 0.7559\n",
      "Starting epoch: 16 | phase: val | ⏰: 13:06:26\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:20<00:00,  6.60it/s, loss=0.00979]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0783 | IoU: 0.5907 | dice: 0.7148 | dice_neg: nan | dice_pos: 0.7148\n",
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 17 | phase: train | ⏰: 13:06:46\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 310/310 [03:01<00:00,  1.71it/s, loss=0.00924]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0739 | IoU: 0.6393 | dice: 0.7620 | dice_neg: nan | dice_pos: 0.7620\n",
      "Starting epoch: 17 | phase: val | ⏰: 13:09:48\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.76it/s, loss=0.00809]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0647 | IoU: 0.6388 | dice: 0.7528 | dice_neg: nan | dice_pos: 0.7528\n",
      "\n",
      "Learning Rate =  0.0005\n",
      "Starting epoch: 18 | phase: train | ⏰: 13:10:36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 310/310 [05:18<00:00,  1.03s/it, loss=0.00868]\n",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0694 | IoU: 0.6624 | dice: 0.7816 | dice_neg: nan | dice_pos: 0.7816\n",
      "Starting epoch: 18 | phase: val | ⏰: 13:15:55\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:48<00:00,  2.74it/s, loss=0.00889]\n",
      "  0%|                                                                                          | 0/310 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0711 | IoU: 0.6531 | dice: 0.7698 | dice_neg: nan | dice_pos: 0.7698\n",
      "Epoch    19: reducing learning rate of group 0 to 5.0000e-05.\n",
      "\n",
      "Learning Rate =  5e-05\n",
      "Starting epoch: 19 | phase: train | ⏰: 13:16:44\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 310/310 [26:05<00:00,  5.05s/it, loss=0.00787]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0630 | IoU: 0.6879 | dice: 0.8033 | dice_neg: nan | dice_pos: 0.8033\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r",
      "  0%|                                                                                          | 0/133 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting epoch: 19 | phase: val | ⏰: 13:42:50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████████████████████████████████████████████████████████████| 133/133 [00:18<00:00,  7.21it/s, loss=0.00806]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss: 0.0645 | IoU: 0.6776 | dice: 0.7942 | dice_neg: nan | dice_pos: 0.7942\n",
      "******** New optimal found, saving state ********\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model_trainer = Trainer(model, lr, epochs, path)\n",
    "print(f\"Current Device is GPU: {torch.cuda.current_device()}\")\n",
    "print(f\"Total memory allocated in GB: {round(torch.cuda.memory_allocated(0)/1024**3,1)}\")\n",
    "print(f\"Total memory cached in GB: {round(torch.cuda.memory_reserved(0)/1024**3,1)}\")\n",
    "model_trainer.start()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b1e3229",
   "metadata": {},
   "source": [
    "After the model training was done the loss, dice_scores and iou_scores were plotted. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "7c7d49f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3gAAAFNCAYAAABSRs15AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABii0lEQVR4nO3dd3hT5/3+8ffjPbGNMcsGs/fGrAxGBiETssmepEmTdGe0TZu0ya9Jm7bfNM0kq5mQnaZZZEHIgLD3NMN4gAfe25ae3x9HgAEbDFiWLd+v69Il6Zyjo48sZHT7WcZai4iIiIiIiLR+Ab4uQERERERERJqGAp6IiIiIiIifUMATERERERHxEwp4IiIiIiIifkIBT0RERERExE8o4ImIiIiIiPgJBTwREZEWzBiz0xhzhq/rEBGR1kEBT0REWpzGhhpjzGRjTEY92xcYY272TnUtU0M/CxERaVsU8ERERJqQMSbI1zWIiEjbpYAnIiItmjEmwBhznzEmzRiTY4x5xRgTcwLnO8cYs8EYU2KMyTTG/KbOvunGmFXGmGJjzDZjzDTP9q7GmA+NMfnGmFRjzKw6j3nAGPOOMeY1Y0wxcL0xJsYY84IxZrfnOR4yxgQ2UM++x7/pqWmFMWZ4A8eGGmMeM8ZkeS6PebZFAp8CXY0xpZ5L1+P9GYmISOulgCciIi3d9Z7LFKAXEAU8cQLnewH4ibU2GhgCfA1gjBkLvALcBcQCE4GdnsfMATKArsAlwF+MMafXOed04B3P414HXgZqgT7ASGAqcKQuo9OBt4H2wBvAB8aY4HqO+z0wHhgBDAfGAvdZa8uAs4Esa22U55LVmB+GiIj4FwU8ERFp6a4C/mmt3W6tLQV+C8w8ga6QNcAgY0w7a22BtXaFZ/tNwIvW2i+stW5rbaa1dpMxphtwCnCPtbbSWrsKeB64ps45F1lrP7DWuoF2OGHrF9baMmttDvB/wMwj1LTcWvuOtbYG+CcQhhPk6vtZ/Nlam2OtzQX+dEgdIiLSxingiYhIS9cVSKtzPw0IAjrhtJLV19IVjBPk6nMxcA6QZoz5xhgzwbO9G7CtgefPt9aWHFJDYp376XVuJ3uef7cxptAYUwg8C3RsoJ6DHu8JiftaC+ur5dCfhbpiiojIfhoILiIiLV0WTmjapztOsMsGAoEOxpgoT+sexhjjOT7t0BMBWGuXAtM9XSDvAN7CCXfpQO8Gnr+9MSa6TsjrDmTWPW2d2+lAFdDBWlvbyNfYbd8NY0wAkOR53vpqSQbW16lj33G2nuNFRKSNUQueiIi0dHOAXxpjehpjooC/AG9aa2uttbuAH4G/GmOijDGhOGPoaoHFh57IGBNijLnKGBPj6Q5ZDLg8u18AbjDGnO6Z2CXRGDPAWpsO/AA8bIwJM8YMw+nO+Xp9xVprdwOfA/8wxrTznKu3MWbSEV7jaGPMRZ5up7/ACYiH1e/5WdxnjEkwxnQA/gi85tmXDcSfyAQ0IiLS+ingiYhIS/ci8CqwENgBVAJ31tl/OU73x1ScVrXTgXOstZUNnO8aYKdnxstbgasBrLVLgBtwxssVAd9woOXwCqAHTmvZ+8D91tovjlDztUAIsAEowJmApcsRjv+v53UUeOq7yBNAD/UQsAxYA6wFVni2Ya3dhBMAt3u6hqrrpohIG2SsVY8OERERXzHGPAD0sdZe7etaRESk9VMLnoiIiIiIiJ9QwBMREREREfET6qIpIiIiIiLiJ9SCJyIiIiIi4icU8ERERERERPxEq1vovEOHDrZHjx6+LkNERERERMQnli9fnmetTahvX6sLeD169GDZsmW+LkNERERERMQnjDFpDe1TF00RERERERE/oYAnIiIiIiLiJ7wa8Iwx04wxm40xqcaYe+vZf5cxZpXnss4Y4zLGtPdmTSIiIiIiIv7Ka2PwjDGBwJPAmUAGsNQY86G1dsO+Y6y1jwKPeo4/H/iltTb/WJ+rpqaGjIwMKisrm6Z48aqwsDCSkpIIDg72dSkiIiIiIn7Fm5OsjAVSrbXbAYwxc4HpwIYGjr8CmHM8T5SRkUF0dDQ9evTAGHNcxUrzsNayd+9eMjIy6Nmzp6/LERERERHxK97sopkIpNe5n+HZdhhjTAQwDXj3eJ6osrKS+Ph4hbtWwBhDfHy8WltFRERERLzAmwGvvrRlGzj2fOD7hrpnGmNuMcYsM8Ysy83Nrf/JFO5aDb1XIiIiIiLe4c2AlwF0q3M/Cchq4NiZHKF7prV2trU2xVqbkpBQ73p+PlVYWMhTTz11XI8955xzKCwsbPTxDzzwAH//+9+PeMzkyZMPWitw586dDBky5LjqExERERGR1sObAW8p0NcY09MYE4IT4j489CBjTAwwCfivF2vxqiMFPJfLdcTHfvLJJ8TGxnqhquZXW1vr6xJERERERNo0rwU8a20tcAcwD9gIvGWtXW+MudUYc2udQy8EPrfWlnmrFm+799572bZtGyNGjOCuu+5iwYIFTJkyhSuvvJKhQ4cCMGPGDEaPHs3gwYOZPXv2/sf26NGDvLw8du7cycCBA5k1axaDBw9m6tSpVFRUHPF5V61axfjx4xk2bBgXXnghBQUFx1T37t27mThxIiNGjGDIkCF8++23AHz22WeMGjWK4cOHc/rppwOQn5/PjBkzGDZsGOPHj2fNmjWA06J4yy23MHXqVK699lpyc3O5+OKLGTNmDGPGjOH7778/pppERERERHyttKqWb7fm8u7yDF+Xcsy8OYsm1tpPgE8O2fbMIff/A/zHm3V42yOPPMK6detYtWoVAAsWLGDJkiWsW7du/0yRL774Iu3bt6eiooIxY8Zw8cUXEx8ff9B5tm7dypw5c3juuee47LLLePfdd7n66qsbfN5rr72Wf//730yaNIk//vGP/OlPf+Kxxx5rdN1vvPEGZ511Fr///e9xuVyUl5eTm5vLrFmzWLhwIT179iQ/3xkWef/99zNy5Eg++OADvv76a6699tr9r3f58uV89913hIeHc+WVV/LLX/6SU045hV27dnHWWWexcePGxv8wRURERESaWVZhBcvSCli2M59lOwvYtKcYt4XosCAuHJlIQEDrmUPCqwHPF/70v/VsyCpu0nMO6tqO+88ffEyPGTt27EHLADz++OO8//77AKSnp7N169bDAl7Pnj0ZMWIEAKNHj2bnzp0Nnr+oqIjCwkImTZoEwHXXXcell14K1D+JSX3bxowZw4033khNTQ0zZsxgxIgRLFiwgIkTJ+6vvX17Z9357777jnffdSY5Pe2009i7dy9FRUUAXHDBBYSHhwPw5ZdfsmHDgZUwiouLKSkpITo6usHXIiIiIiLSXFxuy8bdxSxPK2BZWgHLd+aTVeTM8B4REsiIbrHcMaUPo3u0Z2T32FYV7sAPA15LERkZuf/2ggUL+PLLL1m0aBERERFMnjy53mUCQkND998ODAw8ahfNhsTHxx/UXTM/P58OHTocdtzEiRNZuHAhH3/8Mddccw133XUXsbGx9YZBaw+fAHXfcXVfq9vtZtGiRfsDn4iIiIiIL5VW1bJqVyHL0vJZnlbAirQCyqqdeTI6tQslJbk9s3rEkZLcnoFdogkK9OY0Jd7ndwHvWFvamkJ0dDQlJSUN7i8qKiIuLo6IiAg2bdrE4sWLT/g5Y2JiiIuL49tvv+XUU0/l1Vdf3d+aN3nyZF577TXOOOMMjDG8/PLLTJky5bBzpKWlkZiYyKxZsygrK2PFihX8/ve/5/bbb2fHjh37u2i2b9+eiRMn8vrrr/OHP/yBBQsW0KFDB9q1a3fYOadOncoTTzzBXXfdBTjjBPe1SoqIiIiIeNvuogqW7fR0t0wrYONup7ulMdC/UzQXjkokJbk9o5PjSIoLP7xxozQH0pdAxlKoLILzH/PJ6zhefhfwfCE+Pp6TTz6ZIUOGcPbZZ3PuuecetH/atGk888wzDBs2jP79+zN+/Pgmed6XX36ZW2+9lfLycnr16sVLL70EwC233MKmTZsYPnw4xhhSUlJ4+OGHD3v8ggULePTRRwkODiYqKopXXnmFhIQEZs+ezUUXXYTb7aZjx4588cUXPPDAA9xwww0MGzaMiIgIXn755Xprevzxx7n99tsZNmwYtbW1TJw4kWeeeabeY0VEREREToTLbdm0x9PdcmcBy9MKyCx0esGFBzvdLW+f0ofRyXGMSo6jXVjwISeogex1kL4UMpY4wa4wzdkXEAxJKeB2Q0DradUz9XW9a8lSUlJs3TXeADZu3MjAgQN9VJEcD71nIiIiPlCeD5krIHMZZCyDrJUQ2w1OuhMGTodA/e1fWrayqlpWpRc6LXRp+azcVUhplbNUV8foUMb0cFrmUnrEMbBLO4IP7W5ZmnsgyGUsdT4PtZ5hUVGdodsYSBoL3cZCl+EQ3DKHHRljlltrU+rbp0+xiIiIiD+qrXZaJjKXO2EucxnsTfXsNJAwAPpOdb7svnMjxPVwgt6Iq1rsl1ppe/YUVbIsLX9/oNu4uwSX2+7vbjl9RFdSPOPnDutu6aqBrNUNt851GQajr4OkMU6gi+nm9ONs5RTwRERERFo7a6Fw14GWuYxlsHs1uKqc/ZEdnS+xI66ExBToOhLCPGPp3S7Y9DF8/xh8/GtY8AiM+wmMuRnC43z2kqTtcbktW7JL9o+dW7bzQHfLsOAARnSL5bZJvUnpEcfI7nHEhB/S3bIxrXNjbm7xrXMnSgFPREREpLWpLKrT1XK5c12W6+wLCoMuI2DsLEgc7YwhOlLLREAgDLoABp4PO79zgt7XD8F3j8Ho62H8TyEmsXlel7QpZVW1rE4vdMJcWgEr0wooqdPdMqVHHDee0pOU5DgGdT2ku6WrxulinL7UCXMZS6Bgp7MvIAg6+2frXGMo4ImIiIi0ZK5ayNlwcOtc3hbAM49CfF/oc4YT5BJToNNgCAw+4inrZQz0PNW57FkL3/8LFj8NPz4Lwy6Dk38OCf2b9KVJ21Fd62bznhJWZxSyOr2QNRlFbM0p2T+7Zb+O0Zw/oispyXGM6VFPd8tDW+eyVkJNubNvX+tcyk1OoOs6wm9b5xpDAU9ERESkpbAWijMPjJnLWA67Vx34IhsR74S4oZc4rXOJo7zTjbLzULj4eTjtD7DoCVjxKqx6HfqfAyf/ArqPa/rnFL/hdlt27C1jTUYhq9OLWJVeyIbdxVTXugFoHxnC8KQYpg3pzIjusYw6tLulq9bpYpyx1BPo6mmdG3Vtm2ydawwFPBERERFfqSp1WiLqts6V7nH2BYYc+CKbmOK00MX1aN4vsnHJcM6jMOkeWDLbuWz+BLpPcIJe36mtavp48Y49RZX7W+ZWZzitcyWVTlfL8OBAhibFcN2EZIZ3i2V4Umz9rXOb9rXOLYOsFXVa5zo5QS7lRmd2yzbeOtcYCng+EhUVRWlpaaO377NgwQL+/ve/89FHH+3fdv3113PeeedxySWXeKVWERERaQJuF+RuOrh1LncjWKdVg/a9oOfEA10tOw+BoFDf1rxPZAeY8junm+aKV2DRkzDnckgY6GwbesnxdQuVVqeovIY1mU6IW5VeyJqMQrKLncl8ggIMA7pEc/7wroxIimV4t1j6dIwiMMBAdZkzEVDuGtiaBgVpzoyW2evUOtfEFPCk0VwuF4GBgb4uQ0REpGVx1UB1qfMFtspzXV3iuV9yINRlrXSOAwiLdbpYDjzPCXOJoyEy3qcvo1FCImH8bc5MhOvec8bpfXCrMynLhNudL+ahUb6uUppIZY2L9VnFnq6WhazOKGJHXtn+/b06RHJS7w4MS4pheGIkgyOKCS1Jh8LlToBLqxPk9k0CtE9QuNNC3HmoWueamAJeE7jnnntITk7mpz/9KQAPPPAA0dHR/OQnP2H69OkUFBRQU1PDQw89xPTp0xt1Tmstd999N59++inGGO677z4uv/zyY6rr8ccf55lnniEoKIhBgwYxd+5cSktLufPOO1m2bBnGGO6//34uvvhi5syZw1/+8hestZx77rn89a9/BZwWxV/96lfMmzePf/zjH+zcuZPHH3+c6upqxo0bx1NPPaXQJyIirYfbDTVldcJY6YFwVl1aJ6CVHhLa6m4/ZNu+pQgaEhDkfIkdfsWB1rn43q27VSIwGIZf7ky+svULZ+bNeb+Fb/7qzN457lan1U9aDZfbsjWnZH+QW51eyOY9JdS6ncl8OkcHM7GLizt7lzM4opDkwFzCSjOcALckzRk7uq81Gpx/9zFJEJsM/c92ruN6eK6TITKhdX8GWjAFvCYwc+ZMfvGLX+wPeG+99RafffYZYWFhvP/++7Rr1468vDzGjx/PBRdccHCf4wa89957rFq1itWrV5OXl8eYMWOYOHHiMdX1yCOPsGPHDkJDQyksLATgwQcfJCYmhrVr1wJQUFBAVlYW99xzD8uXLycuLo6pU6fywQcfMGPGDMrKyhgyZAh//vOf2bhxI3/961/5/vvvCQ4O5qc//Smvv/4611577bH9wERERJpCZRFsmQdF6Z7wVV8Yqxvaypxw11iBoU6LVWgUhNS5RHWC0GhnX0gkhETXOS6yzrGe2zGJ/tsqYQz0m+pc0pc6QW/h3+GHf8PIq2HCHdC+p6+rlENYa8koqNjfxXJ1ehHrsgoJrS6km8mlX+hermlXwoCkfJJMLrFVWQSVZMCuathV50TRXZzAlnzSgeC27zq6KwQqaviC//3UP73Xmdq3KXUeCmc/0uDukSNHkpOTQ1ZWFrm5ucTFxdG9e3dqamr43e9+x8KFCwkICCAzM5Ps7Gw6d+581Kf87rvvuOKKKwgMDKRTp05MmjSJpUuXEhMTU+/x9YXGYcOGcdVVVzFjxgxmzJgBwJdffsncuXP3HxMXF8fChQuZPHkyCQkJAFx11VUsXLiQGTNmEBgYyMUXXwzAV199xfLlyxkzZgwAFRUVdOzY8aivRUREpMlUl8GWz5zugVs/B1e1s90E1glikQeuY7rVCWJ19h8U2jz79gc0z32NKTs23cbAzNchdwv88C9Y/jIsexEGX+hMyNJlmK8rbLPySqtYk1HIhp272ZO2mdLsbcRVZdHN5DIuIJergvfSJSiH0IDyAw8qBmrinMDWdSjEnucJcD2c65huEBzmq5ckR+B/Ac9HLrnkEt555x327NnDzJkzAXj99dfJzc1l+fLlBAcH06NHDyorKxt1Pmttvdvj4+MpKCg4aFt+fj4dOhzeDeLjjz9m4cKFfPjhhzz44IOsX78ea+1hYbCh5wIICwvb3wXTWst1113Hww8/3KjXICIi0iRqqyD1K1j3Lmz+1GmFi+rsjAMbfJHzh9igUHX3aikS+sH0J2HKfbD4KVj2kvPe9T7NCXo9J+q9OkZVtS7KqlyUVdVSWlVb5/qQbdXOdVmVi5qKEkYVfEZyyUria3cz3ORymik5cNJgcHvGwQXEDYDYsw5ugYtNhrB2vnvRctz8L+AdoaXNm2bOnMmsWbPIy8vjm2++AaCoqIiOHTsSHBzM/PnzSUtLa/T5Jk6cyLPPPst1111Hfn4+Cxcu5NFHHyUuLo6srCw2btzIwIEDSUtLY/Xq1YwYMeKgx7vdbtLT05kyZQqnnHIKb7zxBqWlpUydOpUnnniCxx57DHC6aI4bN46f//zn5OXlERcXx5w5c7jzzjsPq+n0009n+vTp/PKXv6Rjx47k5+dTUlJCcnLycf/cRERE6uWqhR3fOC11G/8HVUUQ3t4Z8zXkYqdLWIDGgLdo7brA1Afh1F87LXmLn4ZXLoCuI52gN/B8v30Pa11uJ3xVHxzG9gWxsuoDoazu9sO2eR5f42r4j/F1hQQG0CdkL9cEfs75tV8SRRl5QZ2pjE/G1WEs1V37ENKh5/6xcAGRHRS2/ZD/BTwfGTx4MCUlJSQmJtKlSxfA6ep4/vnnk5KSwogRIxgwYECjz3fhhReyaNEihg8fjjGGv/3tb/u7dr722mvccMMNVFZWEhwczPPPP39Y102Xy8XVV19NUVER1lp++ctfEhsby3333cftt9/OkCFDCAwM5P777+eiiy7i4YcfZsqUKVhrOeecc+qdDGbQoEE89NBDTJ06FbfbTXBwME8++aQCnoiINA23G9IXO6096z+A8jxnfNvA82DIJdBrkrpNtjLWWqqDo6kYfQflA28gaO2bxKx8mtC3r6M8ugepfW5ga+fzKHUFUV7toqK6lvJqF+U1LsqrnNv7Jvmoe87Dnqfe565nWwM1Nu611Hc+S3Wtm1JPS9q+kFZV6z784HoEBhgiQwKJCg0i0nOJCg0iITp0/+391yGB+29HhAYRFercjwwJIiokkKicZQQvfQY2fQQuA4Onw/if0qHbmEbVIv7DNPYfdUuRkpJily1bdtC2fa1Z0nroPRMREcD51py10hPq3ndm4gsKh/7TnJa6PmdqnE8zqHG5PQHLaXWqqHY5QWtf4Kobvups339cnUBWUXPw41yHBLQA3JwVsJRbg/7H8IDt5NoYXqw9m9ddp1NMJBEhgZ5LEBEhgQQHHr6Qen2NTvW2QzXQOlXf1sae89ChLsGB5rCAFhkSRGRo4MEBLfTwbaFBAY2afK9BtVXO52bxU7B7tbP8xujrnZlMY5KO/7zS4hljlltrU+rbpxY8ERERaX7ZG5xQt+5dKNgBAcHQ5ww4409OuAuN9nWFfqnW5WbTnhJW7CpgeVoBK3YVsKeostFdAPcJCw7YH8AiQgIJDwkiIjiQrrHB+2+He/ZFhgYRHrzvOCe4RYZMwB38C9LzltBh9VPcs2sud0d+BCk3Ysb/1OneKQ0rzXW6vS57AUqzoUN/OO//YNjlziRB0qYp4ImIiEjz2LsN1r/njKvL2QAmwJlw49RfO90ww+N8XaHfKSqvYUV6ASvSnEC3Kr2Q8moXAJ3ahZKS3J5zh0YQWSd81W1BC99/+8C+8OBAAgKaaNxW8jQYPQ12r8Z8/y9Y9IQzVm/45XDSz50JW+SAPWth8TOw9i1nBtk+ZzoLz/c+TWPpZD8FPBEREfGeokxPqHvX6YoJ0H0CnPN3GDQdorTcTlOx1rI9r8xpmfMEuq05pYAz1mtgl2guS+nGqOQ4RifH0TUm7MS6BzalLsPhkhfhtD84IW/la7DydRhwLqTcCMknt92uum6XszTI4qdh57cQHAEjr3EWk1cAlnr4TcCrb/p/aZla27hPERE5RqW5sOEDp6Vu1w/Oti4j4MwHnTXRYrv5sjq/UVHtYnVG4f5At2JXAQXlNQDEhAczqnss00d0ZVRyHMOTYokMbQVf+9r3hHP/AZPuhSXPwpLnnElDgiOgx6lON94+p0N8b19X6n2VxU7QXfIsFOyEdklw5p9h1LVq7ZYjagWf9KMLCwtj7969xMfHK+S1cNZa9u7dS1hYG/0rnIiIv6oodL6Ir3sXtn8D1gUJA5y10IZc1Da+kHtZVmEFy9MOjJ3bkFW8f4bJPh2jOHNQJ0Z7Wud6dYhqum6UvhCVAKfdB6f8ymm1Sv3SuWyd5+yP6+kJe2dAz1P9a9xZ/nb4cbYT7qpLoNt4OOMBGHA+BPrFV3fxMr+YRbOmpoaMjIxGLyIuvhUWFkZSUhLBwZrqWkQEV60zdqY1rgdWXeYsPL7uXefLt6vaWV9ryMXOpeMgjQs6TjUuNxuyip1At8tpodtd5HzPCQ8OZHi3mP1hbmS3OOIiQ3xccTPZuw22fe38e9uxEGrKITDE6fa7L/B1HNj6/t1Z6wTZxU87n6mAQBh8EYy/FRJH+7o6aYGONIumXwQ8ERGRVil9KcyZeWC9t7B2EBYDoZ7rsHYH3z5o3yHHBUc0z5famkrny/W6d51xQTXlEN3F+TI69GLoOqr1fbluAfLLqp1xc57ZLddkFFJZ46yllhgb7oyb6x7L6OT2DOgSXe/SAW1ObRXsWuRp3fvKmbgHILqr042zzxnQazKEx/qyyiOrqYS1b8OPz0D2OoiId8YcptykmUTliHwW8Iwx04B/AYHA89baR+o5ZjLwGBAM5FlrJx3pnAp4IiLiF3Z+B29cDpEJztTmVcXOmJvKQs/tIuf+vtvu2iOfLyDIE/iOEgTrDZCxzu2GFhF31cCOb5wxdRv/59QUEe9MkjLkEqf1JECBo7HcbktqbumB7pZpBWzPKwMgKMAwODGG0d2d1rlRybF0iQn3ccWtRFEmbPvKCXzbFkBVEZhA6Db2QODrPLxl/Fst2QNLX3CWOijPg46Dnda6oZdCsN5vOTqfBDxjTCCwBTgTyACWAldYazfUOSYW+AGYZq3dZYzpaK3NOdJ5FfBERKTVS/0K5l4Fsd3h2v8e/S/11kJNhRP0Dgp/RQduH7pv/33P7eqSo9cVHHF4SAwOd1pJyvc69wee74yp6zmp4UAoBymtqmV1euH+QLdyVwHFlU5gbx8ZwihPmBudHMewpBjCglthd92WxlULmcsOjN3bN4NrRIcDYa/3aRDZoXnrylrpdMNc957zR5t+05xlDnpOVMu3HBNfBbwJwAPW2rM8938LYK19uM4xPwW6Wmvva+x5FfBERKRV2/wpvHWtszDxNe87k0k0B7fr4MB32O2iA5eD9pVA5yFOS12fM9ruVPWN4HJbduWXsyW7hC17StiSU8qWPSVszSnBbZ3v7/06Ru9fpmB0chw94iM0QVxzKM09MHZv21fOHyww0HXEgbF7iSnemcTEVetMQLT4aUhfDCFRMPJqGHuLJh+S43akgOfNqXgSgfQ69zOAcYcc0w8INsYsAKKBf1lrX/FiTSIiIr6z/n1492boPAyufhci2jffcwcEOlOra3r1E+Z2WzILK9i8p4QtOSVszS5lS3YJqTmlVNW69x+XFBdOv07RnDW4E6N7tGdEt1hiwtXq6RNRCc7i6cMvB7cbdq9yWtJTv4Rv/wELH4XQGOg92dO6dzrEJJ7Yc1YUwIpXYclsKEqH2GQ46y9OuAuLaYpXJVIvbwa8+v4cdWhzYRAwGjgdCAcWGWMWW2u3HHQiY24BbgHo3r27F0oVERHxstVz4YPboNs4uPItpxuktGjWWnYXVbI5u4St2SVsqRPkyqtd+4/rEhNG307RnNQ7nr6dounXKZq+HaNax7pzbVFAACSOci6T7nKC2PZvDkzWsuG/znEdBx3oztl9AgSFNu78eVudSVNWveFMQpR8Ckx7BPqf3Tpny5VWx5u/eTKAuiuZJgFZ9RyTZ60tA8qMMQuB4Thj9/az1s4GZoPTRdNrFYuIiHjDspfgo18663VdMde/1uzyA9ZackqqnK6V2aVszS5hc3YJqdmllFQdmNwmITqUfp2iuHxMN/p1iqZfpyj6dIxWq1xrFx4Hg2c4F2shZ+OBsXs/Pgs//NsZn9pz4oGF1tv3Ovgc1jpdQBc/DalfOEs3DL0Uxt0KXYb54lVJG+bNMXhBOEHtdCATZ5KVK6216+scMxB4AjgLCAGWADOttesaOq/G4ImISKuy+Gn47F7oOxUue0Uz5PlYXqkT5LZmlx7UMldUUbP/mPaRIfTrFOW0xHWKpr8nzMVGtJG15uSAqlJnxtvUL53gVrDT2d6+F/Q50wl7xVlOi13uJojsCGNucpY6iOro09LFv/lkDJ61ttYYcwcwD2eZhBetteuNMbd69j9jrd1ojPkMWAO4cZZSaDDciYiItCrf/hO++pMz8+TFL0KQAkJzKSyvZstBIc4Jcvll1fuPiQkPpl+nKM4d1oV+HaPo19npXtkhqpFd8cT/hUZB/2nOBZyF1veN3VvxCix51tneeRjMeMaZYbaxXTlFvEQLnYuIiDQ1a2H+X2Dh35xuWjOe8c7sfG2M222pcbtxuS21bovLZalxuUkvqPAEuAMtc7klVfsfFxUaRN9OUfTrGO0JcU7rXMfoUM1gKcevptKZFTMo3FlrT/+WpBn5ahZNERGRtsda+OIPzridkVfD+Y+3uokVKmtcZBRUkFlYQWZBBeXVtU6gcjuBan/AOvS+6/AAVut27z+2tpH3Gzrf0f4mHR4cSN9OUUzsm0D/zlH7JzzpGhOmICdNLzgMek32dRUih1HAExERaSpuN3x6Fyx9HsbMgrP/5szY18KUVdWSWVhBRkE5mQUVZBRUkFHoXGcWlJNXWn3UcwQHGoICAggKMAQGGoICnPuBAYagQENggCG4nvtBAQGEBTv363v8YfcD9x1b331DYGAAXdqF0a9TNElx4QQEKMiJSNumgCciItIU3C74389g5Wtw0p1w5oM+67JVXFmzP7hlFpTvb43LKHBCXUF5zUHHhwQFkBgbTlJcOIMGdfLcjiAxLpzE2HCiwoIOCmAKUSIiLZcCnoiIyIly1cD7t8K6d2DSPTD5t14Ld9ZaiitqST8ouHluewJccWXtQY8JCw5wAltsOMOSYkiMcwJcUlw4SbHhdIgKVWgTEfETCngiIiInorYa3rkBNn0Ep98Pp/7qhE5nrSW/rHp/i9u+0Fa3Fa606uAAFxkSuL/FLaVH3P4WuKS4cBLjwomPDNEYNBGRNkIBT0RE5HjVVMBb18LWz2HaX2H8rcf08NKqWt5fkcHm7JI6Ya6CihrXQcdFhwV5AlsE43vFOy1vnla4xNhwYiOCFeBERARQwBMRETk+1WUwZybs+BbOewxSbmj0QwvLq/nPDzt56fudFFXUEBsRTFJcOL0SIpnYL8FpeaszDi4mPNh7r0NERPyKAp6IiMixqiyG1y+FjCVw4TMwfGajHpZbUsUL3+3g1UU7Kat2ceagTtwxpQ/Du8V6t14REWkzFPBERESORXk+vHYx7FkDl7wIgy886kN2F1Xw7DfbmbNkFzUuN+cO68rtU3ozoHO7ZihYRETaEgU8ERGRxirNhVdnQN4WuPw16H/2EQ9P21vG0wu28e6KDKyFi0YlctvkPvTsENk89YqISJujgCciItIYxbvhlelQuAuufBN6n9bgoVuyS3hqfiofrs4iKDCAK8Z255aJvUiKi2jGgkVEpC1SwBMRETmawl3w8gVQlgtXvwM9Tqn3sHWZRTzxdSqfrd9DREggN5/ai5tP6UnHdmHNXLCIiLRVCngiIiJHkr/dCXeVxXDNB9BtzGGHLN2ZzxNfp/LNllzahQXxs9P7csNJPYiLDGn+ekVEpE1TwBMREWlI7mYn3Lmq4boPoeuI/bustXyXmscTX6fy44584iNDuHtaf64Zn0x0mJY1EBER31DAExERqc+etfDKDDABcP3H0GkQAG635atNOTwxP5XV6YV0ahfKH88bxBVjuxMeEujbmkVEpM1TwBMRETlU5gp49UIIiYRrP4QOfXC5LR+v3c1T81PZtKeEbu3D+cuFQ7l4dCKhQQp2IiLSMijgiYjI8autgl2LIL4vxCT6upqmsWuxs4h5eCxc9z9q2nXn/WXpPLNgG9vzyujTMYr/u3w45w/rSlBggK+rFREROYgCnoiIHLs9a2HFq7D2LagocLZ1Ggr9znIuiaMhoBW2am3/BubMhHZdqbzyfd7e5OaZbxaQWVjBoC7tePqqUZw1uDMBAcbXlYqIiNRLAU9ERBqnohDWvg0rX4PdqyAwBAacB0MuhvxtsGUefPd/8O3fISIe+pzphL3epzmtYS3d1i/gzatxx/bg9f6P8/gzW8gtqWJU91gemjGEyf0TMEbBTkREWjZjrfV1DcckJSXFLlu2zNdliIi0DW437PwWVr4KG/8HtZVOS92oa2DopRDR/uDjKwog9Ssn7KV+4dw3gZB8EvSdCv2mQYe+0NKC0sb/Yd++gdzwXlxWfjc7K8I5uU88t0/pw4Re8Qp2IiLSohhjlltrU+rdp4AnIiKHKcqAVW84rXWFaRAaA8MuhZHXQJfhjQtobhdkLHXC3pZ5kLPe2R7X80BXzuSTISjUu6/lKEqWzSXyo5+y1vbimqq7GTOgJ7ef1odR3eN8WpeIiEhDFPBEROToaqtg8yfO2LptXwMWek6EkdfCwPMgOPzEzl+YDls9YW/HQqc1MCQKek12Wvb6ToXoTk3xShplT1ElP773OOfv/AtL7ADe6vd3bj5tGIO6tmu2GkRERI6HAp6IiDQse70T6ta8CRX50C4RRlwFI6+CuB7eec7qcifk7Qt8xZnO9q4joa+nda/LCAho+lkqd+0t5+lvthG68kUeCHyRzZFjCLpyDr0TE5r8uURERLxBAU9ERA5WUQjr3nXG1mWthIBgGHCuM7au15TmnQHTWidkbvnMCXsZSwELUZ2g75lO616vyRAafUJPk5pTwlPzt/Hf1VncHPgJvw18lYqeZxF+1as+7yYqIiJyLBTwRETEmTAl7TtnXN2G/zpdJDsO9kyYchlExvu6QkdZHqR+6Zmo5SuoKnICaI9TDozda9+r0adbl1nEUwtS+XTdHsKCAnmq21dMyZoNg2bAxc9DYLD3XouIiIgXKOCJiLRlRZmw2jNhSsFOCG0HQy9xJkzpOrLlzWhZl6sG0n880LqXt8XZHt/XE/amQffxh4W07OJKPlyVxXsrM9m4u5jo0CCum5DMT+0cIn58DIZfARc8AYFaLUhERFofBTwRkbamttqZMGXla7DtK7Bu6HGqE+oGng8hEb6u8Pjkb4ctnztj93Z+B65qZ4bPPqdR1fMMPq8exlsbK/g+NQ+3heHdYrlwRFcuHJlIzML7YfFTMPoGOPefXhnfJyIi0hwU8ERE2orsDU6oWzMXyvdCdFdnspQRVx5Tt8ZWoaoUV+p8cpb/l4i0r4hx5eO2hg2BfSlMOo0eEy4iacBYZ4zfx7+C5S/BuNtg2sMtu9VSRETkKI4U8LzaN8UYMw34FxAIPG+tfeSQ/ZOB/wI7PJves9b+2Zs1iYj4ncoiz4Qpr0Hmcs+EKec4rXW9T2veCVOagbWWdZnFvL8ykw9Xh5NXeiGxYZdwc98SLoxcx+CchZhdz8CuZ5wZQWO6QfpiOOVXcPofFe5ERMSveS3gGWMCgSeBM4EMYKkx5kNr7YZDDv3WWnuet+oQEfFL1kLa987yBhv+C7UV0HEQnPUwDLsMIjv4usIml1FQzn9XZfH+ykxSc0oJCQzgtAEdmTEykSkDEggNCgQudQ4uyYbUL5yxexnLnGB36q99Wr+IiEhz8GYL3lgg1Vq7HcAYMxeYDhwa8EREpLGKs2DVG7DqdWc8Wmg7GD7TmQmz6yi/a50qqqjh07W7eX9lJj/uyAdgTI84/nLhUM4d2oWYiAZmwIzuBCOvdi4iIiJtiDcDXiKQXud+BjCunuMmGGNWA1nAb6y1671Yk4hI61Nb7bRErXzVWT7AuiH5FJh0Dwy8oPVOmNKA6lo3Czbn8MGqTL7cmEN1rZteHSL59Zn9mDEykW7t/ev1ioiINCVvBrz6/ox86IwuK4Bka22pMeYc4AOg72EnMuYW4BaA7t27N3GZIiIt2IYP4eNfQ1kORHeBU34JI66C+N6+rqxJWWtZsauQD1Zm8tGaLArKa4iPDOHKsd25cGQiw5JiMH7WOikiIuIN3gx4GUC3OveTcFrp9rPWFte5/Ykx5iljTAdrbd4hx80GZoMzi6b3ShYRaSEqi+Gze52umF1GwPQnnQlT/Gzdtp15Zby/MpMPVmWStrec0KAApg7uzEUjEzmlbweCA7WUgYiIyLHw5jeFpUBfY0xPIBOYCVxZ9wBjTGcg21prjTFjgQBgrxdrEhFp+XYthvdugaJ0OPU3MPnewxbybs3yy6r5eI2zCPnKXYUYAxN6xXPHlD5MG9KZ6DD/ea0iIiLNzWsBz1pba4y5A5iHs0zCi9ba9caYWz37nwEuAW4zxtQCFcBM29oW5hMRaSquGvjmr/DtP5yp/W/4FLqP93VVTaKyxsVXG3N4f2UmCzbnUOu29O8Uzb1nD2D6iK50iQn3dYkiIiJ+QQudi4i0BHmp8N4syFrhjLGb9giEtfN1VSfE7bYs2ZnP+ysy+WTdbkoqa+nULpTpIxKZMSKRQV1b9+sTERHxFZ8tdC4iIkdhLSx/Ceb9HoJC4dKXYfAMX1d1QrZml/D+ykz+uyqLzMIKIkICmTakMxeNTGJC73gCAzRZioiIiLco4ImI+EppLnx4h7MEQq8pMOMpaNfV11Udl5ySSj5clcUHqzJZl1lMYIDh1L4duHtaf84c1ImIEP13IyIi0hz0P66IiC9s/swJd5XFTnfMsT+BgNY3Y+Te0irueXcNX2/KwW1haGIMfzxvEOcP70pCdKivyxMREWlzFPBERJpTdZnTHXP5S9BpKFz3P+g40NdVHZe80iqueu5Hdu4t47bJvblwZCJ9Okb7uiwREZE2TQFPRKS5ZC6Hd2dB/nY46Wdw2n3OuLtWKLekiiufW0x6QTkvXT+Gk/p08HVJIiIiggKeiIj3uWrhu/+DBQ9DdBe47kPoOdHXVR23nOJKrnhuMVmFlbx0/Vgm9I73dUkiIiLioYAnIuJN+Tvg/Z9A+o8w5BI49x8QHuvrqo5bdnElV8xezJ7iSv5zwxjG9VK4ExERaUkU8EREvMFaWPU6fHoPmEC46HkYdqmvqzohe4qclruc4kpevnEsY3q093VJIiIicggFPBGRplaeD//7GWz8HySfAhc+A7HdfF3VCdldVMEVsxeTV1rNKzeNZXSywp2IiEhLpIAnItKUUr+CD34K5XvhzD/DhDsgINDXVZ2QzEIn3BWUOeFuVPc4X5ckIiIiDVDAExFpCjUV8OUD8OMzkDAArnobugzzdVUnLKOgnCueW0xheQ2v3jyOEd1ifV2SiIiIHIECnojIidq9Bt6bBbmbYNxtcMb9EBzu66pOWHp+OTNnL6aksobXbx7HsKRYX5ckIiIiR6GAJyJyvNwuWPQEfPUgRMTD1e9Bn9N9XVWT2LXXabkrrarljVnjGZIY4+uSREREpBEU8EREjkdhOrx/K6R9BwPPh/Mfhwj/mHhkZ14ZVzy3mIoaF6/fPE7hTkREpBVRwBMROVZr3oaPfw3WBdOfghFXgjG+rqpJ7Mgr44rZi6mqdfHGzeMZ1LWdr0sSERGRY6CAJyLSWBUF8PFvYN070G0cXPgstO/p66qazLbcUq6YvZhat+WNWeMZ2EXhTkREpLVRwBMRaYwdC+H926B0D5x2H5z8Swj0n1+hqTklXPHcj7jdljmzxtO/c7SvSxIREZHj4D/fTkREvKG2Cr5+CH74N8T3hps+h8TRvq6qSW3NdsIdwNxbxtO3k8KdiIhIa6WAJyLSkOwN8N4tkL0WUm6EqQ9BSKSvq2pSm/eUcOVziwkIMMyZNZ4+HaN8XZKIiIicAAU8EZFDud2w5Fn44n4IawdXvgX9zvJ1VU1u055irnzuR4ICDHNuGU/vBIU7ERGR1k4BT0SkruLd8MFtsH0+9DsbLvg3RCX4uqomtyGrmKueX0xoUCBzbhlPzw7+1TIpIiLSVingiYjss/4D+OgXzri78x6D0df7zfIHda3LLOLqF34kPDiQObPG00PhTkRExG8o4ImIVBbDZ/fCqteh6yi46Dno0MfXVXnF2gwn3EWFBjFn1ni6x0f4uiQRERFpQgp4ItK2FWfBi9OgKB0m3QMT74LAYF9X5RWr0wu55oUfiQ4LZu4t4+nWXuFORETE3yjgiUjbNu/3UJoNN3wK3cf7uhqvWbmrgGtfXEJsRDBzZo0nKU7hTkRExB8F+LoAERGf2bEQ1r8Hp/zSr8Pd8rQCrn1hCXERIcy9ZYLCnYiIiB9TC56ItE2uGvjkLohNhpN/7utqvGZ5Wj7XvbiU+KgQ5swaT9fYcF+XJCIiIl7k1RY8Y8w0Y8xmY0yqMebeIxw3xhjjMsZc4s16RET2WzIbcjfBtEcg2D9Dz9Kd+Vz7whISokN585YJCnciIiJtwFEDnnFcbYz5o+d+d2PM2EY8LhB4EjgbGARcYYwZ1MBxfwXmHWvxIiLHpWQPzH8Y+pwJ/c/2dTVe8eP2vVz34hI6xYQx95bxdI4J83VJIiIi0gwa04L3FDABuMJzvwQnuB3NWCDVWrvdWlsNzAWm13PcncC7QE4jzikicuK+uB9cVXD2X/1ynbtF2/Zy/UtL6RITxtxZ4+nUTuFORESkrWhMwBtnrb0dqASw1hYAIY14XCKQXud+hmfbfsaYROBC4JlGVSsicqLSFsGauXDSnRDf29fVNLkfUvO44T9LSIoLZ+4tE+iocCciItKmNCbg1Xi6UVoAY0wC4G7E4+r7s7g95P5jwD3WWtcRT2TMLcaYZcaYZbm5uY14ahGRerhq4ZPfQLskOPXXvq6myX23NY8b/rOU5PaRzLllPAnRob4uSURERJpZY2bRfBx4H+hojPl/wCXAfY14XAbQrc79JCDrkGNSgLnG6SLVATjHGFNrrf2g7kHW2tnAbICUlJRDQ6KISOMsfwmy18GlL0NIpK+raVILt+Qy65Vl9OwQyes3jyM+SuFORESkLTpiwDPGBAA7gLuB03Fa5WZYazc24txLgb7GmJ5AJjATuLLuAdbannWe6z/AR4eGOxGRJlGWB18/CD0nwaD6hgO3Xgs253DLq8vpkxDFazePo31kY3rRi4iIiD86YsCz1rqNMf+w1k4ANh3Lia21tcaYO3BmxwwEXrTWrjfG3OrZr3F3ItJ8vnwAqsvgnEf9amKVrzdlc+urK+jbKYrXbx5HbITCnYiISFvWmC6anxtjLgbes9YeU/dIa+0nwCeHbKs32Flrrz+Wc4uINFrGclj5qjOxSkJ/X1fTZL7ckM1try9nQOd2vHbTOGIign1dkoiIiPhYYwLer4BIwGWMqfRss9badt4rS0Skibhd8MmvIaozTLrH19U0mc/X7+H2N1YwqEs7XrlpHDHhCnciIiLSiIBnrY1ujkJERLxi5auQtRIueh5C/ePX2WfrdnPHGysZkhjDKzeNpV2Ywp2IiIg4GtOChzHmAmCi5+4Ca+1H3itJRKSJlOfDl3+C5JNh6CW+rqZJfLJ2N3fOWcmwpBhevlHhTkRERA521HXwjDGPAD8HNnguP/dsExFp2b5+CCqL4Oy/+cXEKh+tyeLOOSsZ0S2WVxTuREREpB6NacE7BxhhrXUDGGNeBlYC93qzMBGRE5K1Cpa9CON+Ap2H+LqaE/bh6ix++eYqRnWP5aUbxhIV2qgOGCIiItLGHLUFzyO2zu0YL9QhItJ03G745C6I7ACTf+vrak7YBysz+cXclYxOjuM/CnciIiJyBI35lvAwsNIYMx9nofOJQOv/xiQi/mvNXMhYAtOfgvBYX1dzQt5ZnsHd76xmXM94Xrg+hYgQhTsRERFpWGNm0ZxjjFkAjMEJePdYa/d4uzARkeNSWQRf/BGSxsDwK3xdzQl548dd/P6DtZzcuwPPXZtCeEigr0sSERGRFq4xk6xcCJRbaz+01v4XqDTGzPB6ZSIix2P+w1CWB+f8HQIa2wu95Xn5h5387v21TO6XwPPXKdyJiIhI4zTm28/91tqifXestYXA/V6rSETkeGWvhyWzIeUG6DrC19Uct+cWbuf+D9dz5qBOPHPNaMKCFe5ERESkcRozmKO+EKhBICLSsljrTKwS1g5O+4OvqzluT85P5dF5mzl3aBcemzmC4MDW2wopIiIiza8x3xyWGWP+aYzpbYzpZYz5P2C5twsTETkm696FtO/h9Pshor2vqzlm1lr++cUWHp23mQtHJvIvhTsRERE5Do359nAnUA28CbwNVAK3e7MoEZFjUlUCn98HXUbAqGt9Xc0xs9by18828/hXW7ksJYm/XzqcIIU7EREROQ6NmUWzDM+i5saYQCDSs01EpGX45m9QshsuexUCWtd4NWstf/5oAy99v5Orx3fnzxcMISDA+LosERERaaUaM4vmG8aYdsaYSGA9sNkYc5f3SxMRaYTcLbD4KRh5NXQb4+tqjonbbfnDf9fx0vc7ueHkHjw4XeFORERETkxj+gANstYWAzOAT4DuwDXeLEpEpFGshU/vgpBIOP0BX1dzTFxuy2/fW8tri3dx66Te/PG8QRijcCciIiInpjEBL9gYE4wT8P5rra0BrFerEhFpjI0fwvYFMOU+iErwdTWNVuty85u3V/PmsnR+dnpf7pnWX+FOREREmkRjAt6zwE4gElhojEkGir1ZlIjIUVWXw2e/g05DIOVGX1fTaDUuNz9/cxXvr8zkN1P78asz+ynciYiISJNpzCQrjwOP77tvjNkFTPFmUSIiR/XtP6A4Ay5+DgJbx9KcVbUu7nxjJZ9vyOb35wxk1sRevi5JRERE/Mwxfyuy1lqg1gu1iIg0zt5t8MPjMOxySD7J19U0SmWNi9teW878zbn86YLBXHdSD1+XJCIiIn6odfzZW0Skrs9+C4EhcOaffV1Jo1RUu5j1yjK+35bHXy4cypXjuvu6JBEREfFTCngi0rps/hS2zoOpD0F0Z19Xc1RlVbXc9PJSluzI59FLhnPJ6CRflyQiIiJ+rMGAZ4y56JBNFsgDVllrS7xalYhIfWoq4dN7oEN/GHerr6s5quLKGm54aSmr0gv5v8tHMH1Eoq9LEhERET93pBa88+vZ1h4YZoy5yVr7tZdqEhGp3w+PQ2EaXPshBAb7upojKiqv4doXf2R9VjFPXDGSs4d28XVJIiIi0gY0GPCstTfUt92zTMJbwDhvFSUicpiCNGfmzMEXQq9Jvq7miPLLqrnmhR/Zml3KM1eP5oxBnXxdkoiIiLQRxzOLZppn4XMRkeYz73dgApyxdy1YbkkVVz//Izv3ljH72tFM7t/R1yWJiIhIG3LMAc8Y0x+o8kItIiL1S/0SNn0Ep/8RYlruJCXZxZVc+dxisgorefH6MZzcp4OvSxIREZE25kiTrPwPZ2KVutoDXYCrG3NyY8w04F9AIPC8tfaRQ/ZPBx4E3Dhr6/3CWvtdo6sXEf9XWwWf3A3te8OEO3xdTYOyCiu48rnF5JZU8fKNYxnbs72vSxIREZE26EgteH8/5L4F9gJbrbXVRzuxMSYQeBI4E8gAlhpjPrTWbqhz2FfAh9Zaa4wZhjO2b8CxvAAR8XOLn4L8bXDVuxAU6utq6pWeX84Vzy2mqKKGV28ex6jucb4uSURERNqoI02y8s2+28aYTsAYoB2QC+Q04txjgVRr7XbPOeYC04H9Ac9aW1rn+EgObzEUkbasKBO+eRQGnAd9z/B1NfXakVfGVc8tpqzaxRs3j2doUoyvSxIREZE2LOBoBxhjLgOWAJcClwE/GmMuacS5E4H0OvczPNsOPf+FxphNwMfAjY0pWkTaiM/vA+uCs/6fryupV2pOCZc/u4jKWjdzZinciYiIiO81ZpKV3wNjrLU5AMaYBOBL4J2jPM7Us+2wFjpr7fvA+8aYiTjj8Q77M70x5hbgFoDu3bs3omQRafV2LIT178Hk30JcD19Xc5hNe4q5+vkfAcPcW8bTr1O0r0sSEREROXoLHhCwL9x57G3k4zKAbnXuJwFZDR1srV0I9DbGHDbtnLV2trU2xVqbkpCQ0IinFpFWzVUDn9wFsclw8s99Xc1h1mUWccXsxQQGGN78icKdiIiItByNacH7zBgzD5jjuX858EkjHrcU6GuM6QlkAjOBK+seYIzpA2zzTLIyCgjBCZAi0pYtmQ25m2DmHAgO93U1B1mVXsi1L/xIdFgwb8waR3J8pK9LEhEREdnvqAHPWnuXMeZi4GScbpezPd0qj/a4WmPMHcA8nGUSXrTWrjfG3OrZ/wxwMXCtMaYGqAAut9ZqohWRtqxkD8x/GPqcCf3P9nU1B1m2M5/rX1pK+8gQ3pg1jqS4CF+XJCIiInIQ09ryVEpKil22bJmvyxARb3nvJ87Yu58uhvjevq5mv0Xb9nLTy0vp3C6M12eNo0tMy2pZFBERkbbDGLPcWptS374jLXReQv3LFhjAWmvbNVF9IiKOtEWwZi6c+usWFe6+3ZrLrFeW0S0ugtdnjaNjdJivSxIRERGp15HWwdOsASLSfFy1zsQq7ZKcgNdCzN+Uw09eW06vDpG8fvM44qNa5mLrIiIiItC4SVZERLxv+UuQvRYufRlCWsbEJfPW7+GON1YwoHM7XrlxLHGRIb4uSUREROSIFPBExPfK8uDrB6HnJBg03dfVAPDRmix+MXcVQ5Ni+M8NY4kJD/Z1SSIiIiJHpYAnIr735QNQXQbnPArG+Loa3l+Zwa/fWk1KcntevGEMUaH6VSkiIiKtQ2MWLBcR8Z6M5bDyVRh/GyT093U1vLU0nV+9tZrxveL5z40KdyIiItK6KOCJiO+43fDJryGqM0y6x9fV8OriNO5+dw0T+ybw4vVjiAhRuBMREZHWRd9eRMR3Vr4CWSvhouch1LcT977w3Q4e/GgDZwzsyJNXjSI0KNCn9YiIiIgcDwU8EfGN8nz48k+QfDIMvcRnZVhreeab7fz1s02cPaQz/5o5kpAgdW4QERGR1kkBT0R84+uHoLIIzv6bTyZWsdby1cYcnpifyqr0Qi4Y3pV/XjacoECFOxEREWm9FPBEpPllrYJlL8K4n0DnIc361C635dN1u3ly/jY27i4mKS6ch2YM4Yqx3QkM8P0MniIiIiInQgFPRJqX2w2f3AWRHWDyb5vtaWtcbj5clcWTC1LZnltGr4RI/n7pcKaP6EqwWu1ERETETyjgiUjzWjMXMpbA9KcgPNbrT1dV6+Kd5Rk8vWAbGQUVDOgczRNXjuTsIV3UYiciIiJ+RwFPRJpPZRF88UdIGgPDr/DqU5VX1zJnSTqzF24ju7iK4d1ieeD8wZw+sCOmBSymLiIiIuINCngi0nwWPAJleXDVOxDgnW6RxZU1vLoojRe+20F+WTXjerbnH5eO4OQ+8Qp2IiIi4vcU8ESkeWSvhx+fhZQboOuIJj99QVk1L32/g5d+2ElJZS2T+iVwx2l9GNOjfZM/l4iIiEhLpYAnIt5nLXxyN4S1g9P+0KSnzimp5Plvd/Da4jTKq12cNbgTd0zpy9CkmCZ9HhEREZHWQAFPRLxv3buQ9h2c9xhENE2LWmZhBc9+s425S9Opdbk5f3hXfjq5D/07RzfJ+UVERERaIwU8EfGOmkrYmwq5m+Dz+6DLCBh17QmfdkdeGU8vSOW9FZkYAxeNTOK2yb3p0SHyxGsWERERaeUU8ETkxNRUQN5WyN0MuRud65yNULADrNs5JiQaLn8dAgKP+2k27ynhyfmpfLQmi+DAAK4a151bJvUmMTa8iV6IiIiISOungCcijVNdDnlbPEFu04FLwc4DQc4EQnxv6DQIhlwMCf2h40CI7wNBocf1tGsyCnni61Q+35BNZEggs07txU2n9qRjdFjTvTYRERERP6GAJyIHqy5zglzOpkOCXBpgnWMCgpzQ1nkoDL3MCXIJAzxBLqRJyli6M59/f53Kwi25tAsL4men9+WGk3oQF9k05xcRERHxRwp4Im1VVSnkba4T5DxdLAt3HTgmIBg69IWuI52FyRMGOJf2vZosyNVlreW71Dye+DqVH3fkEx8Zwt3T+nPN+GSiw4Kb/PlERERE/I0Cnoi/qyo50K0yxzNGLnczFNUJcoEhEN8XksbAyGs8LXIDoX1PCPR+sLLW8uXGHJ6Yn8rq9EI6twvjj+cN4oqx3QkPOf5xeyIiIiJtjQKeiL+oLILcLQdPdJK7GYozDhwTGAod+kG3sTD6Wk+L3ECI6wGBzf/rwOW2fLJ2N0/OT2XTnhK6tQ/nLxcO5eLRiYQGKdiJiIiIHCsFPJHWylrIWgmr58CmTw4OckFhTpBLPgk6DjjQtTKuxwnNZNlUalxuPliZydMLtrE9r4zeCZH887LhXDC8K0GBAb4uT0RERKTVUsATaW2KMmHtW7BqjjOGLjAU+k2FMTc5M1Ym9IfY5BYR5A5VWePi7eUZPLNgG5mFFQzs0o6nrhrFWYM7ExhgfF2eiIiISKungCfSGlSXwcaPnNa67QsAC93Gw3mPweALITzWt/UdRXl1LW/8uIvZC7eTU1LFyO6xPDhjMFP6d8QYBTsRERGRpuLVgGeMmQb8CwgEnrfWPnLI/quAezx3S4HbrLWrvVmTSKvhdkPa906o2/BfqC6F2O4w6W4Ydrmz3lwLV1RRw6uLdvLi9zvJL6tmQq94Hrt8BBN6xyvYiYiIiHiB1wKeMSYQeBI4E8gAlhpjPrTWbqhz2A5gkrW2wBhzNjAbGOetmkRahb3bnFC3+k1npsuQaBg8A4ZfCd0nQEDLH6OWW1LFC9/t4LXFaZRW1TKlfwJ3nNaH0cntfV2aiIiIiF/zZgveWCDVWrsdwBgzF5gO7A941tof6hy/GEjyYj0iLVdFAax7D1bPhYwlYAKg1xQ4/Y8w4FwIifB1hY2SUVDO7IXbeXNpOjUuN+cM7cJtk3szuGuMr0sTERERaRO8GfASgfQ69zM4cuvcTcCn9e0wxtwC3ALQvXv3pqpPxLdcNZD6Fax+AzZ/Cq5qZ8mCM/8MQy+Fdl19XWGjpeaU8PSC7fx3VSbGwMWjkvjJpN707BDp69JERERE2hRvBrz6BtjYeg80ZgpOwDulvv3W2tk43TdJSUmp9xwirYK1sGet0wVz7dtQlgsR8ZByIwy/AroMh1Y0Nm1NRiFPzd/GvA17CAsK5NoJPZg1sSddYsJ9XZqIiIhIm+TNgJcBdKtzPwnIOvQgY8ww4HngbGvtXi/WI+I7JXucQLdqDuSsh8AQ6DfNCXV9z4TAYF9X2GjWWn7ckc+T81P5dmse7cKCuHNKH64/uSftI0N8XZ6IiIhIm+bNgLcU6GuM6QlkAjOBK+seYIzpDrwHXGOt3eLFWkSaX00FbPrYGVe37SuwbkhMgXP/AYMvgojWNeGItZavN+Xw5PxUVuwqpENUKPeePYCrxnUnOqz1BFQRERERf+a1gGetrTXG3AHMw1km4UVr7XpjzK2e/c8AfwTigac8U6bXWmtTvFWTiNdZC7sWO+Pq1n8AVcXQLglO+SUMmwkJ/Xxd4TFzuS0fr93NU/NT2bSnhMTYcB6cPphLU7oRFtzyFlMXERERacuMta1rSFtKSopdtmyZr8sQOVj+DljzpjO2rmAnBEfCoAucLpg9Tm0VSxscqqrWxXsrMnn2m23s3FtOn45R/HRyb84f3pXgwNb3ekRERET8hTFmeUMNY15d6FzEr1UWOQuQr5oDu34ADPScCJPuhYHnQ2iUrys8LmVVtcxZsovnvt1OdnEVw5JiePaa0Zw5sBMBAa1nAhgRERGRtkgBT+RYuGph+wKnpW7TR1BbCfF9nfXqhl4Gsd2OeoqWqrC8mpd/SOOlH3ZQWF7DhF7x/OPSEZzcJx7Timb2FBEREWnLFPBEGiN7gzOubs3bULoHwmJh5NUw/EpIHNWqljY4VE5xJS98t4PXFqdRVu3ijIEduW1yH0Ynx/m6NBERERE5Rgp4Ig0pSIN17zqX7HUQEAR9z4LhM6HfWRAU6usKT0h6fjnPLtzGW8syqHW5OX94V26b3JsBndv5ujQREREROU4KeCJ1lebChg+cNevSf3S2JY2Fs/8GQy6GyA4+La8pbMku4ekF2/hwdRaBxnDx6CRundSL5PhIX5cmIiIiIidIAU+kstgZT7f2HWd8nXVBx0HOuLohF0NcD19X2CRWpRfy1PxUPt+QTURIIDee3IObT+1Fp3Zhvi5NRERERJqIAp60TTWVsPVzp6VuyzxwVUFsdzj55zD0Eug02NcVNglrLYu27eXJBal8n7qXmPBgfn56X64/qQdxkSG+Lk9EREREmpgCnrQdrlrY8Y0zpm7j/5xFyCMTYPT1MPRSSEpp1ZOl1OV2W77alMOT81NZlV5IQnQovztnAFeOSyYqVB97EREREX+lb3ri36yFjKVOS93696EsF0LbOevUDb0EekyEQP/5GNS63Hy0ZjdPL9jG5uwSurUP5/9dOISLRyURFhzo6/JERERExMv855utSF3Z650xdevegcJdEBgK/afBkEug71QI9q9xZ5U1Lt5dkcGz32xnV345/TpF8djlIzhvWBeCAgN8XZ6IiIiINBMFPPEfBTs9oe5dyNkAJhB6TYbJv4MB50KY/03/n1NSydwl6by2OI2ckiqGd4vlvnMHcsbATgQE+Ed3UxERERFpPAU8ad1Kc5yul2vfdrpiAnQbD+f8HQbNgKgEn5bnDdZalqcV8MqiND5dt5sal+XUvh147PIRTOgdj/GTcYQiIiIicuwU8KT1qSxyJklZ+44zaYp1Q6chcMYDzrIGsd19XaFXlFfX8t9VWbyyKI2Nu4uJDgvimvE9uHp8d3olRPm6PBERERFpARTwpHWoqYAtnzmhbusXzrIGcT3glF85k6V0HOjrCr1mR14Zry1O4+1l6RRX1jKgczR/uXAoM0Z2JSJEH2EREREROUDfDqXlctXA9m+c7pebPobqEojqBCk3OqEucbTfLGtwKJfbMn9TDq8sTmPhllyCAgxnD+3CtROSSUmOUzdMEREREamXAp60LG43pP/ozH65/n0o3wuhMTB4hmdZg1MhwH+n+y8oq+bNZc6kKRkFFXRqF8ovz+jHFWO70bGdf838KSIiIiJNTwFPWoY9a52WunXvQVE6BIVB/7OdBcj7nAFBob6u0KtWpxfyyqI0/rcmi+paN+N7ted35wzkzEGdCNYyByIiIiLSSAp44lu5m2He7yD1S2dZg96nwWl/gAHnQGi0r6vzqsoaFx+v2c0ri3ayOqOIyJBALk/pxjUTkunXyb9fu4iIiIh4hwKe+EZFASz4KyyZDSFRcMafYOTVENnB15V5XXp+Oa//uIu3lqWTX1ZN74RI/nTBYC4alUh0WLCvyxMRERGRVkwBT5qXqxaWvwTz/wKVhTDqOjjtPr8Pdm635bvUPF5ZlMbXm7IBOHNQJ66b0ENr14mIiIhIk1HAk+azfQF89lvI2eBMljLtYeg81NdVeVVRRQ3vLM/gtcVp7Mgro0NUCD+d3Icrx3Wna2y4r8sTERERET+jgCfet3cbfP4H2Pyxswj5Za/CwPP9dokDgI27i3llURofrMykosbFqO6x/GLmCKYN6UxokP/OAioiIiIivqWAJ95TWQzf/h0WPw0BwXD6H2H87RDsn9P9V9e6+Wz9Hl5dtJOlOwsICw5g+vBErpmQzJDEGF+XJyIiIiJtgAKeND23G1a9Dl/9GcpyYPiVTrhr18XXlXnFnqJK3liyizlLdpFbUkVyfAS/P2cgl6YkERsR4uvyRERERKQNUcCTppW2CD67B3avhqSxcOVcSBzt66qanLWWxdvzeXXxTuatz8ZtLVP6d+SaCclM6ptAQID/dj8VERERkZZLAU+aRmE6fPFHWP8etEuEi56HoZf43Ti70qpa3l+RwauL09iSXUpsRDA3ndKTq8cl0z0+wtfliYiIiEgbp4AnJ6a6DL7/l3MBmHQPnPxzCIn0bV1NLDWnhFcXpfHuikxKq2oZmhjD3y4ZxgXDuxIWrElTRERERKRl8GrAM8ZMA/4FBALPW2sfOWT/AOAlYBTwe2vt371ZjzQha2Ht2/DF/VCSBUMudhYrj+3m68qahLWW1JxSPlu3h3kb9rAus5iQwADOG9aFayYkM6JbrNauExEREZEWx2sBzxgTCDwJnAlkAEuNMR9aazfUOSwf+Bkww1t1iBdkLodP74WMJdBlBFzyIiRP8HVVJ8zttqzOKGTe+mw+X7+H7XllAIzsHsvvzhnARaOS6BAV6uMqRUREREQa5s0WvLFAqrV2O4AxZi4wHdgf8Ky1OUCOMeZcL9YhTaV4N3z1J1g9ByI7wvQnnRkyAwJ8Xdlxq3G5+XF7PvPW7+HzDXvILq4iKMAwoXc8N5zSk6mDOtGpnX8u6yAiIiIi/sebAS8RSK9zPwMY58XnE2+pqYRFT8C3/wR3DZz8Czj11xDWzteVHZeKahcLt+Yyb/0evtqYQ1FFDWHBAUzql8BZgztz+oBOxEQE+7pMEREREZFj5s2AV98AJXtcJzLmFuAWgO7du59ITXIsrIWNH8Ln90HhLhhwHkx9ENr38nVlx6yoooavN2Uzb10232zJpaLGRbuwIM4Y1ImzBndmYt8EwkM0WYqIiIiItG7eDHgZQN0ZN5KArOM5kbV2NjAbICUl5bhCohyj3Wvgs99C2nfQcTBc+yH0muTrqo5JTnEl8zY44+kWbdtLrdvSqV0ol4xO4qzBnRnXqz3Bga23e6mIiIiIyKG8GfCWAn2NMT2BTGAmcKUXn0+aQmkuzH8Ilr8M4XFw7j9g1PUQ2DpW1NiZV8a89XuYt34PK9MLsRZ6dojkplN7Mm1wZ4YnxWoRchERERHxW1771m6trTXG3AHMw1km4UVr7XpjzK2e/c8YYzoDy4B2gNsY8wtgkLW22Ft1SQNqq2HJs/DN36CmHMbdCpPvcUJeC2atZcPuYuatz2beuj1szi4BYEhiO351Rj/OGtKZvh2jtKSBiIiIiLQJXm2WsdZ+AnxyyLZn6tzeg9N1U3zFWtgyD+b9DvK3QZ8z4ay/QEI/X1fWIJfbsjytYH9LXUZBBQEGUnq05w/nDWLqoE50ax/h6zJFRERERJpd6+h3J96Rs8kJdtu+gvi+cOXb0G+qr6uqV1Wtix+27eXz9Xv4YkM2eaXVhAQGcErfDtx5Wh/OGNiJeK1RJyIiIiJtnAJeW1SeDwsegaXPQ0iU02I3ZhYEhfi6soOUVtWyYHMO89ZnM39TDqVVtUSGBDJlQEfOGtyZyf0TiA7TcgYiIiIiIvso4LUlrlpY/hLM/39QWQSjr4cpv4fIDr6ubL/8smq+3JDNvPV7+DY1j+paN/GRIZw7tAvThnTmpD7xhAZpOQMRERERkfoo4LUV2+Y7yx7kboQep8K0R6DzEF9XBUBeaRX/W53FvPV7WLIjH7eFxNhwrh6XzFmDO5HSoz2BmvlSREREROSoFPD8XcFOmPd72PQRxCbD5a85C5b7eFZJt9vy/bY85izZxRcbsqlxWfp1iuL2KX04a3BnBndtp5kvRURERESOkQKev6qpgO8eg+8fAxMAp/8Rxt8OwWE+LSunpJK3l2Uwd+ku0vMriIsI5roJPbh8TDf6dor2aW0iIiIiIq2dAp6/sdZprZv3OyjcBYMvgqkPQozvVqNwuy3fpR5orat1W8b3as9vpvbnrMGdCQvWmDoRERERkaaggOdP8rbCp3fDtq+h4yC47iPoearPyskpruTt5RnMWbKLjIIK2keGcOMpPbl8TDd6J0T5rC4REREREX+lgOcPqkrgm7/B4qchOBym/RXG3AyBzf/2utyWb7fmMmfJLr7cmIPLbTmpdzz3TBvA1MGdNAOmiIiIiIgXKeC1ZtbC2nfg8/ugdA+MuBrOuB+iOjZ7KXuKKnl7WTpzl6aTWVhBfGQIN5/ak5ljutOzQ2Sz1yMiIiIi0hYp4LVWe9bBJ3fBrh+gywhndsxuY5q1BJfbsnBLLm8s2cXXm5zWulP6dOB35wzkzEGdCAkKaNZ6RERERETaOgW81qaiAOb/BZY+D2GxcP6/YOQ1ENB8XR93F1Xw1tIM3ly6i6yiSjpEhXDLxF7MHNON5Hi11omIiIiI+IoCXmvhdsPKV+GrPzkhL+UmmPI7iGjfLE9f63LzzRZnbN3Xm3JwWzi1bwf+cN4gTh+o1joRERERkZZAAa81yFgOn/wGslZA9wlwzqPQeWizPHVWYQVvLk3nrWXp7C6qJCE6lNsm9+bylO50j49olhpERERERKRxFPBastJc+OoBWPkaRHWGi56DoZeCMV592lqXm/mbnda6BZtzsMDEvgncf/5gTh/YkeBAtdaJiIiIiLRECngtkavWGWM3/y9QUwYn/Qwm3Q2h0V592oyCct5ams6by9LJLq6iY3Qot0/pw2Up3ejWXq11IiIiIiItnQJeS7PzO/jkbshZD72mwNl/g4R+Xnu6GpebrzflMGfJLr7ZkgvA5H4JPDi9O6cN6EiQWutERERERFoNBbyWoigTvvgDrHsXYro7yx4MOM9r3THT88v3j63LKamic7sw7jytL5elJJEUp9Y6EREREZHWSAHP12qrYPFT8M2j4K6FSffCyT+HkKYPWTUuN19tzOaNJel8uzUXA0zp35ErxnZncv8EtdaJiIiIiLRyCni+tPVL+Owe2JsK/c+Fs/4ftO95wqe11lJSVcve0mr2llaRV1rN6oxC3l6WQV5pFV1iwvjZaX25fEw3usaGN8ELERERERGRlkABzxcKdsJnv4PNH0N8H7jqXeh7xhEfUlnjYm+ZE9j2llaTV1p18P2yavLLqjyhrppql/ugxwcYOG2A01o3qZ9a60RERERE/JECXnOqLofvH4PvHsMGBFF26h/IGHA9eysgb1WmE872hbR94a3MCWylVbX1njI0KIAOUaF0iAqhY3QYAzu3I95zPz4qhPjIUOKjQugaE05cZEjzvl4REREREWlWCnhNxFpLcWVtnVDmdI3ML6tmb0klidlfcWHOkyS4cvjUnMKfymey54v28MWPB50nMMDQPjKE+MgQOkSF0q19xP6Q1qFOYNt3HRESiPHyungiIiIiItI6KOA1gX9/tZXHv95Kjcsetq+3yeSh0FeZwBrSgnrwz8T/Y2+HMVy2r5XtkPAWEx5MQIACm4iIiIiIHDsFvCYwrFssN53S66BukQkh1XRf+wQRK2djgiPhtL+RnHITvwrUj1xERERERLxDaaMJTOqXwKR+Cc4da2Ht2/DhH6A0G0ZeDaffD1EJvi1SRERERET8ngJeU9q9Bj69G3Ytgq6jYOYbkDTa11WJiIiIiEgb4dW58o0x04wxm40xqcaYe+vZb4wxj3v2rzHGjPJmPV5Tng8f/wZmT4K8LXDBv+HmrxTuRERERESkWXmtBc8YEwg8CZwJZABLjTEfWms31DnsbKCv5zIOeNpz3bosfBSWvQBjZsGU30J4nK8rEhERERGRNsibXTTHAqnW2u0Axpi5wHSgbsCbDrxirbXAYmNMrDGmi7V2txfranqn/gZGXAWdh/i6EhERERERacO82UUzEUivcz/Ds+1Yj2n5IuMV7kRERERExOe8GfDqW8zt0IXiGnMMxphbjDHLjDHLcnNzm6Q4ERERERERf+PNgJcBdKtzPwnIOo5jsNbOttamWGtTEhK03ICIiIiIiEh9vBnwlgJ9jTE9jTEhwEzgw0OO+RC41jOb5nigqNWNvxMREREREWkhvDbJirW21hhzBzAPCARetNauN8bc6tn/DPAJcA6QCpQDN3irHhEREREREX/n1YXOrbWf4IS4utueqXPbArd7swYREREREZG2wqsLnYuIiIiIiEjzUcATERERERHxEwp4IiIiIiIifkIBT0RERERExE8o4ImIiIiIiPgJ40xk2XoYY3KBNF/XUY8OQJ6vi5D99H60LHo/Wh69Jy2L3o+WRe9Hy6L3o2XR+9EyJFtrE+rb0eoCXktljFlmrU3xdR3i0PvRsuj9aHn0nrQsej9aFr0fLYvej5ZF70fLpy6aIiIiIiIifkIBT0RERERExE8o4DWd2b4uQA6i96Nl0fvR8ug9aVn0frQsej9aFr0fLYvejxZOY/BERERERET8hFrwRERERERE/IQC3jEyxkwzxmw2xqQaY+6tZ78xxjzu2b/GGDPKF3W2BcaYbsaY+caYjcaY9caYn9dzzGRjTJExZpXn8kdf1NpWGGN2GmPWen7Wy+rZr89HMzHG9K/z736VMabYGPOLQ47R58PLjDEvGmNyjDHr6mxrb4z5whiz1XMd18Bjj/j/jRy7Bt6PR40xmzy/k943xsQ28Ngj/n6TY9fA+/GAMSazzu+lcxp4rD4fTayB9+PNOu/FTmPMqgYeq89HC6IumsfAGBMIbAHOBDKApcAV1toNdY45B7gTOAcYB/zLWjvOB+X6PWNMF6CLtXaFMSYaWA7MOOT9mAz8xlp7nm+qbFuMMTuBFGttvevj6PPhG57fXZnAOGttWp3tk9Hnw6uMMROBUuAVa+0Qz7a/AfnW2kc8X0zjrLX3HPK4o/5/I8eugfdjKvC1tbbWGPNXgEPfD89xOznC7zc5dg28Hw8Apdbavx/hcfp8eEF978ch+/8BFFlr/1zPvp3o89FiqAXv2IwFUq2126211cBcYPohx0zH+WBYa+1iINYTRKSJWWt3W2tXeG6XABuBRN9WJUehz4dvnA5sqxvupHlYaxcC+Ydsng687Ln9MjCjnoc25v8bOUb1vR/W2s+ttbWeu4uBpGYvrI1q4PPRGPp8eMGR3g9jjAEuA+Y0a1FyXBTwjk0ikF7nfgaHB4rGHCNNzBjTAxgJ/FjP7gnGmNXGmE+NMYObt7I2xwKfG2OWG2NuqWe/Ph++MZOG/1PW56P5dbLW7gbnD1VAx3qO0WfFN24EPm1g39F+v0nTucPTZfbFBrow6/PR/E4Fsq21WxvYr89HC6KAd2xMPdsO7ePamGOkCRljooB3gV9Ya4sP2b0CSLbWDgf+DXzQzOW1NSdba0cBZwO3e7p71KXPRzMzxoQAFwBv17Nbn4+WS5+VZmaM+T1QC7zewCFH+/0mTeNpoDcwAtgN/KOeY/T5aH5XcOTWO30+WhAFvGOTAXSrcz8JyDqOY6SJGGOCccLd69ba9w7db60tttaWem5/AgQbYzo0c5lthrU2y3OdA7yP042mLn0+mt/ZwAprbfahO/T58JnsfV2TPdc59Ryjz0ozMsZcB5wHXGUbmJygEb/fpAlYa7OttS5rrRt4jvp/zvp8NCNjTBBwEfBmQ8fo89GyKOAdm6VAX2NMT89fxWcCHx5yzIfAtc5kgWY8zmDU3c1daFvg6Q/+ArDRWvvPBo7p7DkOY8xYnH/ze5uvyrbDGBPpmewGY0wkMBVYd8hh+nw0vwb/6qrPh898CFznuX0d8N96jmnM/zfSBIwx04B7gAusteUNHNOY32/SBA4Zl30h9f+c9floXmcAm6y1GfXt1Oej5QnydQGtiWeGrTuAeUAg8KK1dr0x5lbP/meAT3BmCEwFyoEbfFVvG3AycA2wts60vb8DusP+9+MS4DZjTC1QAcxs6K+zcsI6Ae978kIQ8Ia19jN9PnzHGBOBM8vcT+psq/t+6PPhZcaYOcBkoIMxJgO4H3gEeMsYcxOwC7jUc2xX4Hlr7TkN/X/ji9fgTxp4P34LhAJfeH5/LbbW3lr3/aCB328+eAl+pYH3Y7IxZgROl8udeH5/6fPhffW9H9baF6hnHLc+Hy2blkkQERERERHxE+qiKSIiIiIi4icU8ERERERERPyEAp6IiIiIiIifUMATERERERHxEwp4IiIiIiIifkIBT0RE2ixjjMsYs6rO5d4mPHcPY4zWghIRkWaldfBERKQtq7DWjvB1ESIiIk1FLXgiIiKHMMbsNMb81RizxHPp49mebIz5yhizxnPd3bO9kzHmfWPMas/lJM+pAo0xzxlj1htjPjfGhPvsRYmISJuggCciIm1Z+CFdNC+vs6/YWjsWeAJ4zLPtCeAVa+0w4HXgcc/2x4FvrLXDgVHAes/2vsCT1trBQCFwsVdfjYiItHnGWuvrGkRERHzCGFNqrY2qZ/tO4DRr7XZjTDCwx1obb4zJA7pYa2s823dbazsYY3KBJGttVZ1z9AC+sNb29dy/Bwi21j7UDC9NRETaKLXgiYiI1M82cLuhY+pTVee2C419FxERL1PAExERqd/lda4XeW7/AMz03L4K+M5z+yvgNgBjTKAxpl1zFSkiIlKX/pIoIiJtWbgxZlWd+59Za/ctlRBqjPkR54+hV3i2/Qx40RhzF5AL3ODZ/nNgtjHmJpyWutuA3d4uXkRE5FAagyciInIIzxi8FGttnq9rERERORbqoikiIiIiIuIn1IInIiIiIiLiJ9SCJyIiIiIi4icU8ERERERERPyEAp6IiIiIiIifUMATERERERHxEwp4IiIiIiIifkIBT0RERERExE/8f1Wq5+TQDfNHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "losses = model_trainer.losses\n",
    "dice_scores = model_trainer.dice_scores # overall dice\n",
    "iou_scores = model_trainer.iou_scores\n",
    "\n",
    "def plot(scores, name):\n",
    "    plt.figure(figsize=(15,5))\n",
    "    plt.plot(range(len(scores[\"train\"])), scores[\"train\"], label=f'train {name}')\n",
    "    plt.plot(range(len(scores[\"train\"])), scores[\"val\"], label=f'val {name}')\n",
    "    plt.title(f'{name} plot'); plt.xlabel('Epoch'); plt.ylabel(f'{name}');\n",
    "    plt.legend(); \n",
    "    plt.show()\n",
    "\n",
    "plot(losses, \"BCE loss\")\n",
    "plot(dice_scores, \"Dice score\")\n",
    "plot(iou_scores, \"IoU score\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "93c0c6dd",
   "metadata": {},
   "source": [
    "From the above results it can be concluded that the loss, Dice and IOU get's reaches plateau after epoch 15. So, it can be concluded that it's not worth training the model after 20 epochs."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f538c609",
   "metadata": {},
   "source": [
    "# Convert to ONNX\n",
    "\n",
    "Lastly the trained model is converted to ONNX format using the Utility script ONNX_Converter.py."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a1027df4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model name :Unet\n",
      "Model checkpoint path is :./Models/model_unet.pth\n",
      "Model output path is :./ONNX_models/unet.onnx\n",
      "Successfully converted the model to onnx format.\n"
     ]
    }
   ],
   "source": [
    "! python ./Utilities/ONNX_converter.py Unet resnet34  imagenet  6 ./Models/model_unet.pth ./ONNX_models/unet.onnx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "286d2760",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}