{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<a href=\"https://colab.research.google.com/github/PacktPublishing/Hands-On-Computer-Vision-with-PyTorch/blob/master/Chapter09/predicting_multiple_instances_of_multiple_classes.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[K     |████████████████████████████████| 61kB 4.8MB/s \n",
      "\u001b[K     |████████████████████████████████| 36.7MB 77kB/s \n",
      "\u001b[K     |████████████████████████████████| 102kB 14.6MB/s \n",
      "\u001b[?25h  Building wheel for contextvars (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
      "  Building wheel for pycocotools (setup.py) ... \u001b[?25l\u001b[?25hdone\n"
     ]
    }
   ],
   "source": [
    "!wget --quiet http://sceneparsing.csail.mit.edu/data/ChallengeData2017/images.tar\n",
    "!wget --quiet http://sceneparsing.csail.mit.edu/data/ChallengeData2017/annotations_instance.tar\n",
    "!tar -xf images.tar\n",
    "!tar -xf annotations_instance.tar\n",
    "!rm images.tar annotations_instance.tar\n",
    "!pip install -qU torch_snippets\n",
    "!wget --quiet https://raw.githubusercontent.com/pytorch/vision/master/references/detection/engine.py\n",
    "!wget --quiet https://raw.githubusercontent.com/pytorch/vision/master/references/detection/utils.py\n",
    "!wget --quiet https://raw.githubusercontent.com/pytorch/vision/master/references/detection/transforms.py\n",
    "!wget --quiet https://raw.githubusercontent.com/pytorch/vision/master/references/detection/coco_eval.py\n",
    "!wget --quiet https://raw.githubusercontent.com/pytorch/vision/master/references/detection/coco_utils.py\n",
    "!pip install -q -U 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch_snippets import *\n",
    "\n",
    "import torchvision\n",
    "from torchvision.models.detection.faster_rcnn import FastRCNNPredictor\n",
    "from torchvision.models.detection.mask_rcnn import MaskRCNNPredictor\n",
    "\n",
    "from engine import train_one_epoch, evaluate\n",
    "import utils\n",
    "import transforms as T\n",
    "device = 'cuda' if torch.cuda.is_available() else 'cpu'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-11-07 13:58:04.425 | INFO     | torch_snippets.loader:Glob:181 - 20210 files found at images/training\n",
      "2020-11-07 13:58:04.469 | INFO     | torch_snippets.loader:Glob:181 - 20210 files found at annotations_instance/training\n"
     ]
    }
   ],
   "source": [
    "all_images = Glob('images/training')\n",
    "all_annots = Glob('annotations_instance/training')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 20210/20210 [03:10<00:00, 106.29it/s]\n"
     ]
    }
   ],
   "source": [
    "classes_list = [4,6]\n",
    "annots = []\n",
    "for ann in Tqdm(all_annots):\n",
    "    _ann = read(ann, 1).transpose(2,0,1)\n",
    "    r,g,b = _ann\n",
    "    if np.array([num in np.unique(r) for num in classes_list]).sum()==0: continue\n",
    "    annots.append(ann)\n",
    "from sklearn.model_selection import train_test_split\n",
    "_annots = stems(annots)\n",
    "trn_items, val_items = train_test_split(_annots, random_state=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_transform(train):\n",
    "    transforms = []\n",
    "    transforms.append(T.PILToTensor())\n",
    "    if train:\n",
    "        transforms.append(T.RandomHorizontalFlip(0.5))\n",
    "    return T.Compose(transforms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_mask(path):\n",
    "    an = read(path, 1).transpose(2,0,1)\n",
    "    r,g,b = an\n",
    "    cls = list(set(np.unique(r)).intersection({4,6}))\n",
    "    print(cls)\n",
    "    masks = []\n",
    "    labels = []\n",
    "    for _cls in cls:\n",
    "      nzs = np.nonzero(r==_cls)\n",
    "      instances = np.unique(g[nzs])\n",
    "      for ix,_id in enumerate(instances):\n",
    "          masks.append(g==_id)\n",
    "          labels.append(classes_list.index(_cls)+1)\n",
    "    return np.array(masks), np.array(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "_id = trn_items[10]\n",
    "img_path = f'images/training/{_id}.jpg'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[4, 6]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(array([[[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]],\n",
       " \n",
       "        [[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]],\n",
       " \n",
       "        [[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]],\n",
       " \n",
       "        ...,\n",
       " \n",
       "        [[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]],\n",
       " \n",
       "        [[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]],\n",
       " \n",
       "        [[False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         ...,\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False],\n",
       "         [False, False, False, ..., False, False, False]]]),\n",
       " array([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "        1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
       "        2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]))"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_mask(img_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2020-11-07 14:01:15.692 | INFO     | torch_snippets.loader:subplots:375 - plotting 8 images in a grid of 2x5 @ (5, 5)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==================================================================\n",
      "Tensor\tShape: torch.Size([3, 512, 683])\tMin: 0.000\tMax: 1.000\tMean: 0.504\tdtype: torch.float32\n",
      "==================================================================\n",
      "Dict Of 6 items\n",
      "==================================================================\n",
      "\tBOXES:\n",
      "\tTensor\tShape: torch.Size([7, 4])\tMin: 1.000\tMax: 475.000\tMean: 278.214\tdtype: torch.float32\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tLABELS:\n",
      "\tTensor\tShape: torch.Size([7])\tMin: 1.000\tMax: 1.000\tMean: 1.000\tdtype: torch.int64\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tMASKS:\n",
      "\tTensor\tShape: torch.Size([7, 512, 683])\tMin: 0.000\tMax: 1.000\tMean: 0.009\tdtype: torch.uint8\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tIMAGE_ID:\n",
      "\tTensor\tShape: torch.Size([1])\tMin: 1.000\tMax: 1.000\tMean: 1.000\tdtype: torch.int64\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tAREA:\n",
      "\tTensor\tShape: torch.Size([7])\tMin: 432.000\tMax: 10366.000\tMean: 6474.571\tdtype: torch.float32\n",
      "==================================================================\n",
      "... ... 1 more items\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x720 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "class MasksDataset(Dataset):\n",
    "    def __init__(self, items, transforms, N):\n",
    "        self.items = items\n",
    "        self.transforms = transforms\n",
    "        self.N = N\n",
    "    def get_mask(self,path):\n",
    "        an = read(path, 1).transpose(2,0,1)\n",
    "        r,g,b = an\n",
    "        cls = list(set(np.unique(r)).intersection({4,6}))\n",
    "        masks = []\n",
    "        labels = []\n",
    "        for _cls in cls:\n",
    "          nzs = np.nonzero(r==_cls)\n",
    "          instances = np.unique(g[nzs])\n",
    "          for ix,_id in enumerate(instances):\n",
    "              masks.append(g==_id)\n",
    "              labels.append(classes_list.index(_cls)+1)\n",
    "        return np.array(masks), np.array(labels)\n",
    "    def __getitem__(self, ix):\n",
    "        _id = self.items[ix]\n",
    "        img_path = f'images/training/{_id}.jpg'\n",
    "        mask_path = f'annotations_instance/training/{_id}.png'\n",
    "        masks, labels = self.get_mask(mask_path)\n",
    "        #print(labels)\n",
    "        obj_ids = np.arange(1, len(masks)+1)\n",
    "        img = Image.open(img_path).convert(\"RGB\")\n",
    "        num_objs = len(obj_ids)\n",
    "        boxes = []\n",
    "        for i in range(num_objs):\n",
    "            obj_pixels = np.where(masks[i])\n",
    "            xmin = np.min(obj_pixels[1])\n",
    "            xmax = np.max(obj_pixels[1])\n",
    "            ymin = np.min(obj_pixels[0])\n",
    "            ymax = np.max(obj_pixels[0])\n",
    "            if (((xmax-xmin)<=10) | (ymax-ymin)<=10):\n",
    "                xmax = xmin+10\n",
    "                ymax = ymin+10\n",
    "            boxes.append([xmin, ymin, xmax, ymax])\n",
    "        boxes = torch.as_tensor(boxes, dtype=torch.float32)\n",
    "        labels = torch.as_tensor(labels, dtype=torch.int64)\n",
    "        masks = torch.as_tensor(masks, dtype=torch.uint8)\n",
    "        area = (boxes[:, 3] - boxes[:, 1]) * (boxes[:, 2] - boxes[:, 0])\n",
    "        iscrowd = torch.zeros((num_objs,), dtype=torch.int64)\n",
    "        image_id = torch.tensor([ix])\n",
    "        target = {}\n",
    "        target[\"boxes\"] = boxes\n",
    "        target[\"labels\"] = labels\n",
    "        target[\"masks\"] = masks\n",
    "        target[\"image_id\"] = image_id\n",
    "        target[\"area\"] = area\n",
    "        target[\"iscrowd\"] = iscrowd\n",
    "        if self.transforms is not None:\n",
    "            img, target = self.transforms(img, target)\n",
    "        if (img.dtype == torch.float32) or (img.dtype == torch.uint8):\n",
    "          img = img/255.\n",
    "        return img, target\n",
    "    def __len__(self):\n",
    "        return self.N\n",
    "    def choose(self):\n",
    "        return self[randint(len(self))]\n",
    "x = MasksDataset(trn_items, get_transform(train=True), N=100)\n",
    "im,targ = x[1]\n",
    "inspect(im,targ)\n",
    "subplots([im, *targ['masks']], sz=10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_model_instance_segmentation(num_classes):\n",
    "    # load an instance segmentation model pre-trained pre-trained on COCO\n",
    "    model = torchvision.models.detection.maskrcnn_resnet50_fpn(pretrained=True)\n",
    "\n",
    "    # get number of input features for the classifier\n",
    "    in_features = model.roi_heads.box_predictor.cls_score.in_features\n",
    "    # replace the pre-trained head with a new one\n",
    "    model.roi_heads.box_predictor = FastRCNNPredictor(in_features, num_classes)\n",
    "\n",
    "    # now get the number of input features for the mask classifier\n",
    "    in_features_mask = model.roi_heads.mask_predictor.conv5_mask.in_channels\n",
    "    hidden_layer = 256\n",
    "    # and replace the mask predictor with a new one\n",
    "    model.roi_heads.mask_predictor = MaskRCNNPredictor(in_features_mask,\n",
    "                                                       hidden_layer,num_classes)\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Downloading: \"https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth\" to /root/.cache/torch/hub/checkpoints/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d4df3b571db049f499241162165eaaad",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=178090079.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "model = get_model_instance_segmentation(3).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = MasksDataset(trn_items, get_transform(train=True), N=3000)\n",
    "dataset_test = MasksDataset(val_items, get_transform(train=False), N=800)\n",
    "\n",
    "# define training and validation data loaders\n",
    "data_loader = torch.utils.data.DataLoader(\n",
    "    dataset, batch_size=2, shuffle=True, num_workers=0,\n",
    "    collate_fn=utils.collate_fn)\n",
    "\n",
    "data_loader_test = torch.utils.data.DataLoader(\n",
    "    dataset_test, batch_size=1, shuffle=False, num_workers=0,\n",
    "    collate_fn=utils.collate_fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_classes = 3\n",
    "model = get_model_instance_segmentation(num_classes).to(device)\n",
    "params = [p for p in model.parameters() if p.requires_grad]\n",
    "optimizer = torch.optim.SGD(params, lr=0.005,\n",
    "                            momentum=0.9, weight_decay=0.0005)\n",
    "# and a learning rate scheduler\n",
    "lr_scheduler = torch.optim.lr_scheduler.StepLR(optimizer,\n",
    "                                                step_size=3,\n",
    "                                                gamma=0.1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==================================================================\n",
      "Dict Of 4 items\n",
      "==================================================================\n",
      "\tBOXES:\n",
      "\tTensor\tShape: torch.Size([100, 4])\tMin: 0.000\tMax: 382.000\tMean: 177.219\tdtype: torch.float32\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tLABELS:\n",
      "\tTensor\tShape: torch.Size([100])\tMin: 1.000\tMax: 2.000\tMean: 1.120\tdtype: torch.int64\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tSCORES:\n",
      "\tTensor\tShape: torch.Size([100])\tMin: 0.355\tMax: 0.712\tMean: 0.409\tdtype: torch.float32\n",
      "==================================================================\n",
      "==================================================================\n",
      "\tMASKS:\n",
      "\tTensor\tShape: torch.Size([100, 1, 340, 382])\tMin: 0.000\tMax: 1.000\tMean: 0.054\tdtype: torch.float32\n",
      "==================================================================\n"
     ]
    }
   ],
   "source": [
    "model.eval()\n",
    "pred = model(dataset[0][0][None].to(device))\n",
    "inspect(pred[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:3103: UserWarning: The default behavior for interpolate/upsample with float scale_factor changed in 1.6.0 to align with other frameworks/libraries, and now uses scale_factor directly, instead of relying on the computed output size. If you wish to restore the old behavior, please set recompute_scale_factor=True. See the documentation of nn.Upsample for details. \n",
      "  warnings.warn(\"The default behavior for interpolate/upsample with float scale_factor changed \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: [0]  [   0/1500]  eta: 0:16:16  lr: 0.000010  loss: 2.2653 (2.2653)  loss_classifier: 0.9195 (0.9195)  loss_box_reg: 0.1731 (0.1731)  loss_mask: 1.1160 (1.1160)  loss_objectness: 0.0399 (0.0399)  loss_rpn_box_reg: 0.0168 (0.0168)  time: 0.6513  data: 0.0119  max mem: 2875\n",
      "Epoch: [0]  [  10/1500]  eta: 0:16:16  lr: 0.000060  loss: 2.5998 (2.6990)  loss_classifier: 0.9031 (0.8928)  loss_box_reg: 0.1800 (0.2189)  loss_mask: 1.3047 (1.4537)  loss_objectness: 0.0399 (0.0971)  loss_rpn_box_reg: 0.0132 (0.0365)  time: 0.6557  data: 0.0690  max mem: 4126\n",
      "Epoch: [0]  [  20/1500]  eta: 0:16:07  lr: 0.000110  loss: 2.3038 (2.2296)  loss_classifier: 0.8112 (0.7332)  loss_box_reg: 0.1800 (0.2395)  loss_mask: 1.0804 (1.1661)  loss_objectness: 0.0265 (0.0665)  loss_rpn_box_reg: 0.0102 (0.0243)  time: 0.6540  data: 0.0871  max mem: 4126\n",
      "Epoch: [0]  [  30/1500]  eta: 0:15:57  lr: 0.000160  loss: 1.4503 (1.9446)  loss_classifier: 0.3333 (0.5951)  loss_box_reg: 0.1977 (0.2451)  loss_mask: 0.6966 (1.0212)  loss_objectness: 0.0301 (0.0623)  loss_rpn_box_reg: 0.0087 (0.0209)  time: 0.6486  data: 0.0963  max mem: 4126\n",
      "Epoch: [0]  [  40/1500]  eta: 0:15:56  lr: 0.000210  loss: 1.1571 (1.7586)  loss_classifier: 0.2501 (0.5046)  loss_box_reg: 0.2067 (0.2340)  loss_mask: 0.6182 (0.9355)  loss_objectness: 0.0446 (0.0620)  loss_rpn_box_reg: 0.0087 (0.0226)  time: 0.6562  data: 0.0884  max mem: 4131\n",
      "Epoch: [0]  [  50/1500]  eta: 0:15:48  lr: 0.000260  loss: 1.1051 (1.6331)  loss_classifier: 0.2395 (0.4632)  loss_box_reg: 0.2037 (0.2460)  loss_mask: 0.5177 (0.8413)  loss_objectness: 0.0446 (0.0616)  loss_rpn_box_reg: 0.0081 (0.0209)  time: 0.6583  data: 0.0791  max mem: 4131\n",
      "Epoch: [0]  [  60/1500]  eta: 0:15:41  lr: 0.000310  loss: 1.1051 (1.5487)  loss_classifier: 0.1952 (0.4314)  loss_box_reg: 0.2243 (0.2512)  loss_mask: 0.4432 (0.7832)  loss_objectness: 0.0462 (0.0616)  loss_rpn_box_reg: 0.0095 (0.0213)  time: 0.6523  data: 0.0721  max mem: 4131\n",
      "Epoch: [0]  [  70/1500]  eta: 0:15:54  lr: 0.000360  loss: 1.0199 (1.4867)  loss_classifier: 0.2294 (0.4076)  loss_box_reg: 0.2438 (0.2603)  loss_mask: 0.4432 (0.7373)  loss_objectness: 0.0462 (0.0599)  loss_rpn_box_reg: 0.0107 (0.0216)  time: 0.7012  data: 0.0947  max mem: 4131\n",
      "Epoch: [0]  [  80/1500]  eta: 0:15:49  lr: 0.000410  loss: 0.9602 (1.4117)  loss_classifier: 0.1960 (0.3806)  loss_box_reg: 0.2150 (0.2553)  loss_mask: 0.4014 (0.6991)  loss_objectness: 0.0297 (0.0560)  loss_rpn_box_reg: 0.0095 (0.0207)  time: 0.7145  data: 0.1064  max mem: 4131\n",
      "Epoch: [0]  [  90/1500]  eta: 0:15:53  lr: 0.000460  loss: 0.9084 (1.3598)  loss_classifier: 0.1752 (0.3622)  loss_box_reg: 0.1835 (0.2572)  loss_mask: 0.3799 (0.6670)  loss_objectness: 0.0330 (0.0536)  loss_rpn_box_reg: 0.0070 (0.0199)  time: 0.7080  data: 0.1020  max mem: 4473\n",
      "Epoch: [0]  [ 100/1500]  eta: 0:15:48  lr: 0.000509  loss: 0.8353 (1.2975)  loss_classifier: 0.1699 (0.3416)  loss_box_reg: 0.2368 (0.2553)  loss_mask: 0.3307 (0.6320)  loss_objectness: 0.0175 (0.0502)  loss_rpn_box_reg: 0.0057 (0.0184)  time: 0.7126  data: 0.0970  max mem: 4473\n",
      "Epoch: [0]  [ 110/1500]  eta: 0:15:48  lr: 0.000559  loss: 0.8168 (1.2722)  loss_classifier: 0.1372 (0.3259)  loss_box_reg: 0.2181 (0.2537)  loss_mask: 0.3334 (0.6144)  loss_objectness: 0.0160 (0.0586)  loss_rpn_box_reg: 0.0064 (0.0196)  time: 0.7096  data: 0.0945  max mem: 4473\n",
      "Epoch: [0]  [ 120/1500]  eta: 0:15:46  lr: 0.000609  loss: 0.9003 (1.2430)  loss_classifier: 0.1705 (0.3144)  loss_box_reg: 0.2616 (0.2566)  loss_mask: 0.3729 (0.5952)  loss_objectness: 0.0218 (0.0569)  loss_rpn_box_reg: 0.0102 (0.0199)  time: 0.7260  data: 0.1024  max mem: 4473\n",
      "Epoch: [0]  [ 130/1500]  eta: 0:15:40  lr: 0.000659  loss: 0.8918 (1.2197)  loss_classifier: 0.1694 (0.3053)  loss_box_reg: 0.2331 (0.2602)  loss_mask: 0.3524 (0.5803)  loss_objectness: 0.0245 (0.0546)  loss_rpn_box_reg: 0.0084 (0.0194)  time: 0.7095  data: 0.0853  max mem: 4473\n",
      "Epoch: [0]  [ 140/1500]  eta: 0:15:34  lr: 0.000709  loss: 0.8490 (1.1986)  loss_classifier: 0.1652 (0.2953)  loss_box_reg: 0.2302 (0.2582)  loss_mask: 0.3517 (0.5654)  loss_objectness: 0.0255 (0.0593)  loss_rpn_box_reg: 0.0074 (0.0204)  time: 0.6977  data: 0.0726  max mem: 4473\n",
      "Epoch: [0]  [ 150/1500]  eta: 0:15:34  lr: 0.000759  loss: 0.7261 (1.1702)  loss_classifier: 0.1417 (0.2851)  loss_box_reg: 0.1823 (0.2537)  loss_mask: 0.3778 (0.5545)  loss_objectness: 0.0190 (0.0573)  loss_rpn_box_reg: 0.0065 (0.0196)  time: 0.7280  data: 0.0850  max mem: 4473\n",
      "Epoch: [0]  [ 160/1500]  eta: 0:15:30  lr: 0.000809  loss: 0.7103 (1.1436)  loss_classifier: 0.1268 (0.2765)  loss_box_reg: 0.1819 (0.2518)  loss_mask: 0.3649 (0.5409)  loss_objectness: 0.0218 (0.0554)  loss_rpn_box_reg: 0.0050 (0.0189)  time: 0.7420  data: 0.0975  max mem: 4473\n",
      "Epoch: [0]  [ 170/1500]  eta: 0:15:21  lr: 0.000859  loss: 0.6961 (1.1174)  loss_classifier: 0.1095 (0.2663)  loss_box_reg: 0.1657 (0.2459)  loss_mask: 0.3624 (0.5320)  loss_objectness: 0.0171 (0.0548)  loss_rpn_box_reg: 0.0060 (0.0183)  time: 0.7016  data: 0.0866  max mem: 4473\n",
      "Epoch: [0]  [ 180/1500]  eta: 0:15:12  lr: 0.000909  loss: 0.7290 (1.1038)  loss_classifier: 0.1072 (0.2593)  loss_box_reg: 0.1932 (0.2462)  loss_mask: 0.3745 (0.5246)  loss_objectness: 0.0171 (0.0549)  loss_rpn_box_reg: 0.0060 (0.0187)  time: 0.6673  data: 0.0650  max mem: 4473\n",
      "Epoch: [0]  [ 190/1500]  eta: 0:15:09  lr: 0.000959  loss: 0.8180 (1.0891)  loss_classifier: 0.1337 (0.2526)  loss_box_reg: 0.2335 (0.2461)  loss_mask: 0.3636 (0.5169)  loss_objectness: 0.0374 (0.0542)  loss_rpn_box_reg: 0.0096 (0.0194)  time: 0.7028  data: 0.0863  max mem: 4473\n",
      "Epoch: [0]  [ 200/1500]  eta: 0:15:06  lr: 0.001009  loss: 0.7390 (1.0742)  loss_classifier: 0.1156 (0.2470)  loss_box_reg: 0.2151 (0.2464)  loss_mask: 0.3466 (0.5074)  loss_objectness: 0.0328 (0.0538)  loss_rpn_box_reg: 0.0175 (0.0196)  time: 0.7542  data: 0.1204  max mem: 4473\n",
      "Epoch: [0]  [ 210/1500]  eta: 0:15:01  lr: 0.001059  loss: 0.7492 (1.0611)  loss_classifier: 0.1222 (0.2409)  loss_box_reg: 0.1683 (0.2427)  loss_mask: 0.3621 (0.5034)  loss_objectness: 0.0313 (0.0540)  loss_rpn_box_reg: 0.0159 (0.0200)  time: 0.7440  data: 0.1013  max mem: 4473\n",
      "Epoch: [0]  [ 220/1500]  eta: 0:14:54  lr: 0.001109  loss: 0.7671 (1.0502)  loss_classifier: 0.1222 (0.2363)  loss_box_reg: 0.1740 (0.2426)  loss_mask: 0.3884 (0.4974)  loss_objectness: 0.0316 (0.0537)  loss_rpn_box_reg: 0.0174 (0.0202)  time: 0.7140  data: 0.0800  max mem: 4473\n",
      "Epoch: [0]  [ 230/1500]  eta: 0:14:52  lr: 0.001159  loss: 0.8191 (1.0468)  loss_classifier: 0.1327 (0.2318)  loss_box_reg: 0.1858 (0.2409)  loss_mask: 0.4003 (0.4993)  loss_objectness: 0.0348 (0.0539)  loss_rpn_box_reg: 0.0130 (0.0209)  time: 0.7488  data: 0.1029  max mem: 4473\n",
      "Epoch: [0]  [ 240/1500]  eta: 0:14:47  lr: 0.001209  loss: 0.7891 (1.0363)  loss_classifier: 0.1025 (0.2271)  loss_box_reg: 0.1499 (0.2388)  loss_mask: 0.4373 (0.4967)  loss_objectness: 0.0292 (0.0530)  loss_rpn_box_reg: 0.0048 (0.0207)  time: 0.7614  data: 0.1126  max mem: 4473\n",
      "Epoch: [0]  [ 250/1500]  eta: 0:14:39  lr: 0.001259  loss: 0.7479 (1.0273)  loss_classifier: 0.1064 (0.2230)  loss_box_reg: 0.1431 (0.2354)  loss_mask: 0.4046 (0.4941)  loss_objectness: 0.0322 (0.0537)  loss_rpn_box_reg: 0.0047 (0.0212)  time: 0.7104  data: 0.0997  max mem: 4473\n",
      "Epoch: [0]  [ 260/1500]  eta: 0:14:34  lr: 0.001309  loss: 0.7479 (1.0168)  loss_classifier: 0.1056 (0.2184)  loss_box_reg: 0.1125 (0.2321)  loss_mask: 0.4035 (0.4913)  loss_objectness: 0.0350 (0.0538)  loss_rpn_box_reg: 0.0143 (0.0211)  time: 0.7231  data: 0.1069  max mem: 4473\n",
      "Epoch: [0]  [ 270/1500]  eta: 0:14:28  lr: 0.001359  loss: 0.6159 (1.0020)  loss_classifier: 0.0764 (0.2136)  loss_box_reg: 0.1067 (0.2286)  loss_mask: 0.3547 (0.4855)  loss_objectness: 0.0277 (0.0531)  loss_rpn_box_reg: 0.0061 (0.0212)  time: 0.7314  data: 0.1119  max mem: 4473\n",
      "Epoch: [0]  [ 280/1500]  eta: 0:14:23  lr: 0.001409  loss: 0.5947 (0.9896)  loss_classifier: 0.0691 (0.2095)  loss_box_reg: 0.1111 (0.2257)  loss_mask: 0.3179 (0.4813)  loss_objectness: 0.0236 (0.0521)  loss_rpn_box_reg: 0.0035 (0.0210)  time: 0.7352  data: 0.1157  max mem: 4473\n",
      "Epoch: [0]  [ 290/1500]  eta: 0:14:16  lr: 0.001459  loss: 0.5959 (0.9811)  loss_classifier: 0.0773 (0.2061)  loss_box_reg: 0.1111 (0.2243)  loss_mask: 0.3370 (0.4790)  loss_objectness: 0.0187 (0.0508)  loss_rpn_box_reg: 0.0035 (0.0209)  time: 0.7322  data: 0.1074  max mem: 4473\n",
      "Epoch: [0]  [ 300/1500]  eta: 0:14:08  lr: 0.001508  loss: 0.6483 (0.9711)  loss_classifier: 0.0810 (0.2030)  loss_box_reg: 0.1158 (0.2222)  loss_mask: 0.3645 (0.4754)  loss_objectness: 0.0108 (0.0498)  loss_rpn_box_reg: 0.0059 (0.0207)  time: 0.6926  data: 0.0852  max mem: 4473\n",
      "Epoch: [0]  [ 310/1500]  eta: 0:14:00  lr: 0.001558  loss: 0.6385 (0.9584)  loss_classifier: 0.0913 (0.1994)  loss_box_reg: 0.1441 (0.2207)  loss_mask: 0.3306 (0.4691)  loss_objectness: 0.0108 (0.0487)  loss_rpn_box_reg: 0.0075 (0.0205)  time: 0.6939  data: 0.0790  max mem: 4473\n",
      "Epoch: [0]  [ 320/1500]  eta: 0:13:52  lr: 0.001608  loss: 0.6305 (0.9511)  loss_classifier: 0.1057 (0.1972)  loss_box_reg: 0.1517 (0.2197)  loss_mask: 0.2950 (0.4647)  loss_objectness: 0.0162 (0.0489)  loss_rpn_box_reg: 0.0075 (0.0207)  time: 0.6922  data: 0.0739  max mem: 4473\n",
      "Epoch: [0]  [ 330/1500]  eta: 0:13:47  lr: 0.001658  loss: 0.5848 (0.9404)  loss_classifier: 0.0951 (0.1941)  loss_box_reg: 0.1205 (0.2167)  loss_mask: 0.3052 (0.4610)  loss_objectness: 0.0169 (0.0482)  loss_rpn_box_reg: 0.0053 (0.0204)  time: 0.7154  data: 0.0885  max mem: 4473\n",
      "Epoch: [0]  [ 340/1500]  eta: 0:13:40  lr: 0.001708  loss: 0.5848 (0.9316)  loss_classifier: 0.0877 (0.1914)  loss_box_reg: 0.1083 (0.2147)  loss_mask: 0.3052 (0.4577)  loss_objectness: 0.0153 (0.0474)  loss_rpn_box_reg: 0.0052 (0.0204)  time: 0.7395  data: 0.0989  max mem: 4703\n",
      "Epoch: [0]  [ 350/1500]  eta: 0:13:34  lr: 0.001758  loss: 0.6230 (0.9242)  loss_classifier: 0.0926 (0.1895)  loss_box_reg: 0.1467 (0.2135)  loss_mask: 0.3320 (0.4542)  loss_objectness: 0.0185 (0.0469)  loss_rpn_box_reg: 0.0075 (0.0202)  time: 0.7328  data: 0.0954  max mem: 4703\n",
      "Epoch: [0]  [ 360/1500]  eta: 0:13:26  lr: 0.001808  loss: 0.6169 (0.9164)  loss_classifier: 0.1083 (0.1870)  loss_box_reg: 0.1563 (0.2118)  loss_mask: 0.3377 (0.4511)  loss_objectness: 0.0158 (0.0464)  loss_rpn_box_reg: 0.0075 (0.0202)  time: 0.6957  data: 0.0797  max mem: 4703\n",
      "Epoch: [0]  [ 370/1500]  eta: 0:13:19  lr: 0.001858  loss: 0.6363 (0.9121)  loss_classifier: 0.1041 (0.1852)  loss_box_reg: 0.1368 (0.2114)  loss_mask: 0.3377 (0.4484)  loss_objectness: 0.0142 (0.0465)  loss_rpn_box_reg: 0.0090 (0.0205)  time: 0.6937  data: 0.0847  max mem: 4703\n",
      "Epoch: [0]  [ 380/1500]  eta: 0:13:13  lr: 0.001908  loss: 0.5914 (0.9040)  loss_classifier: 0.0954 (0.1827)  loss_box_reg: 0.1205 (0.2097)  loss_mask: 0.3184 (0.4451)  loss_objectness: 0.0196 (0.0460)  loss_rpn_box_reg: 0.0134 (0.0204)  time: 0.7308  data: 0.1070  max mem: 4703\n",
      "Epoch: [0]  [ 390/1500]  eta: 0:13:06  lr: 0.001958  loss: 0.6231 (0.9003)  loss_classifier: 0.0954 (0.1813)  loss_box_reg: 0.1349 (0.2091)  loss_mask: 0.3184 (0.4432)  loss_objectness: 0.0218 (0.0459)  loss_rpn_box_reg: 0.0158 (0.0208)  time: 0.7182  data: 0.0857  max mem: 4703\n",
      "Epoch: [0]  [ 400/1500]  eta: 0:12:58  lr: 0.002008  loss: 0.6725 (0.8938)  loss_classifier: 0.0890 (0.1798)  loss_box_reg: 0.1103 (0.2071)  loss_mask: 0.3674 (0.4409)  loss_objectness: 0.0271 (0.0455)  loss_rpn_box_reg: 0.0096 (0.0206)  time: 0.6954  data: 0.0732  max mem: 4703\n",
      "Epoch: [0]  [ 410/1500]  eta: 0:12:54  lr: 0.002058  loss: 0.7020 (0.8899)  loss_classifier: 0.1168 (0.1782)  loss_box_reg: 0.1103 (0.2064)  loss_mask: 0.3925 (0.4397)  loss_objectness: 0.0271 (0.0451)  loss_rpn_box_reg: 0.0092 (0.0204)  time: 0.7515  data: 0.1277  max mem: 4703\n",
      "Epoch: [0]  [ 420/1500]  eta: 0:12:45  lr: 0.002108  loss: 0.6772 (0.8824)  loss_classifier: 0.1004 (0.1763)  loss_box_reg: 0.1310 (0.2045)  loss_mask: 0.3433 (0.4370)  loss_objectness: 0.0143 (0.0444)  loss_rpn_box_reg: 0.0092 (0.0202)  time: 0.7394  data: 0.1303  max mem: 4703\n",
      "Epoch: [0]  [ 430/1500]  eta: 0:12:40  lr: 0.002158  loss: 0.5874 (0.8794)  loss_classifier: 0.0965 (0.1751)  loss_box_reg: 0.1044 (0.2050)  loss_mask: 0.3062 (0.4345)  loss_objectness: 0.0153 (0.0443)  loss_rpn_box_reg: 0.0065 (0.0206)  time: 0.7191  data: 0.1073  max mem: 4806\n",
      "Epoch: [0]  [ 440/1500]  eta: 0:12:35  lr: 0.002208  loss: 0.6087 (0.8792)  loss_classifier: 0.0958 (0.1748)  loss_box_reg: 0.1073 (0.2049)  loss_mask: 0.3330 (0.4342)  loss_objectness: 0.0239 (0.0445)  loss_rpn_box_reg: 0.0085 (0.0207)  time: 0.7793  data: 0.1500  max mem: 4806\n",
      "Epoch: [0]  [ 450/1500]  eta: 0:12:28  lr: 0.002258  loss: 0.7486 (0.8779)  loss_classifier: 0.1144 (0.1737)  loss_box_reg: 0.1403 (0.2045)  loss_mask: 0.4252 (0.4344)  loss_objectness: 0.0332 (0.0444)  loss_rpn_box_reg: 0.0177 (0.0209)  time: 0.7454  data: 0.1456  max mem: 4806\n",
      "Epoch: [0]  [ 460/1500]  eta: 0:12:21  lr: 0.002308  loss: 0.7012 (0.8733)  loss_classifier: 0.1252 (0.1726)  loss_box_reg: 0.1475 (0.2036)  loss_mask: 0.3176 (0.4322)  loss_objectness: 0.0332 (0.0442)  loss_rpn_box_reg: 0.0129 (0.0207)  time: 0.7159  data: 0.0905  max mem: 4806\n",
      "Epoch: [0]  [ 470/1500]  eta: 0:12:13  lr: 0.002358  loss: 0.5469 (0.8662)  loss_classifier: 0.1027 (0.1710)  loss_box_reg: 0.1104 (0.2022)  loss_mask: 0.2885 (0.4287)  loss_objectness: 0.0182 (0.0438)  loss_rpn_box_reg: 0.0065 (0.0205)  time: 0.7151  data: 0.0813  max mem: 4806\n",
      "Epoch: [0]  [ 480/1500]  eta: 0:12:06  lr: 0.002408  loss: 0.5200 (0.8596)  loss_classifier: 0.0771 (0.1689)  loss_box_reg: 0.1013 (0.2003)  loss_mask: 0.2578 (0.4270)  loss_objectness: 0.0109 (0.0431)  loss_rpn_box_reg: 0.0054 (0.0202)  time: 0.7113  data: 0.0901  max mem: 4806\n",
      "Epoch: [0]  [ 490/1500]  eta: 0:11:58  lr: 0.002458  loss: 0.5536 (0.8537)  loss_classifier: 0.0668 (0.1670)  loss_box_reg: 0.0859 (0.1984)  loss_mask: 0.3380 (0.4258)  loss_objectness: 0.0091 (0.0425)  loss_rpn_box_reg: 0.0054 (0.0201)  time: 0.6950  data: 0.0746  max mem: 4806\n",
      "Epoch: [0]  [ 500/1500]  eta: 0:11:52  lr: 0.002507  loss: 0.5943 (0.8511)  loss_classifier: 0.0799 (0.1662)  loss_box_reg: 0.0932 (0.1982)  loss_mask: 0.3623 (0.4242)  loss_objectness: 0.0091 (0.0422)  loss_rpn_box_reg: 0.0069 (0.0203)  time: 0.7020  data: 0.0831  max mem: 4806\n",
      "Epoch: [0]  [ 510/1500]  eta: 0:11:46  lr: 0.002557  loss: 0.6542 (0.8473)  loss_classifier: 0.0898 (0.1652)  loss_box_reg: 0.1179 (0.1968)  loss_mask: 0.3623 (0.4231)  loss_objectness: 0.0172 (0.0420)  loss_rpn_box_reg: 0.0069 (0.0202)  time: 0.7532  data: 0.0920  max mem: 4806\n",
      "Epoch: [0]  [ 520/1500]  eta: 0:11:38  lr: 0.002607  loss: 0.6542 (0.8441)  loss_classifier: 0.0839 (0.1642)  loss_box_reg: 0.1179 (0.1958)  loss_mask: 0.3679 (0.4223)  loss_objectness: 0.0201 (0.0417)  loss_rpn_box_reg: 0.0051 (0.0201)  time: 0.7386  data: 0.0956  max mem: 4806\n",
      "Epoch: [0]  [ 530/1500]  eta: 0:11:32  lr: 0.002657  loss: 0.5638 (0.8384)  loss_classifier: 0.0817 (0.1628)  loss_box_reg: 0.1103 (0.1942)  loss_mask: 0.2719 (0.4202)  loss_objectness: 0.0169 (0.0413)  loss_rpn_box_reg: 0.0048 (0.0199)  time: 0.7267  data: 0.0967  max mem: 4806\n",
      "Epoch: [0]  [ 540/1500]  eta: 0:11:24  lr: 0.002707  loss: 0.5977 (0.8354)  loss_classifier: 0.0756 (0.1616)  loss_box_reg: 0.1096 (0.1933)  loss_mask: 0.2952 (0.4196)  loss_objectness: 0.0160 (0.0408)  loss_rpn_box_reg: 0.0077 (0.0201)  time: 0.7055  data: 0.0812  max mem: 4806\n",
      "Epoch: [0]  [ 550/1500]  eta: 0:11:16  lr: 0.002757  loss: 0.6107 (0.8320)  loss_classifier: 0.1074 (0.1611)  loss_box_reg: 0.1195 (0.1927)  loss_mask: 0.2952 (0.4176)  loss_objectness: 0.0184 (0.0406)  loss_rpn_box_reg: 0.0115 (0.0200)  time: 0.6743  data: 0.0653  max mem: 4806\n",
      "Epoch: [0]  [ 560/1500]  eta: 0:11:09  lr: 0.002807  loss: 0.5871 (0.8276)  loss_classifier: 0.0967 (0.1599)  loss_box_reg: 0.1057 (0.1914)  loss_mask: 0.2833 (0.4161)  loss_objectness: 0.0197 (0.0403)  loss_rpn_box_reg: 0.0048 (0.0198)  time: 0.6891  data: 0.0830  max mem: 4806\n",
      "Epoch: [0]  [ 570/1500]  eta: 0:11:02  lr: 0.002857  loss: 0.5871 (0.8251)  loss_classifier: 0.0874 (0.1590)  loss_box_reg: 0.0940 (0.1897)  loss_mask: 0.3628 (0.4159)  loss_objectness: 0.0234 (0.0407)  loss_rpn_box_reg: 0.0049 (0.0198)  time: 0.7019  data: 0.1036  max mem: 4806\n",
      "Epoch: [0]  [ 580/1500]  eta: 0:10:56  lr: 0.002907  loss: 0.6084 (0.8228)  loss_classifier: 0.1058 (0.1586)  loss_box_reg: 0.0926 (0.1893)  loss_mask: 0.3594 (0.4143)  loss_objectness: 0.0275 (0.0406)  loss_rpn_box_reg: 0.0079 (0.0201)  time: 0.7519  data: 0.1214  max mem: 4806\n",
      "Epoch: [0]  [ 590/1500]  eta: 0:10:48  lr: 0.002957  loss: 0.6084 (0.8203)  loss_classifier: 0.1080 (0.1582)  loss_box_reg: 0.1008 (0.1883)  loss_mask: 0.3126 (0.4135)  loss_objectness: 0.0218 (0.0405)  loss_rpn_box_reg: 0.0077 (0.0199)  time: 0.7229  data: 0.1008  max mem: 4806\n",
      "Epoch: [0]  [ 600/1500]  eta: 0:10:41  lr: 0.003007  loss: 0.6901 (0.8179)  loss_classifier: 0.1055 (0.1571)  loss_box_reg: 0.1014 (0.1873)  loss_mask: 0.3654 (0.4131)  loss_objectness: 0.0256 (0.0403)  loss_rpn_box_reg: 0.0072 (0.0200)  time: 0.6918  data: 0.0726  max mem: 4806\n",
      "Epoch: [0]  [ 610/1500]  eta: 0:10:34  lr: 0.003057  loss: 0.7033 (0.8172)  loss_classifier: 0.1020 (0.1566)  loss_box_reg: 0.1127 (0.1865)  loss_mask: 0.4086 (0.4139)  loss_objectness: 0.0256 (0.0402)  loss_rpn_box_reg: 0.0130 (0.0200)  time: 0.7116  data: 0.0914  max mem: 4806\n",
      "Epoch: [0]  [ 620/1500]  eta: 0:10:26  lr: 0.003107  loss: 0.7033 (0.8152)  loss_classifier: 0.0999 (0.1558)  loss_box_reg: 0.1127 (0.1859)  loss_mask: 0.3753 (0.4128)  loss_objectness: 0.0216 (0.0404)  loss_rpn_box_reg: 0.0114 (0.0202)  time: 0.6785  data: 0.0733  max mem: 4806\n",
      "Epoch: [0]  [ 630/1500]  eta: 0:10:18  lr: 0.003157  loss: 0.5473 (0.8121)  loss_classifier: 0.0851 (0.1550)  loss_box_reg: 0.1226 (0.1851)  loss_mask: 0.2979 (0.4116)  loss_objectness: 0.0187 (0.0403)  loss_rpn_box_reg: 0.0060 (0.0201)  time: 0.6722  data: 0.0796  max mem: 4806\n",
      "Epoch: [0]  [ 640/1500]  eta: 0:10:11  lr: 0.003207  loss: 0.6247 (0.8114)  loss_classifier: 0.0931 (0.1543)  loss_box_reg: 0.1184 (0.1839)  loss_mask: 0.3690 (0.4128)  loss_objectness: 0.0187 (0.0403)  loss_rpn_box_reg: 0.0076 (0.0201)  time: 0.7144  data: 0.0937  max mem: 4806\n",
      "Epoch: [0]  [ 650/1500]  eta: 0:10:04  lr: 0.003257  loss: 0.6247 (0.8069)  loss_classifier: 0.0834 (0.1533)  loss_box_reg: 0.0692 (0.1826)  loss_mask: 0.3537 (0.4111)  loss_objectness: 0.0158 (0.0400)  loss_rpn_box_reg: 0.0073 (0.0199)  time: 0.7148  data: 0.0811  max mem: 4806\n",
      "Epoch: [0]  [ 660/1500]  eta: 0:09:57  lr: 0.003307  loss: 0.5024 (0.8046)  loss_classifier: 0.0776 (0.1528)  loss_box_reg: 0.0821 (0.1822)  loss_mask: 0.3059 (0.4099)  loss_objectness: 0.0181 (0.0398)  loss_rpn_box_reg: 0.0064 (0.0199)  time: 0.6999  data: 0.1002  max mem: 4806\n",
      "Epoch: [0]  [ 670/1500]  eta: 0:09:50  lr: 0.003357  loss: 0.5855 (0.8023)  loss_classifier: 0.0863 (0.1520)  loss_box_reg: 0.0896 (0.1815)  loss_mask: 0.3349 (0.4096)  loss_objectness: 0.0232 (0.0395)  loss_rpn_box_reg: 0.0074 (0.0197)  time: 0.7053  data: 0.1036  max mem: 4806\n",
      "Epoch: [0]  [ 680/1500]  eta: 0:09:43  lr: 0.003407  loss: 0.5613 (0.8002)  loss_classifier: 0.0694 (0.1510)  loss_box_reg: 0.0862 (0.1806)  loss_mask: 0.3604 (0.4089)  loss_objectness: 0.0234 (0.0399)  loss_rpn_box_reg: 0.0091 (0.0198)  time: 0.7109  data: 0.0901  max mem: 4806\n",
      "Epoch: [0]  [ 690/1500]  eta: 0:09:36  lr: 0.003457  loss: 0.5815 (0.7975)  loss_classifier: 0.0669 (0.1502)  loss_box_reg: 0.0894 (0.1799)  loss_mask: 0.3160 (0.4079)  loss_objectness: 0.0210 (0.0398)  loss_rpn_box_reg: 0.0078 (0.0198)  time: 0.7273  data: 0.0990  max mem: 4806\n",
      "Epoch: [0]  [ 700/1500]  eta: 0:09:29  lr: 0.003506  loss: 0.6611 (0.7970)  loss_classifier: 0.0741 (0.1497)  loss_box_reg: 0.0933 (0.1800)  loss_mask: 0.3473 (0.4076)  loss_objectness: 0.0194 (0.0398)  loss_rpn_box_reg: 0.0078 (0.0198)  time: 0.7168  data: 0.0947  max mem: 4806\n",
      "Epoch: [0]  [ 710/1500]  eta: 0:09:21  lr: 0.003556  loss: 0.7792 (0.7966)  loss_classifier: 0.1238 (0.1493)  loss_box_reg: 0.1255 (0.1801)  loss_mask: 0.3670 (0.4070)  loss_objectness: 0.0205 (0.0402)  loss_rpn_box_reg: 0.0187 (0.0200)  time: 0.6990  data: 0.0869  max mem: 4806\n",
      "Epoch: [0]  [ 720/1500]  eta: 0:09:14  lr: 0.003606  loss: 0.6545 (0.7942)  loss_classifier: 0.0924 (0.1484)  loss_box_reg: 0.1069 (0.1791)  loss_mask: 0.3670 (0.4062)  loss_objectness: 0.0313 (0.0405)  loss_rpn_box_reg: 0.0178 (0.0200)  time: 0.7066  data: 0.1016  max mem: 4806\n",
      "Epoch: [0]  [ 730/1500]  eta: 0:09:08  lr: 0.003656  loss: 0.6706 (0.7943)  loss_classifier: 0.0829 (0.1481)  loss_box_reg: 0.1319 (0.1792)  loss_mask: 0.3513 (0.4059)  loss_objectness: 0.0377 (0.0409)  loss_rpn_box_reg: 0.0144 (0.0201)  time: 0.7623  data: 0.1238  max mem: 4806\n",
      "Epoch: [0]  [ 740/1500]  eta: 0:09:01  lr: 0.003706  loss: 0.7649 (0.7935)  loss_classifier: 0.0957 (0.1478)  loss_box_reg: 0.1396 (0.1790)  loss_mask: 0.3896 (0.4059)  loss_objectness: 0.0281 (0.0407)  loss_rpn_box_reg: 0.0125 (0.0201)  time: 0.7487  data: 0.1026  max mem: 4806\n",
      "Epoch: [0]  [ 750/1500]  eta: 0:08:54  lr: 0.003756  loss: 0.8276 (0.7938)  loss_classifier: 0.1181 (0.1477)  loss_box_reg: 0.1396 (0.1791)  loss_mask: 0.3952 (0.4060)  loss_objectness: 0.0223 (0.0408)  loss_rpn_box_reg: 0.0122 (0.0201)  time: 0.7046  data: 0.0808  max mem: 4806\n",
      "Epoch: [0]  [ 760/1500]  eta: 0:08:47  lr: 0.003806  loss: 0.8276 (0.7939)  loss_classifier: 0.1457 (0.1478)  loss_box_reg: 0.1956 (0.1795)  loss_mask: 0.3564 (0.4057)  loss_objectness: 0.0250 (0.0407)  loss_rpn_box_reg: 0.0159 (0.0202)  time: 0.7246  data: 0.0871  max mem: 4806\n",
      "Epoch: [0]  [ 770/1500]  eta: 0:08:39  lr: 0.003856  loss: 0.6741 (0.7913)  loss_classifier: 0.1022 (0.1470)  loss_box_reg: 0.1351 (0.1786)  loss_mask: 0.3482 (0.4049)  loss_objectness: 0.0203 (0.0405)  loss_rpn_box_reg: 0.0133 (0.0204)  time: 0.6879  data: 0.0641  max mem: 4806\n",
      "Epoch: [0]  [ 780/1500]  eta: 0:08:32  lr: 0.003906  loss: 0.5840 (0.7902)  loss_classifier: 0.0830 (0.1466)  loss_box_reg: 0.0955 (0.1782)  loss_mask: 0.3483 (0.4046)  loss_objectness: 0.0214 (0.0403)  loss_rpn_box_reg: 0.0124 (0.0204)  time: 0.6662  data: 0.0699  max mem: 4806\n",
      "Epoch: [0]  [ 790/1500]  eta: 0:08:24  lr: 0.003956  loss: 0.6329 (0.7882)  loss_classifier: 0.0835 (0.1461)  loss_box_reg: 0.0842 (0.1775)  loss_mask: 0.3911 (0.4042)  loss_objectness: 0.0214 (0.0401)  loss_rpn_box_reg: 0.0099 (0.0203)  time: 0.6680  data: 0.0803  max mem: 4806\n",
      "Epoch: [0]  [ 800/1500]  eta: 0:08:16  lr: 0.004006  loss: 0.6329 (0.7876)  loss_classifier: 0.0913 (0.1459)  loss_box_reg: 0.0916 (0.1771)  loss_mask: 0.3740 (0.4042)  loss_objectness: 0.0233 (0.0401)  loss_rpn_box_reg: 0.0106 (0.0204)  time: 0.6492  data: 0.0760  max mem: 4806\n",
      "Epoch: [0]  [ 810/1500]  eta: 0:08:09  lr: 0.004056  loss: 0.6804 (0.7863)  loss_classifier: 0.1123 (0.1456)  loss_box_reg: 0.1111 (0.1770)  loss_mask: 0.3617 (0.4035)  loss_objectness: 0.0183 (0.0399)  loss_rpn_box_reg: 0.0139 (0.0203)  time: 0.6805  data: 0.0866  max mem: 4806\n",
      "Epoch: [0]  [ 820/1500]  eta: 0:08:02  lr: 0.004106  loss: 0.6903 (0.7866)  loss_classifier: 0.1151 (0.1457)  loss_box_reg: 0.1116 (0.1773)  loss_mask: 0.3666 (0.4034)  loss_objectness: 0.0253 (0.0399)  loss_rpn_box_reg: 0.0088 (0.0204)  time: 0.7145  data: 0.1051  max mem: 4806\n",
      "Epoch: [0]  [ 830/1500]  eta: 0:07:55  lr: 0.004156  loss: 0.5958 (0.7836)  loss_classifier: 0.0843 (0.1449)  loss_box_reg: 0.1059 (0.1765)  loss_mask: 0.3537 (0.4023)  loss_objectness: 0.0229 (0.0397)  loss_rpn_box_reg: 0.0082 (0.0202)  time: 0.7089  data: 0.1081  max mem: 4806\n",
      "Epoch: [0]  [ 840/1500]  eta: 0:07:49  lr: 0.004206  loss: 0.7311 (0.7851)  loss_classifier: 0.1090 (0.1450)  loss_box_reg: 0.1363 (0.1772)  loss_mask: 0.3581 (0.4029)  loss_objectness: 0.0137 (0.0398)  loss_rpn_box_reg: 0.0086 (0.0203)  time: 0.7493  data: 0.1055  max mem: 4806\n",
      "Epoch: [0]  [ 850/1500]  eta: 0:07:42  lr: 0.004256  loss: 0.7924 (0.7841)  loss_classifier: 0.1298 (0.1447)  loss_box_reg: 0.1428 (0.1766)  loss_mask: 0.3695 (0.4027)  loss_objectness: 0.0267 (0.0397)  loss_rpn_box_reg: 0.0102 (0.0204)  time: 0.7706  data: 0.0908  max mem: 4806\n",
      "Epoch: [0]  [ 860/1500]  eta: 0:07:35  lr: 0.004306  loss: 0.6277 (0.7840)  loss_classifier: 0.0995 (0.1443)  loss_box_reg: 0.1111 (0.1763)  loss_mask: 0.3605 (0.4023)  loss_objectness: 0.0320 (0.0407)  loss_rpn_box_reg: 0.0105 (0.0205)  time: 0.7240  data: 0.0955  max mem: 4806\n",
      "Epoch: [0]  [ 870/1500]  eta: 0:07:28  lr: 0.004356  loss: 0.6129 (0.7833)  loss_classifier: 0.0880 (0.1439)  loss_box_reg: 0.0886 (0.1758)  loss_mask: 0.3880 (0.4022)  loss_objectness: 0.0435 (0.0409)  loss_rpn_box_reg: 0.0173 (0.0205)  time: 0.7060  data: 0.1138  max mem: 4806\n",
      "Epoch: [0]  [ 880/1500]  eta: 0:07:21  lr: 0.004406  loss: 0.6800 (0.7830)  loss_classifier: 0.0856 (0.1436)  loss_box_reg: 0.0886 (0.1755)  loss_mask: 0.3660 (0.4023)  loss_objectness: 0.0424 (0.0410)  loss_rpn_box_reg: 0.0169 (0.0205)  time: 0.7411  data: 0.1292  max mem: 4806\n",
      "Epoch: [0]  [ 890/1500]  eta: 0:07:14  lr: 0.004456  loss: 0.7078 (0.7820)  loss_classifier: 0.0982 (0.1434)  loss_box_reg: 0.0809 (0.1753)  loss_mask: 0.3614 (0.4019)  loss_objectness: 0.0409 (0.0409)  loss_rpn_box_reg: 0.0134 (0.0205)  time: 0.7716  data: 0.1381  max mem: 4806\n",
      "Epoch: [0]  [ 900/1500]  eta: 0:07:07  lr: 0.004505  loss: 0.6398 (0.7803)  loss_classifier: 0.0835 (0.1430)  loss_box_reg: 0.0809 (0.1747)  loss_mask: 0.3321 (0.4012)  loss_objectness: 0.0302 (0.0410)  loss_rpn_box_reg: 0.0134 (0.0205)  time: 0.7152  data: 0.0963  max mem: 4806\n",
      "Epoch: [0]  [ 910/1500]  eta: 0:06:59  lr: 0.004555  loss: 0.5888 (0.7798)  loss_classifier: 0.0816 (0.1428)  loss_box_reg: 0.0873 (0.1745)  loss_mask: 0.3321 (0.4012)  loss_objectness: 0.0282 (0.0409)  loss_rpn_box_reg: 0.0060 (0.0205)  time: 0.6587  data: 0.0652  max mem: 4806\n",
      "Epoch: [0]  [ 920/1500]  eta: 0:06:52  lr: 0.004605  loss: 0.6186 (0.7787)  loss_classifier: 0.0818 (0.1422)  loss_box_reg: 0.1038 (0.1742)  loss_mask: 0.3529 (0.4007)  loss_objectness: 0.0233 (0.0409)  loss_rpn_box_reg: 0.0060 (0.0207)  time: 0.6931  data: 0.0646  max mem: 4806\n",
      "Epoch: [0]  [ 930/1500]  eta: 0:06:45  lr: 0.004655  loss: 0.6209 (0.7779)  loss_classifier: 0.0757 (0.1419)  loss_box_reg: 0.0895 (0.1738)  loss_mask: 0.3515 (0.4004)  loss_objectness: 0.0312 (0.0412)  loss_rpn_box_reg: 0.0114 (0.0206)  time: 0.7353  data: 0.0807  max mem: 4806\n",
      "Epoch: [0]  [ 940/1500]  eta: 0:06:38  lr: 0.004705  loss: 0.6209 (0.7765)  loss_classifier: 0.0758 (0.1416)  loss_box_reg: 0.0783 (0.1735)  loss_mask: 0.3191 (0.3993)  loss_objectness: 0.0393 (0.0413)  loss_rpn_box_reg: 0.0121 (0.0208)  time: 0.7225  data: 0.0877  max mem: 4806\n",
      "Epoch: [0]  [ 950/1500]  eta: 0:06:31  lr: 0.004755  loss: 0.5828 (0.7753)  loss_classifier: 0.0949 (0.1412)  loss_box_reg: 0.1017 (0.1729)  loss_mask: 0.3132 (0.3992)  loss_objectness: 0.0290 (0.0413)  loss_rpn_box_reg: 0.0104 (0.0207)  time: 0.6951  data: 0.0766  max mem: 4806\n",
      "Epoch: [0]  [ 960/1500]  eta: 0:06:24  lr: 0.004805  loss: 0.5828 (0.7736)  loss_classifier: 0.0931 (0.1408)  loss_box_reg: 0.1032 (0.1729)  loss_mask: 0.3685 (0.3983)  loss_objectness: 0.0214 (0.0410)  loss_rpn_box_reg: 0.0088 (0.0206)  time: 0.6945  data: 0.0751  max mem: 4806\n",
      "Epoch: [0]  [ 970/1500]  eta: 0:06:17  lr: 0.004855  loss: 0.6902 (0.7745)  loss_classifier: 0.1131 (0.1411)  loss_box_reg: 0.1956 (0.1734)  loss_mask: 0.3290 (0.3982)  loss_objectness: 0.0216 (0.0411)  loss_rpn_box_reg: 0.0177 (0.0207)  time: 0.7284  data: 0.0931  max mem: 4806\n",
      "Epoch: [0]  [ 980/1500]  eta: 0:06:10  lr: 0.004905  loss: 0.6902 (0.7737)  loss_classifier: 0.1241 (0.1408)  loss_box_reg: 0.1606 (0.1728)  loss_mask: 0.3727 (0.3982)  loss_objectness: 0.0269 (0.0412)  loss_rpn_box_reg: 0.0179 (0.0208)  time: 0.7381  data: 0.1037  max mem: 4806\n",
      "Epoch: [0]  [ 990/1500]  eta: 0:06:02  lr: 0.004955  loss: 0.6192 (0.7724)  loss_classifier: 0.0974 (0.1404)  loss_box_reg: 0.0965 (0.1724)  loss_mask: 0.3765 (0.3980)  loss_objectness: 0.0229 (0.0410)  loss_rpn_box_reg: 0.0076 (0.0207)  time: 0.6927  data: 0.0881  max mem: 4806\n",
      "Epoch: [0]  [1000/1500]  eta: 0:05:55  lr: 0.005000  loss: 0.6159 (0.7714)  loss_classifier: 0.0883 (0.1400)  loss_box_reg: 0.1022 (0.1720)  loss_mask: 0.3563 (0.3977)  loss_objectness: 0.0268 (0.0410)  loss_rpn_box_reg: 0.0084 (0.0207)  time: 0.6859  data: 0.0788  max mem: 4806\n",
      "Epoch: [0]  [1010/1500]  eta: 0:05:48  lr: 0.005000  loss: 0.6347 (0.7708)  loss_classifier: 0.0883 (0.1400)  loss_box_reg: 0.0977 (0.1719)  loss_mask: 0.3407 (0.3971)  loss_objectness: 0.0270 (0.0409)  loss_rpn_box_reg: 0.0125 (0.0209)  time: 0.7203  data: 0.0897  max mem: 4806\n",
      "Epoch: [0]  [1020/1500]  eta: 0:05:41  lr: 0.005000  loss: 0.6516 (0.7703)  loss_classifier: 0.1098 (0.1397)  loss_box_reg: 0.0984 (0.1719)  loss_mask: 0.3540 (0.3968)  loss_objectness: 0.0255 (0.0409)  loss_rpn_box_reg: 0.0055 (0.0210)  time: 0.7246  data: 0.0860  max mem: 4806\n",
      "Epoch: [0]  [1030/1500]  eta: 0:05:34  lr: 0.005000  loss: 0.5926 (0.7693)  loss_classifier: 0.0982 (0.1395)  loss_box_reg: 0.1243 (0.1718)  loss_mask: 0.3338 (0.3961)  loss_objectness: 0.0198 (0.0409)  loss_rpn_box_reg: 0.0124 (0.0210)  time: 0.7083  data: 0.0852  max mem: 4806\n",
      "Epoch: [0]  [1040/1500]  eta: 0:05:27  lr: 0.005000  loss: 0.5926 (0.7691)  loss_classifier: 0.0997 (0.1395)  loss_box_reg: 0.1257 (0.1716)  loss_mask: 0.3338 (0.3957)  loss_objectness: 0.0247 (0.0411)  loss_rpn_box_reg: 0.0115 (0.0211)  time: 0.7311  data: 0.1160  max mem: 4806\n",
      "Epoch: [0]  [1050/1500]  eta: 0:05:20  lr: 0.005000  loss: 0.6146 (0.7678)  loss_classifier: 0.0975 (0.1392)  loss_box_reg: 0.0948 (0.1712)  loss_mask: 0.3372 (0.3953)  loss_objectness: 0.0321 (0.0411)  loss_rpn_box_reg: 0.0115 (0.0211)  time: 0.6980  data: 0.0934  max mem: 4806\n",
      "Epoch: [0]  [1060/1500]  eta: 0:05:12  lr: 0.005000  loss: 0.6146 (0.7669)  loss_classifier: 0.0859 (0.1389)  loss_box_reg: 0.1303 (0.1710)  loss_mask: 0.3478 (0.3950)  loss_objectness: 0.0265 (0.0410)  loss_rpn_box_reg: 0.0120 (0.0210)  time: 0.6308  data: 0.0533  max mem: 4806\n",
      "Epoch: [0]  [1070/1500]  eta: 0:05:05  lr: 0.005000  loss: 0.5919 (0.7663)  loss_classifier: 0.0816 (0.1387)  loss_box_reg: 0.1089 (0.1709)  loss_mask: 0.4026 (0.3948)  loss_objectness: 0.0178 (0.0409)  loss_rpn_box_reg: 0.0114 (0.0211)  time: 0.6775  data: 0.0611  max mem: 4806\n",
      "Epoch: [0]  [1080/1500]  eta: 0:04:58  lr: 0.005000  loss: 0.6049 (0.7656)  loss_classifier: 0.0932 (0.1385)  loss_box_reg: 0.1088 (0.1706)  loss_mask: 0.3423 (0.3944)  loss_objectness: 0.0206 (0.0409)  loss_rpn_box_reg: 0.0169 (0.0212)  time: 0.7374  data: 0.0991  max mem: 4806\n",
      "Epoch: [0]  [1090/1500]  eta: 0:04:51  lr: 0.005000  loss: 0.6428 (0.7644)  loss_classifier: 0.1078 (0.1382)  loss_box_reg: 0.1082 (0.1702)  loss_mask: 0.3310 (0.3938)  loss_objectness: 0.0358 (0.0410)  loss_rpn_box_reg: 0.0143 (0.0212)  time: 0.7196  data: 0.1115  max mem: 4806\n",
      "Epoch: [0]  [1100/1500]  eta: 0:04:44  lr: 0.005000  loss: 0.5830 (0.7629)  loss_classifier: 0.0833 (0.1378)  loss_box_reg: 0.0968 (0.1696)  loss_mask: 0.3426 (0.3935)  loss_objectness: 0.0333 (0.0409)  loss_rpn_box_reg: 0.0061 (0.0211)  time: 0.6602  data: 0.0766  max mem: 4806\n",
      "Epoch: [0]  [1110/1500]  eta: 0:04:36  lr: 0.005000  loss: 0.5603 (0.7626)  loss_classifier: 0.0801 (0.1378)  loss_box_reg: 0.0964 (0.1697)  loss_mask: 0.3665 (0.3932)  loss_objectness: 0.0264 (0.0409)  loss_rpn_box_reg: 0.0053 (0.0210)  time: 0.6676  data: 0.0661  max mem: 4806\n",
      "Epoch: [0]  [1120/1500]  eta: 0:04:29  lr: 0.005000  loss: 0.6085 (0.7612)  loss_classifier: 0.0993 (0.1374)  loss_box_reg: 0.1075 (0.1692)  loss_mask: 0.3755 (0.3929)  loss_objectness: 0.0257 (0.0407)  loss_rpn_box_reg: 0.0080 (0.0210)  time: 0.6994  data: 0.0808  max mem: 4806\n",
      "Epoch: [0]  [1130/1500]  eta: 0:04:22  lr: 0.005000  loss: 0.6777 (0.7611)  loss_classifier: 0.0809 (0.1372)  loss_box_reg: 0.1075 (0.1691)  loss_mask: 0.3985 (0.3932)  loss_objectness: 0.0195 (0.0407)  loss_rpn_box_reg: 0.0078 (0.0209)  time: 0.7177  data: 0.0928  max mem: 4806\n",
      "Epoch: [0]  [1140/1500]  eta: 0:04:15  lr: 0.005000  loss: 0.6899 (0.7607)  loss_classifier: 0.0799 (0.1370)  loss_box_reg: 0.0977 (0.1689)  loss_mask: 0.3907 (0.3934)  loss_objectness: 0.0195 (0.0406)  loss_rpn_box_reg: 0.0081 (0.0209)  time: 0.7435  data: 0.0953  max mem: 4806\n",
      "Epoch: [0]  [1150/1500]  eta: 0:04:08  lr: 0.005000  loss: 0.6818 (0.7606)  loss_classifier: 0.0894 (0.1368)  loss_box_reg: 0.1110 (0.1690)  loss_mask: 0.3500 (0.3933)  loss_objectness: 0.0296 (0.0406)  loss_rpn_box_reg: 0.0146 (0.0209)  time: 0.6991  data: 0.0814  max mem: 4806\n",
      "Epoch: [0]  [1160/1500]  eta: 0:04:01  lr: 0.005000  loss: 0.6710 (0.7603)  loss_classifier: 0.1109 (0.1366)  loss_box_reg: 0.1472 (0.1689)  loss_mask: 0.3360 (0.3931)  loss_objectness: 0.0308 (0.0406)  loss_rpn_box_reg: 0.0093 (0.0210)  time: 0.6725  data: 0.0702  max mem: 4806\n",
      "Epoch: [0]  [1170/1500]  eta: 0:03:54  lr: 0.005000  loss: 0.6342 (0.7593)  loss_classifier: 0.0946 (0.1363)  loss_box_reg: 0.1047 (0.1686)  loss_mask: 0.3530 (0.3930)  loss_objectness: 0.0279 (0.0405)  loss_rpn_box_reg: 0.0077 (0.0209)  time: 0.6813  data: 0.0794  max mem: 4806\n",
      "Epoch: [0]  [1180/1500]  eta: 0:03:47  lr: 0.005000  loss: 0.6095 (0.7584)  loss_classifier: 0.0738 (0.1360)  loss_box_reg: 0.0948 (0.1682)  loss_mask: 0.3882 (0.3930)  loss_objectness: 0.0171 (0.0403)  loss_rpn_box_reg: 0.0077 (0.0209)  time: 0.6927  data: 0.0915  max mem: 4806\n",
      "Epoch: [0]  [1190/1500]  eta: 0:03:39  lr: 0.005000  loss: 0.6307 (0.7585)  loss_classifier: 0.0888 (0.1359)  loss_box_reg: 0.1075 (0.1683)  loss_mask: 0.3883 (0.3929)  loss_objectness: 0.0154 (0.0404)  loss_rpn_box_reg: 0.0065 (0.0210)  time: 0.6972  data: 0.0903  max mem: 4806\n",
      "Epoch: [0]  [1200/1500]  eta: 0:03:32  lr: 0.005000  loss: 0.6711 (0.7590)  loss_classifier: 0.1032 (0.1359)  loss_box_reg: 0.1110 (0.1681)  loss_mask: 0.4006 (0.3931)  loss_objectness: 0.0275 (0.0406)  loss_rpn_box_reg: 0.0106 (0.0212)  time: 0.7051  data: 0.0933  max mem: 4806\n",
      "Epoch: [0]  [1210/1500]  eta: 0:03:25  lr: 0.005000  loss: 0.6667 (0.7582)  loss_classifier: 0.1092 (0.1357)  loss_box_reg: 0.1069 (0.1678)  loss_mask: 0.3529 (0.3927)  loss_objectness: 0.0410 (0.0407)  loss_rpn_box_reg: 0.0112 (0.0212)  time: 0.7166  data: 0.1017  max mem: 4806\n",
      "Epoch: [0]  [1220/1500]  eta: 0:03:18  lr: 0.005000  loss: 0.6226 (0.7585)  loss_classifier: 0.0817 (0.1354)  loss_box_reg: 0.0939 (0.1674)  loss_mask: 0.3529 (0.3934)  loss_objectness: 0.0410 (0.0410)  loss_rpn_box_reg: 0.0143 (0.0212)  time: 0.7008  data: 0.0876  max mem: 4806\n",
      "Epoch: [0]  [1230/1500]  eta: 0:03:11  lr: 0.005000  loss: 0.5259 (0.7571)  loss_classifier: 0.0770 (0.1351)  loss_box_reg: 0.0811 (0.1671)  loss_mask: 0.3705 (0.3929)  loss_objectness: 0.0263 (0.0408)  loss_rpn_box_reg: 0.0155 (0.0212)  time: 0.7401  data: 0.0985  max mem: 4806\n",
      "Epoch: [0]  [1240/1500]  eta: 0:03:04  lr: 0.005000  loss: 0.5401 (0.7561)  loss_classifier: 0.0737 (0.1348)  loss_box_reg: 0.0855 (0.1665)  loss_mask: 0.3741 (0.3929)  loss_objectness: 0.0175 (0.0409)  loss_rpn_box_reg: 0.0074 (0.0211)  time: 0.7324  data: 0.0946  max mem: 4806\n",
      "Epoch: [0]  [1250/1500]  eta: 0:02:57  lr: 0.005000  loss: 0.5022 (0.7541)  loss_classifier: 0.0730 (0.1343)  loss_box_reg: 0.0787 (0.1658)  loss_mask: 0.3326 (0.3922)  loss_objectness: 0.0191 (0.0407)  loss_rpn_box_reg: 0.0048 (0.0210)  time: 0.6748  data: 0.0652  max mem: 4806\n",
      "Epoch: [0]  [1260/1500]  eta: 0:02:50  lr: 0.005000  loss: 0.4970 (0.7533)  loss_classifier: 0.0769 (0.1343)  loss_box_reg: 0.0871 (0.1656)  loss_mask: 0.3111 (0.3916)  loss_objectness: 0.0232 (0.0407)  loss_rpn_box_reg: 0.0064 (0.0211)  time: 0.6990  data: 0.0955  max mem: 4806\n",
      "Epoch: [0]  [1270/1500]  eta: 0:02:43  lr: 0.005000  loss: 0.5432 (0.7522)  loss_classifier: 0.1043 (0.1340)  loss_box_reg: 0.1128 (0.1651)  loss_mask: 0.3465 (0.3914)  loss_objectness: 0.0202 (0.0407)  loss_rpn_box_reg: 0.0101 (0.0210)  time: 0.7349  data: 0.1074  max mem: 4806\n",
      "Epoch: [0]  [1280/1500]  eta: 0:02:36  lr: 0.005000  loss: 0.6704 (0.7526)  loss_classifier: 0.1043 (0.1341)  loss_box_reg: 0.1128 (0.1653)  loss_mask: 0.3870 (0.3914)  loss_objectness: 0.0203 (0.0407)  loss_rpn_box_reg: 0.0108 (0.0211)  time: 0.7256  data: 0.0898  max mem: 4806\n",
      "Epoch: [0]  [1290/1500]  eta: 0:02:29  lr: 0.005000  loss: 0.6591 (0.7521)  loss_classifier: 0.1238 (0.1341)  loss_box_reg: 0.1524 (0.1654)  loss_mask: 0.3185 (0.3909)  loss_objectness: 0.0203 (0.0406)  loss_rpn_box_reg: 0.0161 (0.0211)  time: 0.7047  data: 0.0900  max mem: 4806\n",
      "Epoch: [0]  [1300/1500]  eta: 0:02:21  lr: 0.005000  loss: 0.6052 (0.7515)  loss_classifier: 0.1073 (0.1341)  loss_box_reg: 0.1270 (0.1653)  loss_mask: 0.3185 (0.3906)  loss_objectness: 0.0187 (0.0406)  loss_rpn_box_reg: 0.0075 (0.0210)  time: 0.6918  data: 0.0917  max mem: 4806\n",
      "Epoch: [0]  [1310/1500]  eta: 0:02:14  lr: 0.005000  loss: 0.6353 (0.7520)  loss_classifier: 0.0983 (0.1342)  loss_box_reg: 0.1271 (0.1656)  loss_mask: 0.3443 (0.3906)  loss_objectness: 0.0195 (0.0406)  loss_rpn_box_reg: 0.0077 (0.0211)  time: 0.7014  data: 0.1022  max mem: 4806\n",
      "Epoch: [0]  [1320/1500]  eta: 0:02:07  lr: 0.005000  loss: 0.6100 (0.7506)  loss_classifier: 0.0886 (0.1340)  loss_box_reg: 0.0919 (0.1652)  loss_mask: 0.3392 (0.3899)  loss_objectness: 0.0183 (0.0404)  loss_rpn_box_reg: 0.0086 (0.0211)  time: 0.7330  data: 0.1149  max mem: 4806\n",
      "Epoch: [0]  [1330/1500]  eta: 0:02:00  lr: 0.005000  loss: 0.6171 (0.7511)  loss_classifier: 0.0866 (0.1338)  loss_box_reg: 0.0919 (0.1651)  loss_mask: 0.3392 (0.3908)  loss_objectness: 0.0163 (0.0403)  loss_rpn_box_reg: 0.0086 (0.0212)  time: 0.7319  data: 0.1035  max mem: 4806\n",
      "Epoch: [0]  [1340/1500]  eta: 0:01:53  lr: 0.005000  loss: 0.6171 (0.7505)  loss_classifier: 0.0778 (0.1334)  loss_box_reg: 0.1092 (0.1646)  loss_mask: 0.4133 (0.3911)  loss_objectness: 0.0163 (0.0402)  loss_rpn_box_reg: 0.0047 (0.0211)  time: 0.7039  data: 0.0839  max mem: 4806\n",
      "Epoch: [0]  [1350/1500]  eta: 0:01:46  lr: 0.005000  loss: 0.5928 (0.7491)  loss_classifier: 0.0735 (0.1330)  loss_box_reg: 0.0954 (0.1641)  loss_mask: 0.3951 (0.3910)  loss_objectness: 0.0130 (0.0400)  loss_rpn_box_reg: 0.0059 (0.0210)  time: 0.7225  data: 0.0799  max mem: 4806\n",
      "Epoch: [0]  [1360/1500]  eta: 0:01:39  lr: 0.005000  loss: 0.5496 (0.7488)  loss_classifier: 0.0739 (0.1329)  loss_box_reg: 0.0954 (0.1639)  loss_mask: 0.3439 (0.3911)  loss_objectness: 0.0159 (0.0400)  loss_rpn_box_reg: 0.0066 (0.0209)  time: 0.7127  data: 0.0787  max mem: 4806\n",
      "Epoch: [0]  [1370/1500]  eta: 0:01:32  lr: 0.005000  loss: 0.6396 (0.7487)  loss_classifier: 0.0974 (0.1327)  loss_box_reg: 0.0935 (0.1636)  loss_mask: 0.3439 (0.3912)  loss_objectness: 0.0266 (0.0404)  loss_rpn_box_reg: 0.0069 (0.0209)  time: 0.6950  data: 0.0908  max mem: 4806\n",
      "Epoch: [0]  [1380/1500]  eta: 0:01:25  lr: 0.005000  loss: 0.6251 (0.7488)  loss_classifier: 0.0926 (0.1326)  loss_box_reg: 0.0927 (0.1636)  loss_mask: 0.3495 (0.3914)  loss_objectness: 0.0296 (0.0405)  loss_rpn_box_reg: 0.0071 (0.0209)  time: 0.7307  data: 0.1083  max mem: 4806\n",
      "Epoch: [0]  [1390/1500]  eta: 0:01:18  lr: 0.005000  loss: 0.7355 (0.7497)  loss_classifier: 0.1143 (0.1327)  loss_box_reg: 0.1200 (0.1637)  loss_mask: 0.3888 (0.3916)  loss_objectness: 0.0392 (0.0407)  loss_rpn_box_reg: 0.0152 (0.0210)  time: 0.7565  data: 0.1329  max mem: 4806\n",
      "Epoch: [0]  [1400/1500]  eta: 0:01:11  lr: 0.005000  loss: 0.7070 (0.7498)  loss_classifier: 0.1143 (0.1324)  loss_box_reg: 0.1196 (0.1634)  loss_mask: 0.4241 (0.3917)  loss_objectness: 0.0562 (0.0412)  loss_rpn_box_reg: 0.0198 (0.0211)  time: 0.7350  data: 0.1180  max mem: 4806\n",
      "Epoch: [0]  [1410/1500]  eta: 0:01:03  lr: 0.005000  loss: 0.6195 (0.7493)  loss_classifier: 0.0679 (0.1323)  loss_box_reg: 0.1146 (0.1633)  loss_mask: 0.3844 (0.3915)  loss_objectness: 0.0355 (0.0411)  loss_rpn_box_reg: 0.0115 (0.0211)  time: 0.7278  data: 0.0870  max mem: 4806\n",
      "Epoch: [0]  [1420/1500]  eta: 0:00:56  lr: 0.005000  loss: 0.6181 (0.7487)  loss_classifier: 0.0997 (0.1321)  loss_box_reg: 0.1194 (0.1630)  loss_mask: 0.3498 (0.3914)  loss_objectness: 0.0311 (0.0411)  loss_rpn_box_reg: 0.0105 (0.0211)  time: 0.7127  data: 0.0778  max mem: 4806\n",
      "Epoch: [0]  [1430/1500]  eta: 0:00:49  lr: 0.005000  loss: 0.6882 (0.7485)  loss_classifier: 0.1034 (0.1319)  loss_box_reg: 0.1194 (0.1628)  loss_mask: 0.3888 (0.3914)  loss_objectness: 0.0293 (0.0413)  loss_rpn_box_reg: 0.0097 (0.0211)  time: 0.6965  data: 0.0831  max mem: 4806\n",
      "Epoch: [0]  [1440/1500]  eta: 0:00:42  lr: 0.005000  loss: 0.7480 (0.7496)  loss_classifier: 0.1151 (0.1321)  loss_box_reg: 0.1424 (0.1632)  loss_mask: 0.4039 (0.3913)  loss_objectness: 0.0384 (0.0417)  loss_rpn_box_reg: 0.0123 (0.0212)  time: 0.7509  data: 0.1298  max mem: 4806\n",
      "Epoch: [0]  [1450/1500]  eta: 0:00:35  lr: 0.005000  loss: 0.7549 (0.7494)  loss_classifier: 0.1305 (0.1320)  loss_box_reg: 0.1315 (0.1629)  loss_mask: 0.4070 (0.3915)  loss_objectness: 0.0412 (0.0418)  loss_rpn_box_reg: 0.0123 (0.0212)  time: 0.7257  data: 0.1191  max mem: 4806\n",
      "Epoch: [0]  [1460/1500]  eta: 0:00:28  lr: 0.005000  loss: 0.7207 (0.7495)  loss_classifier: 0.1062 (0.1320)  loss_box_reg: 0.1065 (0.1628)  loss_mask: 0.4225 (0.3917)  loss_objectness: 0.0412 (0.0418)  loss_rpn_box_reg: 0.0080 (0.0212)  time: 0.7075  data: 0.1174  max mem: 4806\n",
      "Epoch: [0]  [1470/1500]  eta: 0:00:21  lr: 0.005000  loss: 0.6906 (0.7502)  loss_classifier: 0.1062 (0.1320)  loss_box_reg: 0.0896 (0.1629)  loss_mask: 0.4066 (0.3918)  loss_objectness: 0.0336 (0.0421)  loss_rpn_box_reg: 0.0146 (0.0214)  time: 0.7307  data: 0.1343  max mem: 4806\n",
      "Epoch: [0]  [1480/1500]  eta: 0:00:14  lr: 0.005000  loss: 0.6152 (0.7496)  loss_classifier: 0.0930 (0.1319)  loss_box_reg: 0.0904 (0.1628)  loss_mask: 0.3431 (0.3914)  loss_objectness: 0.0339 (0.0422)  loss_rpn_box_reg: 0.0185 (0.0214)  time: 0.7152  data: 0.1114  max mem: 4806\n",
      "Epoch: [0]  [1490/1500]  eta: 0:00:07  lr: 0.005000  loss: 0.5760 (0.7488)  loss_classifier: 0.0921 (0.1317)  loss_box_reg: 0.1034 (0.1625)  loss_mask: 0.3171 (0.3913)  loss_objectness: 0.0250 (0.0421)  loss_rpn_box_reg: 0.0088 (0.0213)  time: 0.7088  data: 0.0932  max mem: 4806\n",
      "Epoch: [0]  [1499/1500]  eta: 0:00:00  lr: 0.005000  loss: 0.5513 (0.7483)  loss_classifier: 0.0854 (0.1315)  loss_box_reg: 0.1034 (0.1624)  loss_mask: 0.3155 (0.3909)  loss_objectness: 0.0142 (0.0421)  loss_rpn_box_reg: 0.0079 (0.0214)  time: 0.6987  data: 0.0691  max mem: 4806\n",
      "Epoch: [0] Total time: 0:17:46 (0.7111 s / it)\n",
      "creating index...\n",
      "index created!\n",
      "Test:  [  0/800]  eta: 0:03:00  model_time: 0.1960 (0.1960)  evaluator_time: 0.0171 (0.0171)  time: 0.2261  data: 0.0122  max mem: 4806\n",
      "Test:  [100/800]  eta: 0:02:24  model_time: 0.1651 (0.1552)  evaluator_time: 0.0273 (0.0325)  time: 0.2121  data: 0.0179  max mem: 4806\n",
      "Test:  [200/800]  eta: 0:02:01  model_time: 0.1502 (0.1535)  evaluator_time: 0.0192 (0.0304)  time: 0.2118  data: 0.0187  max mem: 4806\n",
      "Test:  [300/800]  eta: 0:01:43  model_time: 0.1524 (0.1546)  evaluator_time: 0.0260 (0.0325)  time: 0.2090  data: 0.0177  max mem: 4806\n",
      "Test:  [400/800]  eta: 0:01:23  model_time: 0.1586 (0.1544)  evaluator_time: 0.0226 (0.0339)  time: 0.2338  data: 0.0216  max mem: 4806\n",
      "Test:  [500/800]  eta: 0:01:01  model_time: 0.1451 (0.1533)  evaluator_time: 0.0186 (0.0332)  time: 0.2091  data: 0.0155  max mem: 4806\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "ignored",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-15-16230b13c661>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      9\u001b[0m     \u001b[0mlr_scheduler\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[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m     \u001b[0;31m# evaluate on the test dataset\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 11\u001b[0;31m     \u001b[0mres\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata_loader_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/autograd/grad_mode.py\u001b[0m in \u001b[0;36mdecorate_context\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     24\u001b[0m         \u001b[0;32mdef\u001b[0m \u001b[0mdecorate_context\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[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     25\u001b[0m             \u001b[0;32mwith\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__class__\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---> 26\u001b[0;31m                 \u001b[0;32mreturn\u001b[0m \u001b[0mfunc\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     27\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mcast\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mF\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdecorate_context\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     28\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/content/engine.py\u001b[0m in \u001b[0;36mevaluate\u001b[0;34m(model, data_loader, device)\u001b[0m\n\u001b[1;32m     87\u001b[0m         \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msynchronize\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     88\u001b[0m         \u001b[0mmodel_time\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\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---> 89\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimages\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     90\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     91\u001b[0m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mv\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mto\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcpu_device\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mv\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\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[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    726\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--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\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/usr/local/lib/python3.6/dist-packages/torchvision/models/detection/generalized_rcnn.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, images, targets)\u001b[0m\n\u001b[1;32m     97\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTensor\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     98\u001b[0m             \u001b[0mfeatures\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mOrderedDict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'0'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeatures\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---> 99\u001b[0;31m         \u001b[0mproposals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mproposal_losses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrpn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimages\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeatures\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtargets\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    100\u001b[0m         \u001b[0mdetections\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdetector_losses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mroi_heads\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mproposals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimages\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mimage_sizes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtargets\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    101\u001b[0m         \u001b[0mdetections\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpostprocess\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdetections\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimages\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mimage_sizes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moriginal_image_sizes\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    726\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--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\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/usr/local/lib/python3.6/dist-packages/torchvision/models/detection/rpn.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, images, features, targets)\u001b[0m\n\u001b[1;32m    330\u001b[0m         \u001b[0mfeatures\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\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    331\u001b[0m         \u001b[0mobjectness\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred_bbox_deltas\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfeatures\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 332\u001b[0;31m         \u001b[0manchors\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0manchor_generator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimages\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeatures\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    333\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    334\u001b[0m         \u001b[0mnum_images\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0manchors\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *input, **kwargs)\u001b[0m\n\u001b[1;32m    725\u001b[0m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_slow_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    726\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--> 727\u001b[0;31m             \u001b[0mresult\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0minput\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    728\u001b[0m         for hook in itertools.chain(\n\u001b[1;32m    729\u001b[0m                 \u001b[0m_global_forward_hooks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\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/usr/local/lib/python3.6/dist-packages/torchvision/models/detection/anchor_utils.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, image_list, feature_maps)\u001b[0m\n\u001b[1;32m    145\u001b[0m         \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfeature_maps\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_maps\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    146\u001b[0m         strides = [[torch.tensor(image_size[0] // g[0], dtype=torch.int64, device=device),\n\u001b[0;32m--> 147\u001b[0;31m                     torch.tensor(image_size[1] // g[1], dtype=torch.int64, device=device)] for g in grid_sizes]\n\u001b[0m\u001b[1;32m    148\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_cell_anchors\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    149\u001b[0m         \u001b[0manchors_over_all_feature_maps\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcached_grid_anchors\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgrid_sizes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.6/dist-packages/torchvision/models/detection/anchor_utils.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m    145\u001b[0m         \u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfeature_maps\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_maps\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdevice\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    146\u001b[0m         strides = [[torch.tensor(image_size[0] // g[0], dtype=torch.int64, device=device),\n\u001b[0;32m--> 147\u001b[0;31m                     torch.tensor(image_size[1] // g[1], dtype=torch.int64, device=device)] for g in grid_sizes]\n\u001b[0m\u001b[1;32m    148\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_cell_anchors\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdtype\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    149\u001b[0m         \u001b[0manchors_over_all_feature_maps\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcached_grid_anchors\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgrid_sizes\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\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: "
     ]
    }
   ],
   "source": [
    "num_epochs = 5\n",
    "\n",
    "trn_history = []\n",
    "for epoch in range(num_epochs):\n",
    "    # train for one epoch, printing every 10 iterations\n",
    "    res = train_one_epoch(model, optimizer, data_loader, device, epoch, print_freq=10)\n",
    "    trn_history.append(res)\n",
    "    # update the learning rate\n",
    "    lr_scheduler.step()\n",
    "    # evaluate on the test dataset\n",
    "    res = evaluate(model, data_loader_test, device=device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D>]"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.title('Training Loss')\n",
    "losses = [np.mean(list(trn_history[i].meters['loss'].deque)) for i in range(len(trn_history))]\n",
    "plt.plot(losses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "k=408"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "409"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "k=423"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUoAAAEICAYAAAA5ub1iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAeg0lEQVR4nO3dfZQldX3n8fen6j51z/MwOIwzCKi4Pq1BMqsQ3cRITBTdwJ4Y14eNxEN2srvuri5mDWo05mFdTc5R455EwwluSDQ+xMTAGo0hgHlckVF0FBAcjMiM4MA89fTT7XurvvtH/Xq4dLqpHqZ7+nb4vM6p01W/X9Wt7723+zO/qrq3RhGBmZktLFvpAszMhp2D0syshoPSzKyGg9LMrIaD0syshoPSzKyGg3KVkvROSR9Z6TrMHgsclENM0qsl7ZY0Luk+SZ+T9PwhqOtxkj4m6XuSjkr6e0nPfYT1N0r6sKT7JR2TdJekK09lzQvUdZ6kL0uaTD/Pe4R1nybpxvR890r6t3P6L5L0zfRYN0k6a6Bvu6RrJR2StE/Sf5yz7QslfUXSmKRvS9o10CdJb5P03dT/cUnrB/pfIekf0n6/ME/duaRfT+/VMUm3Str4KF+yxywH5ZCSdAXwfuBdwFbgCcDvAJesZF3JWuAW4AeBzcA1wJ9LWrvA+u9L2zwN2AD8JLB3KQuS1DjB9VvAtcBHgE1Uz+Ha1D7fY18LfIbq+e4CPiLpKal/C/CnwNtT/27gEwMP8RHgH6nex5cC75L0o2nbJvBp4HepXpt/B7xX0g+kbV8L/AzwPODxwAjwvwce+xDV78m7F3iqvwL8EHAhsD491nTd62NzRISnIZuo/mDGgZ9+hHXeCXxkYPmPgfuBo8DfAM8Y6LsYuB04BuwHfiG1b6H64z9C9Qf3t0D2KGseA35wgb5vAJc+wrbPAK5PNXwfeGtqb1OFwPfS9H6gnfpeAOwDfjE97z+k+of/SuBu4CDwSWDzAvv88fRaaKDtu8CL51n3men9GFz3L4FfS/O7gH8Y6FsDTAFPpfoHIoDTB/qvAv4wzW9N/aMD/bcAr0rznwL+x0DfD1EF3eicGn8O+MKctk2p7iet9O/0ap88ohxOFwIdqpHGYn0OOBd4HPAV4KMDfVcDPx8R66j+6G9M7W+iCpvTqf5g30r1R4uk35H0O4vZcTpkbbHwKPGLwP+U9DpJ587Zdh3wV8BfUI2YngzckLrfBlwAnAf8APAc4JcGNj+DagR3FlVY/VfgUuBH0mMdBn57YF97JL06LT4D2BMpUZI9qX1RT5vqtZx9rK/NdkTEBFVYPyOtx8DPh20bEd8HPga8Lh0mX5iez9/NWX9wvk31Xtf5l0AfeHk67XGXpNcv7unZw6x0Unv6pxPwGuD+mnXeycCIck7fRqrA25CWvwv8PLB+znq/SnVI+eSTqHU98HXgLY+wzghVCH8Z6FEF6ktS36uAWxfY7m7g4oHlnwC+k+ZfAMwAnYH+O4CLBpa3pf015nnstwMfn9P2UeCd86zbBL4NvDnN/3ja9+dT/9XAu+ds8/fAz6b5v6M6XO4A51ONnO8cWPffUI2k+2n6DwN9PwfcBZxNdaRxXXpvL5yzv/lGlK9O616d3oNnAQ8AL1rp3/HVNnlEOZwOAlsWe94tjUTeLeluSWPAd1LXlvTzp6gOv++R9Ndp1ALwm1Sh9ZfpIsIJXWCRNAL8X+CLEfG/FlovIqYi4l0R8YPAaVSHxH8saTNwJlUgzufxwD0Dy/ektlkPRMTg+bazgE9LOiLpCFVwFlSj5bnGqUJ+0Hqq0xNz6+9RjVRfSnWY/6b0HPYt8rFeA5wD3At8kOqc5T4ASU8FPk51LrJFNQp9s6SXpm0/TDXi/AJwG3BTap/d9yOZSj9/Nb0He9K+Ll7EtjbAQTmc/h/QpfrjXIxXU13k+TGqUcfZqV0AEXFLRFxCdVj+Z1R/5ETEsYh4U0Q8keoCyxWSLlrMDiW102PtoxqtLkpEjFFdoFrDQ+HxxAVW/x5V+M16Qmo7/nBz1r+XaqS6cWDqRMT+eR77NuBZkgYPa5+V2uere09E/EhEnBYRP5Fq/tLAY81efEHSGuBJs48VEfdExMsi4vSIeC7VP2Cz2z4TuCsiPh8RZUTcCfw58JK0bRkRvxwRZ0fEjvSY+9NUZ89s+YNPZRHb2RwOyiEUEUeBdwC/LelSSaOSmpJeIuk35tlkHVWwHgRGqYIIqK7uSnqNpA1pZDQGlKnvZZKenMLiKNXoq6yrL12p/RTViOWyiHjEbSS9XdK/SrV0gDdQXUC6k+pi0jZJb5TUlrRu4KNGHwN+SdLp6cryO6hGYwv5ENW50LPSfk+XtNCnBL6Qnu9/S/v9L6n9xvlWlvQsSZ30XvwC1WH976fuTwPPlPRT6fm9g+r85zfTtk9Lz6sl6d9THbq/N217K3Bu+oiQJD0JeBkp5CRtlvSk1Pf0tN2vzr7m6WiiAzSALNXYBIiIu6ku0L0tPcenAa9Mr7mdiJU+9ve08ER1yLYbmKA65Ptz4IdS3ztJ5yiprqxeS3Wodw/VYVxQXRhpUV0oOUwVkrcAz0/b/Xeqw/QJqpHh2wf2/SHgQwvU9SPp8SepDjtnp3+9wPq/RHXle4zq/NwXZp9H6n8m1QWcw+l5XpnaO8AHgPvS9AHSOUnSVe85+8mAK6gC+BjVIf27BvpvA14zsPxsqvOmU1QXwJ490PdW4HMDy7+Z6hununD25Dn7/jHgm+mxvgCcPdD3RqpzgxNU5yt3ztn2Fen1OZbeh/eQPn0APCU9n8n03l4xZ9ufTe/F4PT7A/3b0/s/TnWe9edX+vd6NU5KL6aZmS3Ah95mZjWWJSglvVjSnaq+6rXiX1UzMzsZS37oLSmn+tzXi6jOt8x+y+D2Jd2RmdkpshwjyucAeyPi2xExQ/W5rWH4frKZ2aNyQjcSWKTtVJ9nm7UPWPDOMgAttaPDmmUoxcxscaaZYCa6mq9vOYJyUVTdSmoXQIdRnru4zzmbmS2Lm+OGBfuW49B7P9XX0mbtYJ5vEUTEVRGxMyJ2NmkvQxlmZktjOYLyFqpvGpyT7u33Sqov8puZrUpLfugdEf30dbDPAznw4YiY9/uzZmarwbKco4yIzwKfXY7HNjM71fzNHDOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGg5KM7MaDkozsxoOSjOzGrVBKenDkg5I+sZA22ZJ10v6Vvq5KbVL0gck7ZW0R9L5y1m8mdmpsJgR5e8DL57TdiVwQ0ScC9yQlgFeApybpl3AB5emTDOzlVMblBHxN8ChOc2XANek+WuASwfa/yAqXwQ2Stq2VMWama2ER3uOcmtE3Jfm7we2pvntwL0D6+1Lbf+EpF2Sdkva3aP7KMswM1t+J30xJyICiEex3VURsTMidjZpn2wZZmbL5tEG5fdnD6nTzwOpfT9w5sB6O1Kbmdmq9WiD8jrgsjR/GXDtQPtr09XvC4CjA4foZmarUqNuBUkfA14AbJG0D/hl4N3AJyVdDtwDvCKt/lngYmAvMAm8bhlqNjM7pWqDMiJetUDXRfOsG8DrT7YoM7Nh4m/mmJnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1agNSklnSrpJ0u2SbpP0htS+WdL1kr6Vfm5K7ZL0AUl7Je2RdP5yPwkzs+W0mBFlH3hTRDwduAB4vaSnA1cCN0TEucANaRngJcC5adoFfHDJqzYzO4VqgzIi7ouIr6T5Y8AdwHbgEuCatNo1wKVp/hLgD6LyRWCjpG1LXrmZ2SlyQucoJZ0NPBu4GdgaEfelrvuBrWl+O3DvwGb7UpuZ2aq06KCUtBb4E+CNETE22BcRAcSJ7FjSLkm7Je3u0T2RTc3MTqlFBaWkJlVIfjQi/jQ1f3/2kDr9PJDa9wNnDmy+I7U9TERcFRE7I2Jnk/ajrd/MbNkt5qq3gKuBOyLivQNd1wGXpfnLgGsH2l+brn5fABwdOEQ3M1t1GotY53nAzwBfl/TV1PZW4N3AJyVdDtwDvCL1fRa4GNgLTAKvW9KKzcxOsdqgjIi/A7RA90XzrB/A60+yLjOzoeFv5piZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1XBQmpnVcFCamdVwUJqZ1agNSkkdSV+S9DVJt0n6ldR+jqSbJe2V9AlJrdTeTst7U//Zy/sUzMyW12JGlF3ghRHxA8B5wIslXQC8B3hfRDwZOAxcnta/HDic2t+X1jMzW7VqgzIq42mxmaYAXgh8KrVfA1ya5i9Jy6T+iyRpySo2MzvFFnWOUlIu6avAAeB64G7gSET00yr7gO1pfjtwL0DqPwqcNs9j7pK0W9LuHt2TexZmZstoUUEZEUVEnAfsAJ4DPPVkdxwRV0XEzojY2aR9sg9nZrZsTuiqd0QcAW4CLgQ2Smqkrh3A/jS/HzgTIPVvAA4uSbVmZitgMVe9T5e0Mc2PAC8C7qAKzJen1S4Drk3z16VlUv+NERFLWbTZipKqKcuPT2o0jk+D7cfX9Wn6Va1RvwrbgGsk5VTB+smI+Iyk24GPS/p14Fbg6rT+1cAfStoLHAJeuQx1my0/CZShTFUIdtpo7VpizQjlug5Fp0E0M4pWRtnMKJtCEeRTJVk/yPol+dgM2cQ06vVhuktMd4l+H4oCyrLaT5bGK2VJREAZEOXxMmLO8sL1VrWijCgKKItleFEem2qDMiL2AM+ep/3bVOcr57ZPAz+9JNWZrTDlOWRCIyNw+mZ6p69j+vQWM2sz+iOiaIqyDUUbytmhBBACBTQmRmiOr6c1HrQPF3QenCYbn0aT0zDTq0IyyyBLI04JihJ6vYdCszdD9PowcGBWBWFang1RZShPjxflorLVFmcxI0qzx6bZ4MlztH4t3cdvYHx7i6nTM/qjULaqQIxGUDYAQdkIoh1EHpClqRD5eE77wSaj9zcYfXCE1pEe+XQVfmUzp+zkFK2MyEVkAoH6QWO6IJ/qk0310HSvCsuyJJvqEt1uFapFQUQgCfK8qr0sq5GrLQkHpdkCsk4brVuLRkeY2b6Jw09pM3mGmNlYUq4poFVWSQnV8DENI7NmSZ4FUpA3qmFdb13OVKdN2crprW3QPprTmI4UlGJ6Y8bMBihGgqIDRack62XkUw1aRzs0x4PGJDQnS5oTJc2JPvn4DNnYFOoXSCIaOeQZ9At0SNDtPmwUao+eg9JsPlmOztrOsadtZmJrxuQZort9htFNU5zWmSFTUJQZ/SJjeqZJr5dT9rMUjgV5XgVlo1EgoCxFrJ9hutmgtz5nejyjMSVUQNmA6ceVcHqXVrvP2laPdrNPr8iY6rYYH+uQjTVojIvGZE5jKqd1pEHnSIvOoRaaKUHQW98CoHWkS2OqS/WFOVsKDkqzeajZYOJJm7jveaL5hHHO2nyYHWuOsKU9Tifr0StzumWDbtnkwZk1HJxeQ7ffoEgjTAGZgjJEEaKRF/TaOb1Og5mRBt11DWZ61QnNaJaMbJ7icevHaed9mnnBaGOGfpkx2W9xZHSEiU0tut0G0/2cciaHmYzWgzmj3xtBZRB5td/WsSDrlTQaOSiD8AWdpeCgNJuHJLobMrRtmnO2HOQp6w/w1JH7OKN5lFFV3yTrkVNExpFilAf76ykQ3bLJsaJDv6xCcLxoM1U06Zc5k/0WE/0W4zMtZvo5maBfZPTLjLWdLu28Oqc4G5D9MiNCbB6ZZPPIJEAK3ozxmRYPrF/HsZEO6omsH7QPi+xwkE/2oTuzuCvltigOSrMFRCayvKSVFazNu2xujHNGfpQNWZeOSnJBEdBDTJbVn9IMGcfKDr1oMBM5R4pRjpUj9CJnsmhzuD/KRNGmWzToRcZ00WS6aAJVCHaLBmWZ0StzyhCtrGB9a5qRvAdAQyVliKOtDhHigUIUx5q0DuW0Dwej9/doHBwnpqZ8fnIJOSjNFqAyKIqMmTKnFzktFTRV0FHJmkw0qa5OFwQbs2o0WEQwnXUpgSIF6HRUf2YT0WKibFOSUUTGRNliOlpViJYtJos240WbXuTHa2iqCummqkPoXCVFZBwtRujkfVp5wb7YRHagwcihPu37xuDwGOXU9Cl/vf45c1CaLUABUYpemTNVNDnYX8vGbJKOCppRkAlyRFMZObNXv2GUaiRXRtDL+hTRowBKpigGBnk9qvOXJapGl9FgrOwcP6SfDdTZcCwQHfXIVXKkWMOGfAqA+w+vQwU0xgs0NkE5OVl97tKWjIPSbAGRQZ6XNLOCfuTc19tIU1UA9WKK0axPRyUdxKhyMjJyiSaQkVHqoXOExcBhcD7n64yzfSU9etGlmA1aYCaCEugFzERGrqpvunGEddkUR4sR4AnVR4fGe8TEBDHT8/nJJeagNJtPltFvi3Vrp9jSGWdN3mW6bHKoWEtLBQe1ljVZl835OBuzGcgK2gqI/HgQZulrOhmiMeer3rkGbrMwp68YCLmSoBcFJSW9KI+vPB19DmXTlCF6Rzqc9r2S/NBE9RXJovD5ySXmoDRbQDRgfafL1vYxtjSre1d31KMgY6wY4WCxlrGyw/psmo3ZJBuyLhuzGTrKaCqjmUaZJUE2kIYPC8l5zPYXUZIhmsop0fHgLSnpERws1rLn8HZG9jVY/4/j1bnJmZ6/470MHJRmCwnoFdXV517kTJfV1emCjMP9NUyXTe6rbqxFrpJNjQlOy8erUWY+ycZshnWKhwUnVOcuZ2Vzh5NzlDx8ZNijoBslBwvxt8eewrdu287j7yhofOf7FGNjDsll4qA0W4AKmOo1GOuP0MhKpoom3bLBZNniaH+EiX6bTFX7dNEkU9DK+ozkPTY2p9jWOsLjm4c5o3GUjdk069RlTSba6eLP7DnNWUXE8eWHzluWTEfBdAS9gENli7t7Z3DzsSdx7VfPY/tfw/o9D1AeOUrMzKzI6/RY4KA0m09ZohJmipxuuuo90W/TzRp0ywbH+h2miiZliMl+i7Fuh+l+g+leg5mZBnle0mn12DI6ydbRMbZ1xjin/QBPbB3gcfk467IeTUFHIuOhCzc5cCzEA8UIE2WbQ8Vabp08i7vHt3Bgch2HJkaZeHCU9Xc0Oee2GUZuu5fy4CFKf697WTkozRagkvQ5ygZTRYtj/TaZgoZKjsyMMNFr0chKJntNjk236fUa9GYaFOMN1MuYKuFo9zS+HVC0g1jbZ2TDNFs3HGPr6DHWNbqsaVSfkTw0s4YHums5Mj3CgaNr6R4cofVgTudBMfpASftIQXuiz47JPtnEUTQ2QYyPU0xMEf2eQ3KZOSjNFpDPBFMTbb43voHR5gyTvRaZqptdHJ3q0O1Vfz7Tky3iaAv1RT4tOodFY7LavnOopHWsTDcWEjBC0Rnlu+1tx692qwga00FjsqAz2eecyRk08QCa6hKTU9Xt0gbuT1mS7kcZpQPyFHFQms2jnOmx6WuHaHQ3Mb3uDKZyyPrprmqCdjcY6QdI5N2S5ngfArJeWd3+bKYPvT6anCa6M9VFlqIgioHPN87e4TzioRvxRkk5e0dzh+DQcFCazacsKG6/i9HbYfQENw3A157/eTmh/4XRzOyxyEFpZlbDQWlmVsNBaWZWw0FpZlbDQWlmVsNBaWZWw0FpZlbDQWlmVsNBaWZWw0FpZlbDQWlmVsNBaWZWw0FpZlbDQWlmVmPRQSkpl3SrpM+k5XMk3Sxpr6RPSGql9nZa3pv6z16e0s3MTo0TGVG+AbhjYPk9wPsi4snAYeDy1H45cDi1vy+tZ2a2ai0qKCXtAF4K/F5aFvBC4FNplWuAS9P8JWmZ1H9RWt/MbFVa7Ijy/cCbqf5XTYDTgCMR0U/L+4DtaX47cC9A6j+a1n8YSbsk7Za0u0f3UZZvZrb8aoNS0suAAxHx5aXccURcFRE7I2Jnk/ZSPrSZ2ZJazH8u9jzgJyVdDHSA9cBvARslNdKocQewP62/HzgT2CepAWwADi555WZmp0jtiDIi3hIROyLibOCVwI0R8RrgJuDlabXLgGvT/HVpmdR/Y4T/300zW71O5nOUvwhcIWkv1TnIq1P71cBpqf0K4MqTK9HMbGVpGAZ767U5nquLVroMM3sMuzluYCwOzfsJHX8zx8yshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0syshoPSzKyGg9LMrIaD0sysxqKCUtJ3JH1d0lcl7U5tmyVdL+lb6eem1C5JH5C0V9IeSecv5xMwM1tuJzKi/NGIOC8idqblK4EbIuJc4Ia0DPAS4Nw07QI+uFTFmpmthJM59L4EuCbNXwNcOtD+B1H5IrBR0raT2I+Z2YpabFAG8JeSvixpV2rbGhH3pfn7ga1pfjtw78C2+1Lbw0jaJWm3pN09uo+idDOzU6OxyPWeHxH7JT0OuF7SNwc7IyIkxYnsOCKuAq4CWK/NJ7StmdmptKgRZUTsTz8PAJ8GngN8f/aQOv08kFbfD5w5sPmO1GZmtirVBqWkNZLWzc4DPw58A7gOuCytdhlwbZq/Dnhtuvp9AXB04BDdzGzVWcyh91bg05Jm1/+jiPgLSbcAn5R0OXAP8Iq0/meBi4G9wCTwuiWv2szsFFLEyp8elHQMuHOl61ikLcCDK13EIqyWOmH11Lpa6oTVU+sw1XlWRJw+X8diL+YstzsHPp851CTtXg21rpY6YfXUulrqhNVT62qp019hNDOr4aA0M6sxLEF51UoXcAJWS62rpU5YPbWuljph9dS6Kuocios5ZmbDbFhGlGZmQ8tBaWZWY8WDUtKLJd2Z7l95Zf0Wy1rLhyUdkPSNgbahvO+mpDMl3STpdkm3SXrDMNYrqSPpS5K+lur8ldR+jqSbUz2fkNRK7e20vDf1n30q6hyoN5d0q6TPDHmdq+IesZI2SvqUpG9KukPShcNYZ62IWLEJyIG7gScCLeBrwNNXsJ4fBs4HvjHQ9hvAlWn+SuA9af5i4HOAgAuAm09xrduA89P8OuAu4OnDVm/a39o03wRuTvv/JPDK1P4h4D+l+f8MfCjNvxL4xCl+Xa8A/gj4TFoe1jq/A2yZ0zZU733a9zXAz6X5FrBxGOusfR4runO4EPj8wPJbgLescE1nzwnKO4FtaX4b1YfjAX4XeNV8661Q3dcCLxrmeoFR4CvAc6m+jdGY+3sAfB64MM030no6RfXtoLoJ9QuBz6Q/2KGrM+1zvqAcqvce2AD849zXZdjqXMy00ofei7p35Qo7qftungrpsO/ZVKO1oas3Hc5+leoOU9dTHUUciYj+PLUcrzP1HwVOOxV1Au8H3gyUafm0Ia0TluEescvgHOAB4P+k0xm/l26sM2x11lrpoFxVovpnbqg+TyVpLfAnwBsjYmywb1jqjYgiIs6jGrE9B3jqCpf0T0h6GXAgIr680rUs0vMj4nyq/3rl9ZJ+eLBzSN77BtWprA9GxLOBCR76L2OAoamz1koH5Wq4d+XQ3ndTUpMqJD8aEX+amoe23og4AtxEdQi7UdLsvQYGazleZ+rfABw8BeU9D/hJSd8BPk51+P1bQ1gnsGruEbsP2BcRN6flT1EF57DVWWulg/IW4Nx0ZbFFdVL8uhWuaa6hvO+mJAFXA3dExHuHtV5Jp0vamOZHqM6j3kEVmC9foM7Z+l8O3JhGHcsqIt4SETsi4myq38MbI+I1w1YnrJ57xEbE/cC9kv5FaroIuH3Y6lyUlT5JSnWl6y6q81ZvW+FaPgbcB/So/jW8nOq80w3At4C/AjandQX8dqr768DOU1zr86kOWfYAX03TxcNWL/As4NZU5zeAd6T2JwJforpv6R8D7dTeSct7U/8TV+D34AU8dNV76OpMNX0tTbfN/t0M23uf9n0esDu9/38GbBrGOusmf4XRzKzGSh96m5kNPQelmVkNB6WZWQ0HpZlZDQelmVkNB6WZWQ0HpZlZjf8PH3zQaacpZpQAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUoAAAEICAYAAAA5ub1iAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3df5RkZ33f+ff3ufdWVVf/nN8azQgJIdkOtmN+KPww2HFQ7AXWazhr1gFzQIclRzkbb9YOzsYC2zHx2Ti2s8fYJDZEJ+CImIAxDouCwQ4ReO0ki4wwNgaEQNiARujXjGZ6+ldV3br3u388T1VX93TPbUkz0z2Zz+ucOl313Hurnqqe+fTz3Oe5T5m7IyIi2wu7XQERkb1OQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEF5mTKzt5rZb+12PUSuBArKPczMftTM7jGzZTN7yMw+ZmYv3gP1Omxm7zOzb5rZopn9VzN7/nn2XzCzd5vZw2a2ZGZfNrPbLmWdt6nXs8zsM2a2mn4+awfH3Ghmvck/Uhb9tJl9w8zOmtn7zWxuYvv/bWZfSe/9S2b2+olt35N+v5M3N7MfTttvSXU7a2YnzOyXzSzfVKdXm9m9ZrZiZl81s+9J5a/d9Lyr6bmfeyE+vyuJgnKPMrM3Ab8K/AJwBHga8BvAK3azXskM8GngucB+4A7g98xsZpv935aO+WvAPPBDwP0XskKbw2MH+7eADwO/BewjvocPp/Lz+XXie5/0euB1wIuAq4Ep4F9ObF8B/ifie78F+DUz+24Ad/9jd58Z3YAfBJaB30/HdoGfAA4CzwduBv7RxPv4fuCXgDcAs8D3An+Znvu9m57776dtf9r4AclG7q7bHrsR/0MtA//LefZ5K/BbE49/B3gYWAT+CPj2iW0vB74ILAEPAv8olR8EPgKcAR4H/hgIT7LOZ4HnbrPt88Arz3PstwMfT3V4BHhLKm8T/1h8M91+FWinbd8HnAB+Kr3vf0f8w38b8FXgFPABYP82r/kD6bOwibJvAC89Tz1fnZ5z82f/QeD/nHj83UAP6G7zPHcCP7nNtt8EfvM8dXgT8B8nHv834I07/B19Evi53f73fTne1KLcm14IdIAPPYFjPgbcCBwmthjeO7HtXcDfc/dZ4DuAT6TynySGzSFiq/UtgAOY2W+Y2W/s5IVTl7XF9q3ETwH/zMzeYGY3bjp2FvjPxBbU1cANwF1p808DLwCeBXwX8DzgZyYOv4rYor0WuBX4B8Argb+Znus0sQU4eq3PmdmPpoffDnzOU4Ikn0vlW73HOeDniUG15S6b7reJv4/NzzMF/A3gC1tsmwZeRWzdbud7R8eaWQbcBBwys/tT1/xfpdfY/NzXpmPfc57nlu3sdlLrdu4NeC3wcMM+b2WiVbNp2wIx8ObT428Afw+Y27TfzxO7nzc8hbrOAX8BvPk8+0wRQ/gzQEkM1Jelba8BPrvNcV8FXj7x+H8Avpbufx8wADoT2+8Fbp54fDS9Xr7Fc/8s8P5NZe8F3rpNXX4N+KmtPnvg7wJfBq4j9gbuTJ//C7d4njuIfxRsi22vA/5qq21p+/9K/MN2MD2+Or3OPem9HgT+K/DPtnm/f7jb/7Yv15talHvTKeDgTs+7mVlmZr+YTuSfBb6WNh1MP3+Y2P3+upn9v2b2wlT+L4ih9Z/M7C+f6ABLarn8R+BT7v7Pt9vP3dfc/Rfc/bnAAWL39XfMbD9wDTEQt3I18PWJx19PZSOPuXtv4vG1wIfM7IyZnSEGZ0VsLW+2TAz5SXPE0xMbpBbz3yaea93Ku4H3AX9IbO19MpWf2PQ8/4LYov8RT+m1yS3Ae7baZmavBP458Q/MyVS8ln7+S3d/KJX/CvF3vdnrOX9LVc5nt5Nat3NvxFbJCvCq8+zzVlKrhtgSuRd4OrHbN2pR3rDpmAL4h8ADWzzfdwCPMtEia6hjG/gDYivsCZ3XJA7sOHEw6DXAn26z3+YW5Q+wsUV5YtP+9wEv2mEdfoAYZJPnKL/OFucoiYMpK8RzoQ8TQ3btPPUePXeYKPunxHO1B7Y55hpgCDxji20vBR4DnrfFtgeA1088/p/Z1EInDjKtALO7/W/7cr2pRbkHufsi8E+AXzezV5pZ18wKM3uZmf3yFofMAn1iS7RLHCkH4uhumiYy7+4lcdClTtt+0MxuMDMjDgJVo23nY2YFcQBjDbjF3c97jJn9rJn9jVSXDvDjxAGk+4iDSUfN7CfMrG1msxNTjd4H/IyZHTKzg+kzOd/c0XcSz4Vem173kJltN0vgD9P7/T/S6/7vqfwTW+x7O/AM4rnSZ6XX+T3iqQDMbL+ZPSNNE3omsVX386PPxczeDPwo8Lfd/dQ29Xkd8N/cfUPr2sxeQvxj9MPu/idbHPebwD+wOGVrH/EP4Uc27XML8Lvufk5rWXZot5Nat+1vxHOV97Demvk94LvTtrey3qKcIZ5rXCK2il5PalESB1l+nziwcZY4teXF6bh/SOymrxBbQD878drvBN65Tb3+Znr+VWLranT7nm32/xlia+oscWT7D0fvI23/DuIAzun0Pm9L5R3g7cBD6fZ20jlJtm5RBuJgy33ps/gq8AsT278AvHbi8bOJ503XiANgz57Y9hbgY9u8n/Fnnx5/S3rN1fT5v2nT/k78Qzb5Wb1l0z5fYovRa2I3frjp2I9NbC+I08bOpM9u/BlNfIZn2GFPQbetb5Y+TBER2Ya63iIiDS5KUJrZS83svjS3a9cvVRMReSoueNc7TYL9MvD9xPNenwZe4+5fvKAvJCJyiVyMFuXzgPvd/S/dfQC8n71xfbKIyJPyhBYS2KFjxLldIyeIF/Nvq2Vt7zB9EaoiIrIzPVYYeN+22nYxgnJHzOxW4vW5dOjyfLt5t6oiIsLdfte22y5G1/tB4lUGI8dT2Qbufru73+TuNxW0L0I1REQujIsRlJ8GbjSzp6e1/V5NXCRAROSydMG73u4+TJeD/QGQAe9293OWlBIRuVxclHOU7v5R4KMX47lFRC41XZkjItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItJAQSki0kBBKSLSQEEpItKgMSjN7N1m9qiZfX6ibL+ZfdzMvpJ+7kvlZmZvN7P7zexzZvaci1l5EZFLYSctyn8LvHRT2W3AXe5+I3BXegzwMuDGdLsVeMeFqaaIyO5pDEp3/yPg8U3FrwDuSPfvAF45Uf4ejz4FLJjZ0QtVWRGR3fBkz1EecfeH0v2HgSPp/jHggYn9TqSyc5jZrWZ2j5ndU9J/ktUQEbn4nvJgjrs74E/iuNvd/SZ3v6mg/VSrISJy0TzZoHxk1KVOPx9N5Q8C10zsdzyViYhctp5sUN4J3JLu3wJ8eKL89Wn0+wXA4kQXXUTkspQ37WBm7wO+DzhoZieAnwN+EfiAmb0R+DrwI2n3jwIvB+4HVoE3XIQ6i4hcUo1B6e6v2WbTzVvs68CPPdVKiYjsJboyR0SkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpIGCUkSkgYJSRKSBglJEpEFjUJrZNWb2STP7opl9wcx+PJXvN7OPm9lX0s99qdzM7O1mdr+Zfc7MnnOx34SIyMW0kxblEPhJd38m8ALgx8zsmcBtwF3ufiNwV3oM8DLgxnS7FXjHBa+1iMgl1BiU7v6Qu/9pur8E3AscA14B3JF2uwN4Zbr/CuA9Hn0KWDCzoxe85iIil8gTOkdpZtcBzwbuBo64+0Np08PAkXT/GPDAxGEnUpmIyGVpx0FpZjPA7wI/4e5nJ7e5uwP+RF7YzG41s3vM7J6S/hM5VETkktpRUJpZQQzJ97r7f0jFj4y61Onno6n8QeCaicOPp7IN3P12d7/J3W8qaD/Z+ouIXHQ7GfU24F3Ave7+KxOb7gRuSfdvAT48Uf76NPr9AmBxoosuInLZyXewz4uA1wF/YWZ/lsreAvwi8AEzeyPwdeBH0raPAi8H7gdWgTdc0BqLiFxijUHp7v8FsG0237zF/g782FOsl4jInrGTFqWIyFMTMqzIsVYLaxVYUUBRQJ7hWQDb1Bbb/Hg77hv3d18vm3y6eoux5tG+7lDV2EPZti+joBSRi8MMyzLIMrJ9C9SH99E/PM3K1S1WrzL6+51yvsKmh4S8JphjwTFzzEg/Y8C5G+6Wnna9rK5jWQixrBoGqirEOTgW87OuDK8M3Nbn5tQGNdgwYH0jXzXKt3W2fSsKShG58EKGZakV2W5TH97H0rfMs3hdxso1NVPHz/LMgye5cfZRbpx6hAPZMrNhjcxqSs+pMAoqOqEEYKnusFK3yfDxfkv1FGerDpnFspZVPDqc5ZHhPLUHZrIe06HPSt1mcdil9IzCKoowpF8XLFUdloYdHh90eXhljm9sn5MKShG5wFI3O7Tb2NwsPttl9do5zjwjY+X6ktkjy9yw/yTPmDnJ4dZZOjagEwZMhzifeqVuc6bq0gklc/QAODWc4ZHhPB0bciiPgXeqmuFkOUsRYlnHSk4O53iwv4/ajcOtJeossFq3WKymqN2Yz9fo2JByIvqmspKF9tr2IzEoKEXkQjIjTHUIc7PUBxZYvXaWs9fkrFwD5TU9rjt6iutnT/GM7mMcKRYpbEjLKlbqNqXn9OqCM1WXpTo27zo2BGCp6rBctQnmLFUdijBktWrTr3Oo2qxWbTKrOV12WRm2GXqgX+c8bPPUGJUbmTm1G6tVi36ds1a3WB62GNQ5a8PivG9LQSkiT40ZWMCCYXlOOHSA8up9rByf4uzTAqvHa/Kjq3zLoVN858I3OdY+zVX5IoVVrNQtzlRdgjkZNat1m4cG85wqpwFohyG1B1aqFmtVQcCZykqCeQzDKieYU4SKgLM0bLPYn6LGmMpLOlnJsA4MPSO3ivlWj9m8F4OyKlgdthjWGUM//5RyBaWIPHmjc5GtApvqYNNd+tcd5MyNbZaeBoODQ1r7e1y9f5EbZk9yXeckC9kqhVX0vODxaobFYZfMarphwFLV4aH+PN9cmcfM6eYDag8sDjqsDFpkoaZblBShojcs6FcZmTnTxYAs1Cz2OyyuThFCzcJUj7m20a9ylgctsjQINJWVDOuM1WGL1WELgKoO570IW0EpIk9MmopjWYa121i3CwuzlFfPs3qkzdnrAstPH7Lv2CLfOr/I9TMnOdpaZF++wkK2Qu2BnhepO93h7LBD6dm45Xi636XGGFYZvWFB7UZZB6raqOqMwXB9Gk8wGDj0hzm1x5HwLMQR9JVBQW8YW5wAbs7pfpeVMl4yXWOUVcZqWbC81sbq7d+yglLkStc0Z9HCxF2DLBvPh+TAPsrDs6wdabN8dcbqUWd4rMcNxx/jOfsf4GntU1yVL9IJAyoP1AQGMG69lZ7Rrwv6dU6/zuhVBb0qpz/MqdJ0oNqNwTCjrDLcbT34gLoO6S1sbA7WbpgbZbXxreShJgs+DtVhHSjLnEE/V4tSRJLUVSZsDEebCEvfNGHbLIVjmhNp3Sl8pku1r8vK8SmWjmesXuUMDw6YPbjCDftP8l3zD/LXph5kOvQpqKgJlAQGHluDwWKrr7CKYDVFqChCTLVla1O5YUA7H1LVgX6ZU5Y5IdS0WkOy4Kz2W/R7BWZO0RqSZzWDMqcsMwzIi4o8rxgOM8oyvm6rNaTIK6o6MBhk1FUG5o1rnykoRa4QlueEhXlsZhov1v/rW1VvvEoFoKqhqiCEePVMkeOdNnW3oH+ww/KxnJWjRv9IRfvwEtfvP811s6f41u4jXF2c5lB+loWwNn6NnudQw8Cz8eh26RntUHK4tUSvLlirCvKiYiorOTKVU6ZBlkGVMdPq069yhnXAU1e8XZSYOVUVqOtAvwrUVQA3aodBP7YUsyyG66jVWQ4zQnA6nRL3IYNBTjkoMLUoRSTMTFNfexUrT5thMBOoC6gnEmAcFDWEIYRhLKgLo2pB1TGGXRjMO+WRkoOHz/KdC6e4YfoxntF5lGPFaQ5lS3TTlJ7R4MmoC71CPAlYejY+JwkQcDqhJLOaXl1QWEXfaurMYlhmgaWyTe1GwKnThMceUFXxSpy4Iq5RDQP1IIstxACW1dQhXsGTZT7uqrt7fBbzeFVPOH+TUkEpcoWw6WmWrp/h5HcFBoeH5DMlU90+7byilQ8xYr5UdWBtUDAoYzwUxZBOMWQ6HzKVl8y31zjcXuaq9iJHizMcK06zEFbp2JC2VWTmVG7jgASosNj99pxATSeUVHWgrHKWq3bseltFZYGlusOZwRTB0lQgnF5VsNjvYMB0MaCdDVmsOwz6BV5D3qrI8orhIIN+Oqc6VZEVKZwHOUODkFUURYW70esXeL1+WaRrwrmI0G7R2x8or+lz/bGT3DD3GNdPnWRfvsJcWKOwisxqKg+s1C2W6ikyarqhz3QYkKVh4cKGzIYec9ansJqAk5mT4QSgxOKIsgcqjApjtW6nSeUZlQcqArUHijCkm1qIgXjOcioMqIo4QbwVhtRuHOlANx8QxmUBM6ddDBkMM7LUIsyymkEnp67idszJMifL6nE3vaoCZpDnFVUV6K+0sLMFodrug1NQilwxPBh1YRSdIcemz/DXZ07wnZ0HOBDWmA8VLTMCkGH0vKZMx3XMKDBKnIE7GVCYUVig9JqeO5Mza2p3ehi9dM12lQZxKoxATUUgo6ZIXfQiq6hTqGZekxU1c3lv3MUGmMt7HCWOZkPsvk/nffa3O/SqnEEdz1+WdRYHf6o4jWhYZWShppXHFOyXeRo9Twtm1AZlIOubRr1FJA7aZD1nsNTiZG+G/mxBx0q6VjFtgcICIX3pQWE1VUqODCNuqShSWTAjSy3HjI1BWWHjQBvJrKZjJQRoeZX2C1Qeu+QDz8fHjKb/1G6Unrr/Noxdc8L4HOfBoqDXjlOLVut4KWK/yunXOUMPDOsQpxbVMRyHHphuDajd6A1zVvstzDKsVVN1a3W9RQR8eZX5vxownGpzb3mcYM7+fJlvaz3EdOjRTiGZmYGHc74npiCj3mZWdrWpNRbMKagIGB2GlMTuccdL6m2+gaZKoVvYkCwFcpValYVVtKjGZaMVhioPlMTu/Kis9HzD+dHSc3peUHsgWB2v9KmmOFnO8HBvjm8uz/PImVl8++UoFZQiVwrv9Wg/tMRCOzCYL/jmsTnOHOrS84LK4+o5WZpPmZlRpalC2eSE9HOuia7JzCgmWpW1OwU1GHTwOFJtFYVXlD55VU08Ipvo88agrCmsJsPH4RmAbFOLr3KoWQ/YUaiOQnJy1L1O50pH+52pOzxezfBA5wBfaR/mvvwI31BQighVha31KZaGZP0Cd6NjZWzBmZGZjbveNRtbjoGwoWwUngWByuP5zWq82nh8hoz1AZ6aOFhTTrRIRwEZJoPSnAKnsI37TIb16PkqPLUxfbxvBdTm4zZrBmCj/dYVrNKiIrTqcbf+j7LtT1IqKEWuEF4OYa1HvrhGa7HDmaUp/qp/iKuKRa7OVulalUaeY9Oqpo5nJt3ZfCF05ZOtQKPGx2FWeQo69w2d7HrTucua9csR1ysJxflmficB4miMxxDMiGGabd4nKVI9IQZsFmo6tsZsGNCxkopw3rmUCkqRK4XXeH+ArfbJe1D3Ms6UU6zU7XFbMTOjph4H4SgsJ23uigdzgvs4iDAY+HqrMAClx5bjZDBu1aJsWT3uYo9Cb3NrclI1sd9WMoyw6Vr2MD4X6mRecThbZilfHH+dxFYUlCJXCK8dvMaGVfyyLd8YQKNADASwmtH3cU0G1ej+ODytpvT1QMow8Hh+cuM5To9dd18/dwiTYRnPQcbpSeuvNZquVLGxBQtQjwd8tg7LyZAch/hIql8gjvCPVlffjoJS5EqS0s8qYGgsllMsVR1WPaPvQ9rkYLFFOQ4nr85pVY664qVXlB7PXgZ3SuL5w8HE3MrSndKh5xmlr6/7GEfG6xSccUWf2KWG1kQ4T05Tisel86O+cW2PDV3rieA+d/x+FMwGVPStorDheRdRUlCKXGncsRqsMgZVNp6rCLGrnZFRU1NvGpzZ+BxhYxd8k5r1KUObR6uBtKL5xit6RvtmrHext+o6T3bx10PXxi3PwHorcquQ3DCiP/kk56GgFLmCuDtWO6kRB6Qlz/ANE85juVGnc4+bR723MxrYCaQuscWBFAyqdHzAaVm9IRwDbLgyaPT6m197VL8417MafyFYNhGO1bg7HzYMMI2PSyqcMk0XKj3f6uvAxxSUIlcQy+KyaRBblCtlm+Wqw6rnrHpJQU1BvDSxHJ0D9Ao2TbApLHan+16zUnua/hPPP5YOA4/rT47CuPZ4SeMgnU2MlzBWtKjHC2n0x+cMPYamj1YKWg/o2BKNy6hV+IZW7ygkR2W11VvM+5x4D2SUOwh/UFCKXDEsxC8B8yyM86NOV7fEq1w8nvPzePni+rzIc/vOpddkGAN3eikUC2pa1Aw8sOo5pWfjSeWjr38YLdw7WmAjM0+BWqfWoaduu5/T5d7K6HzkZuNBnB08R3ye8++noBS5kqTRDzfw4GRWU4QhLWqK0aj1VodxzlnKsfH8yIlDK4wBGZ0UnpU5PY9fCdsJJR0raVGPB3RaVlOkbvrkaHcx8TUU40Gacae9Hi+cMdn1HtVjq/OTG+sdO/RluvzxfBSUIlcKC7F1mGVgsevdGxacLGd5sJqnsNN0rRrPf5wcta58q0sInVXPWEnXUmcer8YZkMUVzMlYJX6R1+g84MCz8RNPhz6zoce0DSipyTxe7jgd6vF5TXxTPE8MLMUrc9bnYk52vc/XGh1Nf4rXs1fpksnheT86BaXIlcQCZPHcYSiNpX6LxwfTPDacY9oGzIYenXSFzjnGo9jr11D3PKPnBb26IEvXZw9SWek5mdVpKCbQqye73k6v9nhVTFqEgzSoVDnjsD6f+nyjLxO2nh4UXy9zjXqLyFaqmmzg5CuBM4vT3N85yFy+xplON7bwQn+8iG9hQzoWu8ojGU7HhmkpNhu3HjPq8X49L6gwpq2kG/rjBXsHnjEdBsyGNTqhpEWVnq+ik1qzk7E2Oe1nMhgD8YqgcXiPu+T1hoU91ifRb0zeOp2DrXB1vUVknVcV9PvYyhozX1sm60+z+kCHUwvH+OjcMYZdp5qq8amKYmbA/EyPhak1pvKSmbzPQmuNhXyVfcUK+7MVFrLVcZc1s5rSc86mUaKWVbSsIhC/BwegEwbj/UtiFzwLNSGthl56oLB6HJi4n7s4RxoJDxaDs5rock92yUf7hvFpBD8nLDMzsvQ9PKFh9FtBKXKl8BofDGB5mWw4ZObkWaY7bXyqRTVVUE3lVFOBwUxOf75gbd8MZ2ec4bRTd2vCdEmnO2Cu2+NId5nj3TMcaC2zL1+JXfZQjteSHHW5J0ed4zqRIY54p0sZY8szrXueysMo9Gzz4hukEfjYkqzGS7udG4KbNW3fbo3MEQWlyJXCHR8O8TXHBmUMnRCwLJBnGXmeYyEw3SrwdgvvtikXOqwdKujtyxh2c4ZTUzw+M8fD+w7whYWjzM2ucmh6hcNTSxxqLXOgWKFI32kTrGY29JjNelRuLFbT9DxnNvTYny9TE1ip21QYc6HHQliN144TVyavqOmYUxBncVbpPTCxStB28ZZhE9elj7ri63vX41XWnZKMgWfnPVPZGJRm1gH+CGin/T/o7j9nZk8H3g8cAD4DvM7dB2bWBt4DPBc4Bfwdd/9a0+uIyMXnVZwx6FWaQG4TwzbBsBSeZBmWBdqdDq0TXea6bbyVU7cyqqmccjZjMN2hP9/lgYWD/OWcM5yu8emK0KrIsposr2kVQ6Za8fu3h1VG7VBkNa2sYiovY7e+6LO/tcKBYoX9+QrzqVs/HfpM22B8PnS0mG+I301LYevLq8XrzdevyhndL8jG5ypH4Ri3jxYNtjQ96fwzKXfSouwDL3H3ZTMrgP9iZh8D3gS8zd3fb2bvBN4IvCP9PO3uN5jZq4FfAv7OTn6JInLxeT3Zdqo2TMFx29hGs+UVOHUaywIWAnmek2cZ7VaBtVrU8zMM901RzuWU3UDZLag6BVU7fRd4DksBPEu3kL4W1qBuwXCmwroVxVRJtzNgobvGke4S10yd5mARu/UHsmVmwxpzoRcHmdJybZ30HT4F8QvP6g1Jt313fBSS1Q5HzWEHQenuDiynh0W6OfAS4EdT+R3AW4lB+Yp0H+CDwL8yM0vPIyK7ydcHPbbevnFbbIRVeBUD1MIgTjFKrU9bPEvxSJtWq4iXRmYZ3i6oOy28COMJ7nUeqDoZdRHGE8KrTqA/lzGYzRl22/Sm4BtzNV/ff5Av7z/E4ZlljnbPcrS9yOHWWfZny8xlvfEofMdKuqGkQ0XHHNzHk9ULC+Pu9+aR7/gdkE5FdeG63gBmlhG71zcAvw58FTjj7qNZmieAY+n+MeABAHcfmtkisXt+ctNz3grcCtChu5NqiMhTNRpcsXDuZO7zSfuOghNi69OGcdX0zd32kOfra6CZkWUZrTyP2wGC4UXOTLdN3Smo2zlVO9DfX7B8dZulq1qcPjTLIwdneWh2jk42pJOXTGXxNpv3OFgscShfGk9pmg1xxfKFMKQ7cdVOGK++fu7CHhlGxyqmQ/8pd71x9wp4lpktAB8Cvm1HH+75n/N24HaAOduv1qbIpTLqXk92s59IaE4csyE4z/dapGvNJ8uDYXlOVuRkeU4rz+nMz9J+fJ7O4y1WjrZYOrPAmfmZcSvUMsfymiyvmOqUzHb6LHTWONJZ4kj7LE9rn+L61qMczpbphiGzNqRjYbzO5pP1hEa93f2MmTZtkmUAAApkSURBVH0SeCGwYGZ5alUeBx5Muz0IXAOcMLMcmCcO6ojIXjVqYe6kpTm5WEbTGbWJARTf8MXZVXytqsKHQ8wMzzJsOKRTVbTOTDP12BRrh3IGM62JesabB6hzOFPAY7POFw5UFPt6XLVviRvnH+MZ3ce4tn2Sa4pTXJWtsD+UzNr681RpjmaJ0/OMlbr9lEe9DwFlCskp4PuJAzSfBF5FHPm+BfhwOuTO9Pj/S9s/ofOTInuLBVsf1JkMxs0hmabjjO9v3vZEnHN8FcOzdhywMIzfFDkcki2v0j3TZeqbHep2HtfPrOtYF7M4IBQMN6Pq5gzmc/pz05w5MMMfH7iKTxwYMn1olesPnOLZCw/wvOmvcn3++HgdzNGIeS+tvN6riw1fjbHZTlqUR4E70nnKAHzA3T9iZl8E3m9m/xfwWeBdaf93Af/OzO4HHgde/QQ+ShG5RMZhuV0r8lK0b9zHrU73WB+raqzXh6VlrMjjnMnaGbW3bNOCF3mR02m18Okp6oVpBvs7rB7OWbl6ni8cneVLRw/z2aPX8B1z3+RY+zTHitPMhjWmbUCF8fBwgRODA7h/Y9tq2l5o7M3Zfn++3bzb1RD575/ZxnOTI15fmmBskuo3WjtzW6PTBLA+kJRlWKuFtVv4/nl6V8+yeqRg7WCgf8DpH6qYPrLC9QdOcai9zGzRA+DR3izfXJnn7je8n/4DD2zZrNSVOSJXkokW3J6U6uc1Wy4YPN5nJO0zLlnrYVmGrawy9fgina918W6butti2C0o56Z5rDvLw4WNFy8OQ8gGTntRX1crIpebnbRwtzrvWVd4OaBeWoplKUxzILfA1HbPVa1u+zIKShH579tkmD7J1vT5l8wQEREFpYhIEwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEiDHQelmWVm9lkz+0h6/HQzu9vM7jez3zazVipvp8f3p+3XXZyqi4hcGk+kRfnjwL0Tj38JeJu73wCcBt6Yyt8InE7lb0v7iYhctnYUlGZ2HPgfgX+THhvwEuCDaZc7gFem+69Ij0nbb077i4hclnbaovxV4B8DdXp8ADjj7sP0+ARwLN0/BjwAkLYvpv03MLNbzeweM7unpP8kqy8icvE1BqWZ/SDwqLt/5kK+sLvf7u43uftNBe0L+dQiIhdUvoN9XgT8kJm9HOgAc8CvAQtmlqdW43HgwbT/g8A1wAkzy4F54NQFr7mIyCXS2KJ09ze7+3F3vw54NfAJd38t8EngVWm3W4APp/t3psek7Z9wd7+gtRYRuYSeyjzKnwLeZGb3E89BviuVvws4kMrfBNz21KooIrK7bC809uZsvz/fbt7taojIFexuv4uz/viWM3R0ZY6ISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEgDBaWISIMdBaWZfc3M/sLM/szM7kll+83s42b2lfRzXyo3M3u7md1vZp8zs+dczDcgInKxPZEW5d9y92e5+03p8W3AXe5+I3BXegzwMuDGdLsVeMeFqqyIyG54Kl3vVwB3pPt3AK+cKH+PR58CFszs6FN4HRGRXbXToHTgP5nZZ8zs1lR2xN0fSvcfBo6k+8eAByaOPZHKNjCzW83sHjO7p6T/JKouInJp5Dvc78Xu/qCZHQY+bmZfmtzo7m5m/kRe2N1vB24HmLP9T+hYEZFLaUctSnd/MP18FPgQ8DzgkVGXOv18NO3+IHDNxOHHU5mIyGWpMSjNbNrMZkf3gR8APg/cCdySdrsF+HC6fyfw+jT6/QJgcaKLLiJy2dlJ1/sI8CEzG+3/7939983s08AHzOyNwNeBH0n7fxR4OXA/sAq84YLXWkTkEjL33T89aGZLwH27XY8dOgic3O1K7MDlUk+4fOp6udQTLp+67qV6Xuvuh7basNPBnIvtvon5mXuamd1zOdT1cqknXD51vVzqCZdPXS+XeuoSRhGRBgpKEZEGeyUob9/tCjwBl0tdL5d6wuVT18ulnnD51PWyqOeeGMwREdnL9kqLUkRkz1JQiog02PWgNLOXmtl9af3K25qPuKh1ebeZPWpmn58o25PrbprZNWb2STP7opl9wcx+fC/W18w6ZvYnZvbnqZ7/NJU/3czuTvX5bTNrpfJ2enx/2n7dpajnRH0zM/usmX1kj9fzslgj1swWzOyDZvYlM7vXzF64F+vZyN137QZkwFeB64EW8OfAM3exPt8LPAf4/ETZLwO3pfu3Ab+U7r8c+BhgwAuAuy9xXY8Cz0n3Z4EvA8/ca/VNrzeT7hfA3en1PwC8OpW/E/jf0v2/D7wz3X818NuX+HN9E/DvgY+kx3u1nl8DDm4q21O/+/TadwB/N91vAQt7sZ6N72NXXxxeCPzBxOM3A2/e5Tpdtyko7wOOpvtHiZPjAf418Jqt9tulen8Y+P69XF+gC/wp8Hzi1Rj55n8HwB8AL0z387SfXaL6HScuQv0S4CPpP+yeq2d6za2Cck/97oF54K82fy57rZ47ue1213tHa1fusqe07ualkLp9zya21vZcfVN39s+IK0x9nNiLOOPuwy3qMq5n2r4IHLgU9QR+FfjHQJ0eH9ij9YSLsEbsRfB04DHgN9PpjH+TFtbZa/VstNtBeVnx+GduT82nMrMZ4HeBn3D3s5Pb9kp93b1y92cRW2zPA75tl6t0DjP7QeBRd//Mbtdlh17s7s8hfvXKj5nZ905u3CO/+5x4Kusd7v5sYIX1r4wB9kw9G+12UF4Oa1fu2XU3zawghuR73f0/pOI9W193PwN8ktiFXTCz0VoDk3UZ1zNtnwdOXYLqvQj4ITP7GvB+Yvf71/ZgPYHLZo3YE8AJd787Pf4gMTj3Wj0b7XZQfhq4MY0stognxe/c5TpttifX3TQzA94F3Ovuv7JX62tmh8xsId2fIp5HvZcYmK/app6j+r8K+ERqdVxU7v5mdz/u7tcR/x1+wt1fu9fqCZfPGrHu/jDwgJl9ayq6GfjiXqvnjuz2SVLiSNeXieetfnqX6/I+4CGgJP41fCPxvNNdwFeA/wzsT/sa8Oup3n8B3HSJ6/piYpflc8CfpdvL91p9gb8OfDbV8/PAP0nl1wN/Qly39HeAdirvpMf3p+3X78K/g+9jfdR7z9Uz1enP0+0Lo/83e+13n177WcA96ff//wD79mI9m266hFFEpMFud71FRPY8BaWISAMFpYhIAwWliEgDBaWISAMFpYhIAwWliEiD/x/zznAkmRcIxwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.eval()\n",
    "k+=1\n",
    "im = dataset_test[k][0]\n",
    "show(im,sz=5)\n",
    "with torch.no_grad():\n",
    "    prediction = model([im.to(device)])\n",
    "    for i in range(len(prediction[0]['masks'])):\n",
    "        plt.imshow(Image.fromarray(prediction[0]['masks'][i, 0].mul(255).byte().cpu().numpy()))\n",
    "        plt.title('Class: '+str(prediction[0]['labels'][i].cpu().numpy())+' Score:'+str(prediction[0]['scores'][i].cpu().numpy()))\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "python3",
   "language": "python",
   "name": "python3"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "11c306e3ffa2480d9e4a672e7971a9ee": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "ProgressStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "ProgressStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "bar_color": null,
      "description_width": "initial"
     }
    },
    "2327d674927f4c8196a9d548eabd25e6": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "323ece2fee2141359f2171724dec6785": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "DescriptionStyleModel",
     "state": {
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "DescriptionStyleModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "StyleView",
      "description_width": ""
     }
    },
    "70e0c557f9384eedba80c1b185d57756": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "7913e298350c4856a7cd2ff7e9ea4736": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HTMLModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HTMLModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HTMLView",
      "description": "",
      "description_tooltip": null,
      "layout": "IPY_MODEL_70e0c557f9384eedba80c1b185d57756",
      "placeholder": "​",
      "style": "IPY_MODEL_323ece2fee2141359f2171724dec6785",
      "value": " 170M/170M [00:11&lt;00:00, 15.0MB/s]"
     }
    },
    "a46d38d28bec4d56b3122f682e967fd2": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "FloatProgressModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "FloatProgressModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "ProgressView",
      "bar_style": "success",
      "description": "100%",
      "description_tooltip": null,
      "layout": "IPY_MODEL_2327d674927f4c8196a9d548eabd25e6",
      "max": 178090079,
      "min": 0,
      "orientation": "horizontal",
      "style": "IPY_MODEL_11c306e3ffa2480d9e4a672e7971a9ee",
      "value": 178090079
     }
    },
    "aeccac7e650b4e79bfe7f4b7afbf1057": {
     "model_module": "@jupyter-widgets/base",
     "model_name": "LayoutModel",
     "state": {
      "_model_module": "@jupyter-widgets/base",
      "_model_module_version": "1.2.0",
      "_model_name": "LayoutModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/base",
      "_view_module_version": "1.2.0",
      "_view_name": "LayoutView",
      "align_content": null,
      "align_items": null,
      "align_self": null,
      "border": null,
      "bottom": null,
      "display": null,
      "flex": null,
      "flex_flow": null,
      "grid_area": null,
      "grid_auto_columns": null,
      "grid_auto_flow": null,
      "grid_auto_rows": null,
      "grid_column": null,
      "grid_gap": null,
      "grid_row": null,
      "grid_template_areas": null,
      "grid_template_columns": null,
      "grid_template_rows": null,
      "height": null,
      "justify_content": null,
      "justify_items": null,
      "left": null,
      "margin": null,
      "max_height": null,
      "max_width": null,
      "min_height": null,
      "min_width": null,
      "object_fit": null,
      "object_position": null,
      "order": null,
      "overflow": null,
      "overflow_x": null,
      "overflow_y": null,
      "padding": null,
      "right": null,
      "top": null,
      "visibility": null,
      "width": null
     }
    },
    "d4df3b571db049f499241162165eaaad": {
     "model_module": "@jupyter-widgets/controls",
     "model_name": "HBoxModel",
     "state": {
      "_dom_classes": [],
      "_model_module": "@jupyter-widgets/controls",
      "_model_module_version": "1.5.0",
      "_model_name": "HBoxModel",
      "_view_count": null,
      "_view_module": "@jupyter-widgets/controls",
      "_view_module_version": "1.5.0",
      "_view_name": "HBoxView",
      "box_style": "",
      "children": [
       "IPY_MODEL_a46d38d28bec4d56b3122f682e967fd2",
       "IPY_MODEL_7913e298350c4856a7cd2ff7e9ea4736"
      ],
      "layout": "IPY_MODEL_aeccac7e650b4e79bfe7f4b7afbf1057"
     }
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}