yolov5-face/evaluate.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import torchvision.models as models\n",
    "import torch.onnx as onnx\n",
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from imutils import paths\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import PIL\n",
    "import time\n",
    "import cv2\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from models.experimental import attempt_load\n",
    "from utils.datasets import letterbox\n",
    "from utils.general import check_img_size, non_max_suppression_face, apply_classifier, scale_coords, xywh2xyxy, xyxy2xywh, \\\n",
    "    strip_optimizer, set_logging, increment_path\n",
    "from utils.plots import plot_one_box\n",
    "from utils.torch_utils import select_device, load_classifier, time_synchronized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_dir(dirName):\n",
    "    # Create a target directory & all intermediate \n",
    "    # directories if they don't exists\n",
    "    \n",
    "    if not os.path.exists(dirName):\n",
    "        os.makedirs(dirName, exist_ok = True)\n",
    "        print(\"[INFO] Directory \" ,dirName,  \" created\")\n",
    "    else:\n",
    "        print(\"[INFO] Directory \" ,dirName,  \" already exists\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "iCartoonFace test set size: 10000\n",
      "AnimalWeb test set size: 3752\n",
      "TFW-indoor test set size: 1080\n",
      "TFW-outdoor test set size: 800\n",
      "[INFO] Directory  Object-Detection-Metrics/results/test/exp_999/icartoon/detections_50_50  created\n",
      "[INFO] Directory  Object-Detection-Metrics/results/test/exp_999/animalweb/detections_50_50  created\n",
      "[INFO] Directory  Object-Detection-Metrics/results/test/exp_999/tfw_i/detections_50_50  created\n",
      "[INFO] Directory  Object-Detection-Metrics/results/test/exp_999/tfw_o/detections_50_50  created\n"
     ]
    }
   ],
   "source": [
    "exp_id = 999            # experiment ID\n",
    "iou_thr = 0.5           # IoU threshold\n",
    "con_thr = 0.5           # confidence threshold\n",
    "img_size = 640          # image sizes used at evaluation\n",
    "\n",
    "set_ = 'test'            # val / test / train\n",
    "weights = \"/workspace/yolov5-face-my/runs/train/exp1_l6/weights/best.pt\"    # path to the model weights\n",
    "\n",
    "# dataset directories\n",
    "ic_dir = '/workspace/dataset/icartoon/'\n",
    "aw_dir = '/workspace/dataset/animalweb/'\n",
    "tf_dir = '/workspace/dataset/tfw_split/'\n",
    "\n",
    "# Get paths to the dataset images\n",
    "ic_paths = list(paths.list_images(os.path.join(ic_dir, set_, 'images')))\n",
    "aw_paths = list(paths.list_images(os.path.join(aw_dir, set_, 'images')))\n",
    "tfi_paths = list(paths.list_images(os.path.join(tf_dir, set_, 'indoor', 'images')))\n",
    "tfo_paths = list(paths.list_images(os.path.join(tf_dir, set_, 'outdoor', 'images')))\n",
    "print(\"iCartoonFace {} set size: {}\". format(set_, len(ic_paths)))\n",
    "print(\"AnimalWeb {} set size: {}\". format(set_, len(aw_paths)))\n",
    "print(\"TFW-indoor {} set size: {}\". format(set_, len(tfi_paths)))\n",
    "print(\"TFW-outdoor {} set size: {}\". format(set_, len(tfo_paths)))\n",
    "\n",
    "# Create directories for saving evaluation results\n",
    "ic_res_path = \"Object-Detection-Metrics/results/{}/exp_{}/icartoon/detections_{}_{}\".format(set_, exp_id, \n",
    "                                                                                         int(iou_thr * 100),\n",
    "                                                                                         int(con_thr * 100))\n",
    "aw_res_path = \"Object-Detection-Metrics/results/{}/exp_{}/animalweb/detections_{}_{}\".format(set_, exp_id, \n",
    "                                                                                         int(iou_thr * 100),\n",
    "                                                                                         int(con_thr * 100))\n",
    "tfi_res_path = \"Object-Detection-Metrics/results/{}/exp_{}/tfw_i/detections_{}_{}\".format(set_, exp_id, \n",
    "                                                                                         int(iou_thr * 100),\n",
    "                                                                                         int(con_thr * 100))\n",
    "tfo_res_path = \"Object-Detection-Metrics/results/{}/exp_{}/tfw_o/detections_{}_{}\".format(set_, exp_id, \n",
    "                                                                                         int(iou_thr * 100),\n",
    "                                                                                         int(con_thr * 100))\n",
    "make_dir(ic_res_path)\n",
    "make_dir(aw_res_path)\n",
    "make_dir(tfi_res_path)\n",
    "make_dir(tfo_res_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fusing layers... \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(device(type='cuda', index=1),\n",
       " '/workspace/yolov5-face-my/runs/train/exp1_l6/weights/best.pt')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load weights \n",
    "device = torch.device(\"cuda:1\" if torch.cuda.is_available() else \"cpu\")\n",
    "model = attempt_load(weights, map_location=device)\n",
    "device, weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def scale_coords_landmarks(img1_shape, coords, img0_shape, ratio_pad=None):\n",
    "    # Rescale coords (xyxy) from img1_shape to img0_shape\n",
    "    \n",
    "    if ratio_pad is None:  # calculate from img0_shape\n",
    "        gain = min(img1_shape[0] / img0_shape[0], img1_shape[1] / img0_shape[1])  # gain  = old / new\n",
    "        pad = (img1_shape[1] - img0_shape[1] * gain) / 2, (img1_shape[0] - img0_shape[0] * gain) / 2  # wh padding\n",
    "    else:\n",
    "        gain = ratio_pad[0][0]\n",
    "        pad = ratio_pad[1]\n",
    "\n",
    "    coords[:, [0, 2, 4, 6, 8]] -= pad[0]  # x padding\n",
    "    coords[:, [1, 3, 5, 7, 9]] -= pad[1]  # y padding\n",
    "    coords[:, :10] /= gain\n",
    "    coords[:, 0].clamp_(0, img0_shape[1])  # x1\n",
    "    coords[:, 1].clamp_(0, img0_shape[0])  # y1\n",
    "    coords[:, 2].clamp_(0, img0_shape[1])  # x2\n",
    "    coords[:, 3].clamp_(0, img0_shape[0])  # y2\n",
    "    coords[:, 4].clamp_(0, img0_shape[1])  # x3\n",
    "    coords[:, 5].clamp_(0, img0_shape[0])  # y3\n",
    "    coords[:, 6].clamp_(0, img0_shape[1])  # x4\n",
    "    coords[:, 7].clamp_(0, img0_shape[0])  # y4\n",
    "    coords[:, 8].clamp_(0, img0_shape[1])  # x5\n",
    "    coords[:, 9].clamp_(0, img0_shape[0])  # y5\n",
    "    \n",
    "    return coords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_results(img, xywh, conf, landmarks):\n",
    "    # Draw detected bounding box with landmarks\n",
    "    \n",
    "    h,w,c = img.shape\n",
    "    \n",
    "    tl = 1 or round(0.002 * (h + w) / 2) + 1    # line/font thickness\n",
    "    clors = [(255,0,0),(0,255,0),(0,0,255),(255,255,0),(0,255,255)]    # colors for landmarks\n",
    "\n",
    "    # Convert YOLO coordinates to x1, y1, x2, y2 format\n",
    "    x1 = int(xywh[0] * w - 0.5 * xywh[2] * w)\n",
    "    y1 = int(xywh[1] * h - 0.5 * xywh[3] * h)\n",
    "    x2 = int(xywh[0] * w + 0.5 * xywh[2] * w)\n",
    "    y2 = int(xywh[1] * h + 0.5 * xywh[3] * h)\n",
    "    \n",
    "    # Draw bounding box\n",
    "    cv2.rectangle(img, (x1,y1), (x2, y2), (0,255,0), thickness=tl, lineType=cv2.LINE_AA)\n",
    "\n",
    "    # Draw facial landmarks\n",
    "    for i in range(5):\n",
    "        point_x = int(landmarks[2 * i] * w)\n",
    "        point_y = int(landmarks[2 * i + 1] * h)\n",
    "        cv2.circle(img, (point_x, point_y), tl+1, clors[i], -1)\n",
    "\n",
    "    # Put confidence threshold\n",
    "    tf = max(tl - 1, 1)  # font thickness\n",
    "    label = str(conf)[:5]\n",
    "    cv2.putText(img, label, (x1, y1 - 2), 0, tl / 3, [225, 255, 255], thickness=tf, lineType=cv2.LINE_AA)\n",
    "    \n",
    "    return img"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def detect_one(model, image_path, device, imgsz=640, conf_thres=0.5, iou_thres=0.5, augment=None, show=False):\n",
    "    # Detect face region and facial landmarks on given images\n",
    "    \n",
    "    # Load image and create a copy\n",
    "    orgimg = cv2.imread(image_path)\n",
    "    img0 = orgimg.copy()\n",
    "    \n",
    "    # Preprocess the image     \n",
    "    stride = int(model.stride.max())    \n",
    "    if imgsz <= 0:                    \n",
    "        imgsz = dynamic_resize(img0.shape)\n",
    "    imgsz = check_img_size(imgsz, s=64)\n",
    "    img = letterbox(img0, imgsz)[0]\n",
    "    img = img[:, :, ::-1].transpose(2, 0, 1)\n",
    "    img = np.ascontiguousarray(img)\n",
    "    img = torch.from_numpy(img).to(device)\n",
    "    img = img.float()\n",
    "    img /= 255.0\n",
    "    if img.ndimension() == 3:\n",
    "        img = img.unsqueeze(0)\n",
    "\n",
    "    # Inference\n",
    "    if augment is not None:\n",
    "        pred = model(img, augment=True, aug_setting=augment)[0]\n",
    "    else:\n",
    "        pred = model(img, augment=False)[0]   \n",
    "    \n",
    "    # Apply non max suppression\n",
    "    pred = non_max_suppression_face(pred, conf_thres, iou_thres)[0]\n",
    "    gn = torch.tensor(img0.shape)[[1, 0, 1, 0]].to(device)                        # normalization gain whwh\n",
    "    gn_lks = torch.tensor(img0.shape)[[1, 0, 1, 0, 1, 0, 1, 0, 1, 0]].to(device)  # normalization gain landmarks\n",
    "    \n",
    "    # Scale and store predictions\n",
    "    boxes = []\n",
    "    landmarks = []\n",
    "    conf_scores = []\n",
    "    h, w, c = img0.shape\n",
    "    if pred is not None:\n",
    "        pred[:, :4] = scale_coords(img.shape[2:], pred[:, :4], img0.shape).round()\n",
    "        pred[:, 5:15] = scale_coords_landmarks(img.shape[2:], pred[:, 5:15], img0.shape).round()\n",
    "        for j in range(pred.size()[0]):\n",
    "            xywh = (xyxy2xywh(pred[j, :4].view(1, 4)) / gn).view(-1)\n",
    "            xywh = xywh.data.cpu().numpy()\n",
    "            conf = pred[j, 4].cpu().numpy()\n",
    "            points = (pred[j, 5:15].view(1, 10) / gn_lks).view(-1).tolist()\n",
    "            class_num = pred[j, 15].cpu().numpy()\n",
    "            x1 = int(xywh[0] * w - 0.5 * xywh[2] * w)\n",
    "            y1 = int(xywh[1] * h - 0.5 * xywh[3] * h)\n",
    "            x2 = int(xywh[0] * w + 0.5 * xywh[2] * w)\n",
    "            y2 = int(xywh[1] * h + 0.5 * xywh[3] * h)\n",
    "                \n",
    "            boxes.append([x1, y1, x2-x1, y2-y1])\n",
    "            landmarks.append(points)\n",
    "            conf_scores.append(conf)\n",
    "            \n",
    "            if show:\n",
    "                orgimg = show_results(orgimg, xywh, conf, points)\n",
    "            \n",
    "    if show:\n",
    "        img = PIL.Image.fromarray(cv2.cvtColor(orgimg, cv2.COLOR_BGR2RGB))\n",
    "        img.show()\n",
    "        \n",
    "    return boxes, landmarks, conf_scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<PIL.Image.Image image mode=RGB size=464x348 at 0x7F6C8287B198>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "([[235, 43, 29, 37]],\n",
       " [[0.5237069129943848,\n",
       "   0.16091954708099365,\n",
       "   0.5538793206214905,\n",
       "   0.16091954708099365,\n",
       "   0.5387930870056152,\n",
       "   0.18103447556495667,\n",
       "   0.5280172228813171,\n",
       "   0.1982758641242981,\n",
       "   0.5495689511299133,\n",
       "   0.1982758641242981]],\n",
       " [array(    0.87606, dtype=float32)])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "detect_one(model, tfo_paths[1], device, imgsz=img_size, show=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def evaluate(imagePaths, res_path, model, device, img_size, con_thr, iou_thr, augment=None):\n",
    "    # Evaluate model performance on a dataset\n",
    "    \n",
    "    for ind, imagePath in enumerate(imagePaths, 1):\n",
    "        print(\"Processing image: {}/{}\".format(ind, len(imagePaths)))\n",
    "        \n",
    "        # Detect\n",
    "        boxes, pred_landmarks, scores = detect_one(model, imagePath, device, \n",
    "                                                                 img_size, con_thr, iou_thr, augment=augment)\n",
    "\n",
    "        # Open the file to  write the results in absolute Object-Detection-Metrics format\n",
    "        # \"face [conf_score] [x1] [x2] [w] [h]\"        \n",
    "        imageName = imagePath.split('/')[-1].split('.')[0]\n",
    "        det_file = open(\"{}/{}.txt\".format(res_path, imageName), \"w\")\n",
    "    \n",
    "        # Iterate through the detected bounding boxes and conf scores\n",
    "        # and write to the txt file\n",
    "        for pred_box, conf_score in zip(boxes, scores):\n",
    "            det_file.write('face')\n",
    "            det_file.write(\" \")\n",
    "            det_file.write(str(conf_score))\n",
    "            det_file.write(\" \")\n",
    "            det_file.write(str(pred_box[0]))\n",
    "            det_file.write(\" \")\n",
    "            det_file.write(str(pred_box[1]))\n",
    "            det_file.write(\" \")\n",
    "            det_file.write(str(pred_box[2]))\n",
    "            det_file.write(\" \")\n",
    "            det_file.write(str(pred_box[3]))\n",
    "            det_file.write(\"\\n\")\n",
    "                \n",
    "        det_file.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1/800\n",
      "Processing image: 2/800\n",
      "Processing image: 3/800\n",
      "Processing image: 4/800\n",
      "Processing image: 5/800\n",
      "Processing image: 6/800\n",
      "Processing image: 7/800\n",
      "Processing image: 8/800\n",
      "Processing image: 9/800\n",
      "Processing image: 10/800\n",
      "Processing image: 11/800\n",
      "Processing image: 12/800\n",
      "Processing image: 13/800\n",
      "Processing image: 14/800\n",
      "Processing image: 15/800\n",
      "Processing image: 16/800\n",
      "Processing image: 17/800\n",
      "Processing image: 18/800\n",
      "Processing image: 19/800\n",
      "Processing image: 20/800\n",
      "Processing image: 21/800\n",
      "Processing image: 22/800\n",
      "Processing image: 23/800\n",
      "Processing image: 24/800\n",
      "Processing image: 25/800\n",
      "Processing image: 26/800\n",
      "Processing image: 27/800\n",
      "Processing image: 28/800\n",
      "Processing image: 29/800\n",
      "Processing image: 30/800\n",
      "Processing image: 31/800\n",
      "Processing image: 32/800\n",
      "Processing image: 33/800\n",
      "Processing image: 34/800\n",
      "Processing image: 35/800\n",
      "Processing image: 36/800\n",
      "Processing image: 37/800\n",
      "Processing image: 38/800\n",
      "Processing image: 39/800\n",
      "Processing image: 40/800\n",
      "Processing image: 41/800\n",
      "Processing image: 42/800\n",
      "Processing image: 43/800\n",
      "Processing image: 44/800\n",
      "Processing image: 45/800\n",
      "Processing image: 46/800\n",
      "Processing image: 47/800\n",
      "Processing image: 48/800\n",
      "Processing image: 49/800\n",
      "Processing image: 50/800\n",
      "Processing image: 51/800\n",
      "Processing image: 52/800\n",
      "Processing image: 53/800\n",
      "Processing image: 54/800\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-13-c430cd13e853>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtfo_paths\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtfo_res_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdevice\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimg_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcon_thr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0miou_thr\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maugment\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\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<ipython-input-12-f50748f0f090>\u001b[0m in \u001b[0;36mevaluate\u001b[0;34m(imagePaths, res_path, model, device, img_size, con_thr, iou_thr, augment)\u001b[0m\n\u001b[1;32m      7\u001b[0m         \u001b[0;31m# Detect\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m         boxes, pred_landmarks, scores = detect_one(model, imagePath, device, \n\u001b[0;32m----> 9\u001b[0;31m                                                                  img_size, con_thr, iou_thr, augment=augment)\n\u001b[0m\u001b[1;32m     10\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     11\u001b[0m         \u001b[0;31m# Open the file to  write the results in absolute Object-Detection-Metrics format\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-9-05c09dd984dd>\u001b[0m in \u001b[0;36mdetect_one\u001b[0;34m(model, image_path, device, imgsz, conf_thres, iou_thres, augment, show)\u001b[0m\n\u001b[1;32m     22\u001b[0m     \u001b[0;31m# Inference\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     23\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0maugment\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 24\u001b[0;31m         \u001b[0mpred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maugment\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maug_setting\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0maugment\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     25\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     26\u001b[0m         \u001b[0mpred\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maugment\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.6/site-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/workspace/yolov5-face/models/yolo.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, x, augment, profile, aug_setting, feature_vis)\u001b[0m\n\u001b[1;32m    174\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0msi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mf\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    175\u001b[0m                 \u001b[0mxi\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mscale_img\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mflip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfi\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfi\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msame_shape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 176\u001b[0;31m                 \u001b[0myi\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward_once\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeature_vis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mfeature_vis\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m  \u001b[0;31m# forward\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    177\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    178\u001b[0m                 \u001b[0myi\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m...\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m:\u001b[0m\u001b[0;36m4\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m/=\u001b[0m \u001b[0msi\u001b[0m  \u001b[0;31m# de-scale\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/workspace/yolov5-face/models/yolo.py\u001b[0m in \u001b[0;36mforward_once\u001b[0;34m(self, x, profile, feature_vis)\u001b[0m\n\u001b[1;32m    200\u001b[0m                 \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'%10.1f%10.0f%10.1fms %-40s'\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mo\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtype\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    201\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 202\u001b[0;31m             \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# run\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    203\u001b[0m             \u001b[0my\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msave\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# save output\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    204\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.6/site-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/workspace/yolov5-face/models/common.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m    109\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    110\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\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--> 111\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv3\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdim\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    112\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    113\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mShuffleV2Block\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mModule\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/opt/conda/lib/python3.6/site-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/opt/conda/lib/python3.6/site-packages/torch/nn/modules/container.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m    115\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\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    116\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 117\u001b[0;31m             \u001b[0minput\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodule\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\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    118\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    119\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/opt/conda/lib/python3.6/site-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/workspace/yolov5-face/models/common.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     77\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     78\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\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---> 79\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mx\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcv1\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     80\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     81\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mBottleneckCSP\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mModule\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/opt/conda/lib/python3.6/site-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/workspace/yolov5-face/models/common.py\u001b[0m in \u001b[0;36mfuseforward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     48\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     49\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mfuseforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\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---> 50\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconv\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     51\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     52\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mStemBlock\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mModule\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/opt/conda/lib/python3.6/site-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/opt/conda/lib/python3.6/site-packages/torch/nn/modules/conv.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m    421\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    422\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m)\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[0m\n\u001b[0;32m--> 423\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_conv_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweight\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    424\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    425\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mConv3d\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_ConvNd\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/opt/conda/lib/python3.6/site-packages/torch/nn/modules/conv.py\u001b[0m in \u001b[0;36m_conv_forward\u001b[0;34m(self, input, weight)\u001b[0m\n\u001b[1;32m    418\u001b[0m                             _pair(0), self.dilation, self.groups)\n\u001b[1;32m    419\u001b[0m         return F.conv2d(input, weight, self.bias, self.stride,\n\u001b[0;32m--> 420\u001b[0;31m                         self.padding, self.dilation, self.groups)\n\u001b[0m\u001b[1;32m    421\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    422\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m)\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[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "evaluate(tfo_paths, tfo_res_path, model, device, img_size, con_thr, iou_thr, augment=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1/432\n",
      "Processing image: 2/432\n",
      "Processing image: 3/432\n",
      "Processing image: 4/432\n",
      "Processing image: 5/432\n",
      "Processing image: 6/432\n",
      "Processing image: 7/432\n",
      "Processing image: 8/432\n",
      "Processing image: 9/432\n",
      "Processing image: 10/432\n",
      "Processing image: 11/432\n",
      "Processing image: 12/432\n",
      "Processing image: 13/432\n",
      "Processing image: 14/432\n",
      "Processing image: 15/432\n",
      "Processing image: 16/432\n",
      "Processing image: 17/432\n",
      "Processing image: 18/432\n",
      "Processing image: 19/432\n",
      "Processing image: 20/432\n",
      "Processing image: 21/432\n",
      "Processing image: 22/432\n",
      "Processing image: 23/432\n",
      "Processing image: 24/432\n",
      "Processing image: 25/432\n",
      "Processing image: 26/432\n",
      "Processing image: 27/432\n",
      "Processing image: 28/432\n",
      "Processing image: 29/432\n",
      "Processing image: 30/432\n",
      "Processing image: 31/432\n",
      "Processing image: 32/432\n",
      "Processing image: 33/432\n",
      "Processing image: 34/432\n",
      "Processing image: 35/432\n",
      "Processing image: 36/432\n",
      "Processing image: 37/432\n",
      "Processing image: 38/432\n",
      "Processing image: 39/432\n",
      "Processing image: 40/432\n",
      "Processing image: 41/432\n",
      "Processing image: 42/432\n",
      "Processing image: 43/432\n",
      "Processing image: 44/432\n",
      "Processing image: 45/432\n",
      "Processing image: 46/432\n",
      "Processing image: 47/432\n",
      "Processing image: 48/432\n",
      "Processing image: 49/432\n",
      "Processing image: 50/432\n",
      "Processing image: 51/432\n",
      "Processing image: 52/432\n",
      "Processing image: 53/432\n",
      "Processing image: 54/432\n",
      "Processing image: 55/432\n",
      "Processing image: 56/432\n",
      "Processing image: 57/432\n",
      "Processing image: 58/432\n",
      "Processing image: 59/432\n",
      "Processing image: 60/432\n",
      "Processing image: 61/432\n",
      "Processing image: 62/432\n",
      "Processing image: 63/432\n",
      "Processing image: 64/432\n",
      "Processing image: 65/432\n",
      "Processing image: 66/432\n",
      "Processing image: 67/432\n",
      "Processing image: 68/432\n",
      "Processing image: 69/432\n",
      "Processing image: 70/432\n",
      "Processing image: 71/432\n",
      "Processing image: 72/432\n",
      "Processing image: 73/432\n",
      "Processing image: 74/432\n",
      "Processing image: 75/432\n",
      "Processing image: 76/432\n",
      "Processing image: 77/432\n",
      "Processing image: 78/432\n",
      "Processing image: 79/432\n",
      "Processing image: 80/432\n",
      "Processing image: 81/432\n",
      "Processing image: 82/432\n",
      "Processing image: 83/432\n",
      "Processing image: 84/432\n",
      "Processing image: 85/432\n",
      "Processing image: 86/432\n",
      "Processing image: 87/432\n",
      "Processing image: 88/432\n",
      "Processing image: 89/432\n",
      "Processing image: 90/432\n",
      "Processing image: 91/432\n",
      "Processing image: 92/432\n",
      "Processing image: 93/432\n",
      "Processing image: 94/432\n",
      "Processing image: 95/432\n",
      "Processing image: 96/432\n",
      "Processing image: 97/432\n",
      "Processing image: 98/432\n",
      "Processing image: 99/432\n",
      "Processing image: 100/432\n",
      "Processing image: 101/432\n",
      "Processing image: 102/432\n",
      "Processing image: 103/432\n",
      "Processing image: 104/432\n",
      "Processing image: 105/432\n",
      "Processing image: 106/432\n",
      "Processing image: 107/432\n",
      "Processing image: 108/432\n",
      "Processing image: 109/432\n",
      "Processing image: 110/432\n",
      "Processing image: 111/432\n",
      "Processing image: 112/432\n",
      "Processing image: 113/432\n",
      "Processing image: 114/432\n",
      "Processing image: 115/432\n",
      "Processing image: 116/432\n",
      "Processing image: 117/432\n",
      "Processing image: 118/432\n",
      "Processing image: 119/432\n",
      "Processing image: 120/432\n",
      "Processing image: 121/432\n",
      "Processing image: 122/432\n",
      "Processing image: 123/432\n",
      "Processing image: 124/432\n",
      "Processing image: 125/432\n",
      "Processing image: 126/432\n",
      "Processing image: 127/432\n",
      "Processing image: 128/432\n",
      "Processing image: 129/432\n",
      "Processing image: 130/432\n",
      "Processing image: 131/432\n",
      "Processing image: 132/432\n",
      "Processing image: 133/432\n",
      "Processing image: 134/432\n",
      "Processing image: 135/432\n",
      "Processing image: 136/432\n",
      "Processing image: 137/432\n",
      "Processing image: 138/432\n",
      "Processing image: 139/432\n",
      "Processing image: 140/432\n",
      "Processing image: 141/432\n",
      "Processing image: 142/432\n",
      "Processing image: 143/432\n",
      "Processing image: 144/432\n",
      "Processing image: 145/432\n",
      "Processing image: 146/432\n",
      "Processing image: 147/432\n",
      "Processing image: 148/432\n",
      "Processing image: 149/432\n",
      "Processing image: 150/432\n",
      "Processing image: 151/432\n",
      "Processing image: 152/432\n",
      "Processing image: 153/432\n",
      "Processing image: 154/432\n",
      "Processing image: 155/432\n",
      "Processing image: 156/432\n",
      "Processing image: 157/432\n",
      "Processing image: 158/432\n",
      "Processing image: 159/432\n",
      "Processing image: 160/432\n",
      "Processing image: 161/432\n",
      "Processing image: 162/432\n",
      "Processing image: 163/432\n",
      "Processing image: 164/432\n",
      "Processing image: 165/432\n",
      "Processing image: 166/432\n",
      "Processing image: 167/432\n",
      "Processing image: 168/432\n",
      "Processing image: 169/432\n",
      "Processing image: 170/432\n",
      "Processing image: 171/432\n",
      "Processing image: 172/432\n",
      "Processing image: 173/432\n",
      "Processing image: 174/432\n",
      "Processing image: 175/432\n",
      "Processing image: 176/432\n",
      "Processing image: 177/432\n",
      "Processing image: 178/432\n",
      "Processing image: 179/432\n",
      "Processing image: 180/432\n",
      "Processing image: 181/432\n",
      "Processing image: 182/432\n",
      "Processing image: 183/432\n",
      "Processing image: 184/432\n",
      "Processing image: 185/432\n",
      "Processing image: 186/432\n",
      "Processing image: 187/432\n",
      "Processing image: 188/432\n",
      "Processing image: 189/432\n",
      "Processing image: 190/432\n",
      "Processing image: 191/432\n",
      "Processing image: 192/432\n",
      "Processing image: 193/432\n",
      "Processing image: 194/432\n",
      "Processing image: 195/432\n",
      "Processing image: 196/432\n",
      "Processing image: 197/432\n",
      "Processing image: 198/432\n",
      "Processing image: 199/432\n",
      "Processing image: 200/432\n",
      "Processing image: 201/432\n",
      "Processing image: 202/432\n",
      "Processing image: 203/432\n",
      "Processing image: 204/432\n",
      "Processing image: 205/432\n",
      "Processing image: 206/432\n",
      "Processing image: 207/432\n",
      "Processing image: 208/432\n",
      "Processing image: 209/432\n",
      "Processing image: 210/432\n",
      "Processing image: 211/432\n",
      "Processing image: 212/432\n",
      "Processing image: 213/432\n",
      "Processing image: 214/432\n",
      "Processing image: 215/432\n",
      "Processing image: 216/432\n",
      "Processing image: 217/432\n",
      "Processing image: 218/432\n",
      "Processing image: 219/432\n",
      "Processing image: 220/432\n",
      "Processing image: 221/432\n",
      "Processing image: 222/432\n",
      "Processing image: 223/432\n",
      "Processing image: 224/432\n",
      "Processing image: 225/432\n",
      "Processing image: 226/432\n",
      "Processing image: 227/432\n",
      "Processing image: 228/432\n",
      "Processing image: 229/432\n",
      "Processing image: 230/432\n",
      "Processing image: 231/432\n",
      "Processing image: 232/432\n",
      "Processing image: 233/432\n",
      "Processing image: 234/432\n",
      "Processing image: 235/432\n",
      "Processing image: 236/432\n",
      "Processing image: 237/432\n",
      "Processing image: 238/432\n",
      "Processing image: 239/432\n",
      "Processing image: 240/432\n",
      "Processing image: 241/432\n",
      "Processing image: 242/432\n",
      "Processing image: 243/432\n",
      "Processing image: 244/432\n",
      "Processing image: 245/432\n",
      "Processing image: 246/432\n",
      "Processing image: 247/432\n",
      "Processing image: 248/432\n",
      "Processing image: 249/432\n",
      "Processing image: 250/432\n",
      "Processing image: 251/432\n",
      "Processing image: 252/432\n",
      "Processing image: 253/432\n",
      "Processing image: 254/432\n",
      "Processing image: 255/432\n",
      "Processing image: 256/432\n",
      "Processing image: 257/432\n",
      "Processing image: 258/432\n",
      "Processing image: 259/432\n",
      "Processing image: 260/432\n",
      "Processing image: 261/432\n",
      "Processing image: 262/432\n",
      "Processing image: 263/432\n",
      "Processing image: 264/432\n",
      "Processing image: 265/432\n",
      "Processing image: 266/432\n",
      "Processing image: 267/432\n",
      "Processing image: 268/432\n",
      "Processing image: 269/432\n",
      "Processing image: 270/432\n",
      "Processing image: 271/432\n",
      "Processing image: 272/432\n",
      "Processing image: 273/432\n",
      "Processing image: 274/432\n",
      "Processing image: 275/432\n",
      "Processing image: 276/432\n",
      "Processing image: 277/432\n",
      "Processing image: 278/432\n",
      "Processing image: 279/432\n",
      "Processing image: 280/432\n",
      "Processing image: 281/432\n",
      "Processing image: 282/432\n",
      "Processing image: 283/432\n",
      "Processing image: 284/432\n",
      "Processing image: 285/432\n",
      "Processing image: 286/432\n",
      "Processing image: 287/432\n",
      "Processing image: 288/432\n",
      "Processing image: 289/432\n",
      "Processing image: 290/432\n",
      "Processing image: 291/432\n",
      "Processing image: 292/432\n",
      "Processing image: 293/432\n",
      "Processing image: 294/432\n",
      "Processing image: 295/432\n",
      "Processing image: 296/432\n",
      "Processing image: 297/432\n",
      "Processing image: 298/432\n",
      "Processing image: 299/432\n",
      "Processing image: 300/432\n",
      "Processing image: 301/432\n",
      "Processing image: 302/432\n",
      "Processing image: 303/432\n",
      "Processing image: 304/432\n",
      "Processing image: 305/432\n",
      "Processing image: 306/432\n",
      "Processing image: 307/432\n",
      "Processing image: 308/432\n",
      "Processing image: 309/432\n",
      "Processing image: 310/432\n",
      "Processing image: 311/432\n",
      "Processing image: 312/432\n",
      "Processing image: 313/432\n",
      "Processing image: 314/432\n",
      "Processing image: 315/432\n",
      "Processing image: 316/432\n",
      "Processing image: 317/432\n",
      "Processing image: 318/432\n",
      "Processing image: 319/432\n",
      "Processing image: 320/432\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 321/432\n",
      "Processing image: 322/432\n",
      "Processing image: 323/432\n",
      "Processing image: 324/432\n",
      "Processing image: 325/432\n",
      "Processing image: 326/432\n",
      "Processing image: 327/432\n",
      "Processing image: 328/432\n",
      "Processing image: 329/432\n",
      "Processing image: 330/432\n",
      "Processing image: 331/432\n",
      "Processing image: 332/432\n",
      "Processing image: 333/432\n",
      "Processing image: 334/432\n",
      "Processing image: 335/432\n",
      "Processing image: 336/432\n",
      "Processing image: 337/432\n",
      "Processing image: 338/432\n",
      "Processing image: 339/432\n",
      "Processing image: 340/432\n",
      "Processing image: 341/432\n",
      "Processing image: 342/432\n",
      "Processing image: 343/432\n",
      "Processing image: 344/432\n",
      "Processing image: 345/432\n",
      "Processing image: 346/432\n",
      "Processing image: 347/432\n",
      "Processing image: 348/432\n",
      "Processing image: 349/432\n",
      "Processing image: 350/432\n",
      "Processing image: 351/432\n",
      "Processing image: 352/432\n",
      "Processing image: 353/432\n",
      "Processing image: 354/432\n",
      "Processing image: 355/432\n",
      "Processing image: 356/432\n",
      "Processing image: 357/432\n",
      "Processing image: 358/432\n",
      "Processing image: 359/432\n",
      "Processing image: 360/432\n",
      "Processing image: 361/432\n",
      "Processing image: 362/432\n",
      "Processing image: 363/432\n",
      "Processing image: 364/432\n",
      "Processing image: 365/432\n",
      "Processing image: 366/432\n",
      "Processing image: 367/432\n",
      "Processing image: 368/432\n",
      "Processing image: 369/432\n",
      "Processing image: 370/432\n",
      "Processing image: 371/432\n",
      "Processing image: 372/432\n",
      "Processing image: 373/432\n",
      "Processing image: 374/432\n",
      "Processing image: 375/432\n",
      "Processing image: 376/432\n",
      "Processing image: 377/432\n",
      "Processing image: 378/432\n",
      "Processing image: 379/432\n",
      "Processing image: 380/432\n",
      "Processing image: 381/432\n",
      "Processing image: 382/432\n",
      "Processing image: 383/432\n",
      "Processing image: 384/432\n",
      "Processing image: 385/432\n",
      "Processing image: 386/432\n",
      "Processing image: 387/432\n",
      "Processing image: 388/432\n",
      "Processing image: 389/432\n",
      "Processing image: 390/432\n",
      "Processing image: 391/432\n",
      "Processing image: 392/432\n",
      "Processing image: 393/432\n",
      "Processing image: 394/432\n",
      "Processing image: 395/432\n",
      "Processing image: 396/432\n",
      "Processing image: 397/432\n",
      "Processing image: 398/432\n",
      "Processing image: 399/432\n",
      "Processing image: 400/432\n",
      "Processing image: 401/432\n",
      "Processing image: 402/432\n",
      "Processing image: 403/432\n",
      "Processing image: 404/432\n",
      "Processing image: 405/432\n",
      "Processing image: 406/432\n",
      "Processing image: 407/432\n",
      "Processing image: 408/432\n",
      "Processing image: 409/432\n",
      "Processing image: 410/432\n",
      "Processing image: 411/432\n",
      "Processing image: 412/432\n",
      "Processing image: 413/432\n",
      "Processing image: 414/432\n",
      "Processing image: 415/432\n",
      "Processing image: 416/432\n",
      "Processing image: 417/432\n",
      "Processing image: 418/432\n",
      "Processing image: 419/432\n",
      "Processing image: 420/432\n",
      "Processing image: 421/432\n",
      "Processing image: 422/432\n",
      "Processing image: 423/432\n",
      "Processing image: 424/432\n",
      "Processing image: 425/432\n",
      "Processing image: 426/432\n",
      "Processing image: 427/432\n",
      "Processing image: 428/432\n",
      "Processing image: 429/432\n",
      "Processing image: 430/432\n",
      "Processing image: 431/432\n",
      "Processing image: 432/432\n"
     ]
    }
   ],
   "source": [
    "evaluate(tfi_paths, tfi_res_path, model, device, img_size, con_thr, iou_thr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1/2062\n",
      "Processing image: 2/2062\n",
      "Processing image: 3/2062\n",
      "Processing image: 4/2062\n",
      "Processing image: 5/2062\n",
      "Processing image: 6/2062\n",
      "Processing image: 7/2062\n",
      "Processing image: 8/2062\n",
      "Processing image: 9/2062\n",
      "Processing image: 10/2062\n",
      "Processing image: 11/2062\n",
      "Processing image: 12/2062\n",
      "Processing image: 13/2062\n",
      "Processing image: 14/2062\n",
      "Processing image: 15/2062\n",
      "Processing image: 16/2062\n",
      "Processing image: 17/2062\n",
      "Processing image: 18/2062\n",
      "Processing image: 19/2062\n",
      "Processing image: 20/2062\n",
      "Processing image: 21/2062\n",
      "Processing image: 22/2062\n",
      "Processing image: 23/2062\n",
      "Processing image: 24/2062\n",
      "Processing image: 25/2062\n",
      "Processing image: 26/2062\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:49: RuntimeWarning: invalid value encountered in true_divide\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 27/2062\n",
      "Processing image: 28/2062\n",
      "Processing image: 29/2062\n",
      "Processing image: 30/2062\n",
      "Processing image: 31/2062\n",
      "Processing image: 32/2062\n",
      "Processing image: 33/2062\n",
      "Processing image: 34/2062\n",
      "Processing image: 35/2062\n",
      "Processing image: 36/2062\n",
      "Processing image: 37/2062\n",
      "Processing image: 38/2062\n",
      "Processing image: 39/2062\n",
      "Processing image: 40/2062\n",
      "Processing image: 41/2062\n",
      "Processing image: 42/2062\n",
      "Processing image: 43/2062\n",
      "Processing image: 44/2062\n",
      "Processing image: 45/2062\n",
      "Processing image: 46/2062\n",
      "Processing image: 47/2062\n",
      "Processing image: 48/2062\n",
      "Processing image: 49/2062\n",
      "Processing image: 50/2062\n",
      "Processing image: 51/2062\n",
      "Processing image: 52/2062\n",
      "Processing image: 53/2062\n",
      "Processing image: 54/2062\n",
      "Processing image: 55/2062\n",
      "Processing image: 56/2062\n",
      "Processing image: 57/2062\n",
      "Processing image: 58/2062\n",
      "Processing image: 59/2062\n",
      "Processing image: 60/2062\n",
      "Processing image: 61/2062\n",
      "Processing image: 62/2062\n",
      "Processing image: 63/2062\n",
      "Processing image: 64/2062\n",
      "Processing image: 65/2062\n",
      "Processing image: 66/2062\n",
      "Processing image: 67/2062\n",
      "Processing image: 68/2062\n",
      "Processing image: 69/2062\n",
      "Processing image: 70/2062\n",
      "Processing image: 71/2062\n",
      "Processing image: 72/2062\n",
      "Processing image: 73/2062\n",
      "Processing image: 74/2062\n",
      "Processing image: 75/2062\n",
      "Processing image: 76/2062\n",
      "Processing image: 77/2062\n",
      "Processing image: 78/2062\n",
      "Processing image: 79/2062\n",
      "Processing image: 80/2062\n",
      "Processing image: 81/2062\n",
      "Processing image: 82/2062\n",
      "Processing image: 83/2062\n",
      "Processing image: 84/2062\n",
      "Processing image: 85/2062\n",
      "Processing image: 86/2062\n",
      "Processing image: 87/2062\n",
      "Processing image: 88/2062\n",
      "Processing image: 89/2062\n",
      "Processing image: 90/2062\n",
      "Processing image: 91/2062\n",
      "Processing image: 92/2062\n",
      "Processing image: 93/2062\n",
      "Processing image: 94/2062\n",
      "Processing image: 95/2062\n",
      "Processing image: 96/2062\n",
      "Processing image: 97/2062\n",
      "Processing image: 98/2062\n",
      "Processing image: 99/2062\n",
      "Processing image: 100/2062\n",
      "Processing image: 101/2062\n",
      "Processing image: 102/2062\n",
      "Processing image: 103/2062\n",
      "Processing image: 104/2062\n",
      "Processing image: 105/2062\n",
      "Processing image: 106/2062\n",
      "Processing image: 107/2062\n",
      "Processing image: 108/2062\n",
      "Processing image: 109/2062\n",
      "Processing image: 110/2062\n",
      "Processing image: 111/2062\n",
      "Processing image: 112/2062\n",
      "Processing image: 113/2062\n",
      "Processing image: 114/2062\n",
      "Processing image: 115/2062\n",
      "Processing image: 116/2062\n",
      "Processing image: 117/2062\n",
      "Processing image: 118/2062\n",
      "Processing image: 119/2062\n",
      "Processing image: 120/2062\n",
      "Processing image: 121/2062\n",
      "Processing image: 122/2062\n",
      "Processing image: 123/2062\n",
      "Processing image: 124/2062\n",
      "Processing image: 125/2062\n",
      "Processing image: 126/2062\n",
      "Processing image: 127/2062\n",
      "Processing image: 128/2062\n",
      "Processing image: 129/2062\n",
      "Processing image: 130/2062\n",
      "Processing image: 131/2062\n",
      "Processing image: 132/2062\n",
      "Processing image: 133/2062\n",
      "Processing image: 134/2062\n",
      "Processing image: 135/2062\n",
      "Processing image: 136/2062\n",
      "Processing image: 137/2062\n",
      "Processing image: 138/2062\n",
      "Processing image: 139/2062\n",
      "Processing image: 140/2062\n",
      "Processing image: 141/2062\n",
      "Processing image: 142/2062\n",
      "Processing image: 143/2062\n",
      "Processing image: 144/2062\n",
      "Processing image: 145/2062\n",
      "Processing image: 146/2062\n",
      "Processing image: 147/2062\n",
      "Processing image: 148/2062\n",
      "Processing image: 149/2062\n",
      "Processing image: 150/2062\n",
      "Processing image: 151/2062\n",
      "Processing image: 152/2062\n",
      "Processing image: 153/2062\n",
      "Processing image: 154/2062\n",
      "Processing image: 155/2062\n",
      "Processing image: 156/2062\n",
      "Processing image: 157/2062\n",
      "Processing image: 158/2062\n",
      "Processing image: 159/2062\n",
      "Processing image: 160/2062\n",
      "Processing image: 161/2062\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:51: RuntimeWarning: invalid value encountered in less\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 162/2062\n",
      "Processing image: 163/2062\n",
      "Processing image: 164/2062\n",
      "Processing image: 165/2062\n",
      "Processing image: 166/2062\n",
      "Processing image: 167/2062\n",
      "Processing image: 168/2062\n",
      "Processing image: 169/2062\n",
      "Processing image: 170/2062\n",
      "Processing image: 171/2062\n",
      "Processing image: 172/2062\n",
      "Processing image: 173/2062\n",
      "Processing image: 174/2062\n",
      "Processing image: 175/2062\n",
      "Processing image: 176/2062\n",
      "Processing image: 177/2062\n",
      "Processing image: 178/2062\n",
      "Processing image: 179/2062\n",
      "Processing image: 180/2062\n",
      "Processing image: 181/2062\n",
      "Processing image: 182/2062\n",
      "Processing image: 183/2062\n",
      "Processing image: 184/2062\n",
      "Processing image: 185/2062\n",
      "Processing image: 186/2062\n",
      "Processing image: 187/2062\n",
      "Processing image: 188/2062\n",
      "Processing image: 189/2062\n",
      "Processing image: 190/2062\n",
      "Processing image: 191/2062\n",
      "Processing image: 192/2062\n",
      "Processing image: 193/2062\n",
      "Processing image: 194/2062\n",
      "Processing image: 195/2062\n",
      "Processing image: 196/2062\n",
      "Processing image: 197/2062\n",
      "Processing image: 198/2062\n",
      "Processing image: 199/2062\n",
      "Processing image: 200/2062\n",
      "Processing image: 201/2062\n",
      "Processing image: 202/2062\n",
      "Processing image: 203/2062\n",
      "Processing image: 204/2062\n",
      "Processing image: 205/2062\n",
      "Processing image: 206/2062\n",
      "Processing image: 207/2062\n",
      "Processing image: 208/2062\n",
      "Processing image: 209/2062\n",
      "Processing image: 210/2062\n",
      "Processing image: 211/2062\n",
      "Processing image: 212/2062\n",
      "Processing image: 213/2062\n",
      "Processing image: 214/2062\n",
      "Processing image: 215/2062\n",
      "Processing image: 216/2062\n",
      "Processing image: 217/2062\n",
      "Processing image: 218/2062\n",
      "Processing image: 219/2062\n",
      "Processing image: 220/2062\n",
      "Processing image: 221/2062\n",
      "Processing image: 222/2062\n",
      "Processing image: 223/2062\n",
      "Processing image: 224/2062\n",
      "Processing image: 225/2062\n",
      "Processing image: 226/2062\n",
      "Processing image: 227/2062\n",
      "Processing image: 228/2062\n",
      "Processing image: 229/2062\n",
      "Processing image: 230/2062\n",
      "Processing image: 231/2062\n",
      "Processing image: 232/2062\n",
      "Processing image: 233/2062\n",
      "Processing image: 234/2062\n",
      "Processing image: 235/2062\n",
      "Processing image: 236/2062\n",
      "Processing image: 237/2062\n",
      "Processing image: 238/2062\n",
      "Processing image: 239/2062\n",
      "Processing image: 240/2062\n",
      "Processing image: 241/2062\n",
      "Processing image: 242/2062\n",
      "Processing image: 243/2062\n",
      "Processing image: 244/2062\n",
      "Processing image: 245/2062\n",
      "Processing image: 246/2062\n",
      "Processing image: 247/2062\n",
      "Processing image: 248/2062\n",
      "Processing image: 249/2062\n",
      "Processing image: 250/2062\n",
      "Processing image: 251/2062\n",
      "Processing image: 252/2062\n",
      "Processing image: 253/2062\n",
      "Processing image: 254/2062\n",
      "Processing image: 255/2062\n",
      "Processing image: 256/2062\n",
      "Processing image: 257/2062\n",
      "Processing image: 258/2062\n",
      "Processing image: 259/2062\n",
      "Processing image: 260/2062\n",
      "Processing image: 261/2062\n",
      "Processing image: 262/2062\n",
      "Processing image: 263/2062\n",
      "Processing image: 264/2062\n",
      "Processing image: 265/2062\n",
      "Processing image: 266/2062\n",
      "Processing image: 267/2062\n",
      "Processing image: 268/2062\n",
      "Processing image: 269/2062\n",
      "Processing image: 270/2062\n",
      "Processing image: 271/2062\n",
      "Processing image: 272/2062\n",
      "Processing image: 273/2062\n",
      "Processing image: 274/2062\n",
      "Processing image: 275/2062\n",
      "Processing image: 276/2062\n",
      "Processing image: 277/2062\n",
      "Processing image: 278/2062\n",
      "Processing image: 279/2062\n",
      "Processing image: 280/2062\n",
      "Processing image: 281/2062\n",
      "Processing image: 282/2062\n",
      "Processing image: 283/2062\n",
      "Processing image: 284/2062\n",
      "Processing image: 285/2062\n",
      "Processing image: 286/2062\n",
      "Processing image: 287/2062\n",
      "Processing image: 288/2062\n",
      "Processing image: 289/2062\n",
      "Processing image: 290/2062\n",
      "Processing image: 291/2062\n",
      "Processing image: 292/2062\n",
      "Processing image: 293/2062\n",
      "Processing image: 294/2062\n",
      "Processing image: 295/2062\n",
      "Processing image: 296/2062\n",
      "Processing image: 297/2062\n",
      "Processing image: 298/2062\n",
      "Processing image: 299/2062\n",
      "Processing image: 300/2062\n",
      "Processing image: 301/2062\n",
      "Processing image: 302/2062\n",
      "Processing image: 303/2062\n",
      "Processing image: 304/2062\n",
      "Processing image: 305/2062\n",
      "Processing image: 306/2062\n",
      "Processing image: 307/2062\n",
      "Processing image: 308/2062\n",
      "Processing image: 309/2062\n",
      "Processing image: 310/2062\n",
      "Processing image: 311/2062\n",
      "Processing image: 312/2062\n",
      "Processing image: 313/2062\n",
      "Processing image: 314/2062\n",
      "Processing image: 315/2062\n",
      "Processing image: 316/2062\n",
      "Processing image: 317/2062\n",
      "Processing image: 318/2062\n",
      "Processing image: 319/2062\n",
      "Processing image: 320/2062\n",
      "Processing image: 321/2062\n",
      "Processing image: 322/2062\n",
      "Processing image: 323/2062\n",
      "Processing image: 324/2062\n",
      "Processing image: 325/2062\n",
      "Processing image: 326/2062\n",
      "Processing image: 327/2062\n",
      "Processing image: 328/2062\n",
      "Processing image: 329/2062\n",
      "Processing image: 330/2062\n",
      "Processing image: 331/2062\n",
      "Processing image: 332/2062\n",
      "Processing image: 333/2062\n",
      "Processing image: 334/2062\n",
      "Processing image: 335/2062\n",
      "Processing image: 336/2062\n",
      "Processing image: 337/2062\n",
      "Processing image: 338/2062\n",
      "Processing image: 339/2062\n",
      "Processing image: 340/2062\n",
      "Processing image: 341/2062\n",
      "Processing image: 342/2062\n",
      "Processing image: 343/2062\n",
      "Processing image: 344/2062\n",
      "Processing image: 345/2062\n",
      "Processing image: 346/2062\n",
      "Processing image: 347/2062\n",
      "Processing image: 348/2062\n",
      "Processing image: 349/2062\n",
      "Processing image: 350/2062\n",
      "Processing image: 351/2062\n",
      "Processing image: 352/2062\n",
      "Processing image: 353/2062\n",
      "Processing image: 354/2062\n",
      "Processing image: 355/2062\n",
      "Processing image: 356/2062\n",
      "Processing image: 357/2062\n",
      "Processing image: 358/2062\n",
      "Processing image: 359/2062\n",
      "Processing image: 360/2062\n",
      "Processing image: 361/2062\n",
      "Processing image: 362/2062\n",
      "Processing image: 363/2062\n",
      "Processing image: 364/2062\n",
      "Processing image: 365/2062\n",
      "Processing image: 366/2062\n",
      "Processing image: 367/2062\n",
      "Processing image: 368/2062\n",
      "Processing image: 369/2062\n",
      "Processing image: 370/2062\n",
      "Processing image: 371/2062\n",
      "Processing image: 372/2062\n",
      "Processing image: 373/2062\n",
      "Processing image: 374/2062\n",
      "Processing image: 375/2062\n",
      "Processing image: 376/2062\n",
      "Processing image: 377/2062\n",
      "Processing image: 378/2062\n",
      "Processing image: 379/2062\n",
      "Processing image: 380/2062\n",
      "Processing image: 381/2062\n",
      "Processing image: 382/2062\n",
      "Processing image: 383/2062\n",
      "Processing image: 384/2062\n",
      "Processing image: 385/2062\n",
      "Processing image: 386/2062\n",
      "Processing image: 387/2062\n",
      "Processing image: 388/2062\n",
      "Processing image: 389/2062\n",
      "Processing image: 390/2062\n",
      "Processing image: 391/2062\n",
      "Processing image: 392/2062\n",
      "Processing image: 393/2062\n",
      "Processing image: 394/2062\n",
      "Processing image: 395/2062\n",
      "Processing image: 396/2062\n",
      "Processing image: 397/2062\n",
      "Processing image: 398/2062\n",
      "Processing image: 399/2062\n",
      "Processing image: 400/2062\n",
      "Processing image: 401/2062\n",
      "Processing image: 402/2062\n",
      "Processing image: 403/2062\n",
      "Processing image: 404/2062\n",
      "Processing image: 405/2062\n",
      "Processing image: 406/2062\n",
      "Processing image: 407/2062\n",
      "Processing image: 408/2062\n",
      "Processing image: 409/2062\n",
      "Processing image: 410/2062\n",
      "Processing image: 411/2062\n",
      "Processing image: 412/2062\n",
      "Processing image: 413/2062\n",
      "Processing image: 414/2062\n",
      "Processing image: 415/2062\n",
      "Processing image: 416/2062\n",
      "Processing image: 417/2062\n",
      "Processing image: 418/2062\n",
      "Processing image: 419/2062\n",
      "Processing image: 420/2062\n",
      "Processing image: 421/2062\n",
      "Processing image: 422/2062\n",
      "Processing image: 423/2062\n",
      "Processing image: 424/2062\n",
      "Processing image: 425/2062\n",
      "Processing image: 426/2062\n",
      "Processing image: 427/2062\n",
      "Processing image: 428/2062\n",
      "Processing image: 429/2062\n",
      "Processing image: 430/2062\n",
      "Processing image: 431/2062\n",
      "Processing image: 432/2062\n",
      "Processing image: 433/2062\n",
      "Processing image: 434/2062\n",
      "Processing image: 435/2062\n",
      "Processing image: 436/2062\n",
      "Processing image: 437/2062\n",
      "Processing image: 438/2062\n",
      "Processing image: 439/2062\n",
      "Processing image: 440/2062\n",
      "Processing image: 441/2062\n",
      "Processing image: 442/2062\n",
      "Processing image: 443/2062\n",
      "Processing image: 444/2062\n",
      "Processing image: 445/2062\n",
      "Processing image: 446/2062\n",
      "Processing image: 447/2062\n",
      "Processing image: 448/2062\n",
      "Processing image: 449/2062\n",
      "Processing image: 450/2062\n",
      "Processing image: 451/2062\n",
      "Processing image: 452/2062\n",
      "Processing image: 453/2062\n",
      "Processing image: 454/2062\n",
      "Processing image: 455/2062\n",
      "Processing image: 456/2062\n",
      "Processing image: 457/2062\n",
      "Processing image: 458/2062\n",
      "Processing image: 459/2062\n",
      "Processing image: 460/2062\n",
      "Processing image: 461/2062\n",
      "Processing image: 462/2062\n",
      "Processing image: 463/2062\n",
      "Processing image: 464/2062\n",
      "Processing image: 465/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 466/2062\n",
      "Processing image: 467/2062\n",
      "Processing image: 468/2062\n",
      "Processing image: 469/2062\n",
      "Processing image: 470/2062\n",
      "Processing image: 471/2062\n",
      "Processing image: 472/2062\n",
      "Processing image: 473/2062\n",
      "Processing image: 474/2062\n",
      "Processing image: 475/2062\n",
      "Processing image: 476/2062\n",
      "Processing image: 477/2062\n",
      "Processing image: 478/2062\n",
      "Processing image: 479/2062\n",
      "Processing image: 480/2062\n",
      "Processing image: 481/2062\n",
      "Processing image: 482/2062\n",
      "Processing image: 483/2062\n",
      "Processing image: 484/2062\n",
      "Processing image: 485/2062\n",
      "Processing image: 486/2062\n",
      "Processing image: 487/2062\n",
      "Processing image: 488/2062\n",
      "Processing image: 489/2062\n",
      "Processing image: 490/2062\n",
      "Processing image: 491/2062\n",
      "Processing image: 492/2062\n",
      "Processing image: 493/2062\n",
      "Processing image: 494/2062\n",
      "Processing image: 495/2062\n",
      "Processing image: 496/2062\n",
      "Processing image: 497/2062\n",
      "Processing image: 498/2062\n",
      "Processing image: 499/2062\n",
      "Processing image: 500/2062\n",
      "Processing image: 501/2062\n",
      "Processing image: 502/2062\n",
      "Processing image: 503/2062\n",
      "Processing image: 504/2062\n",
      "Processing image: 505/2062\n",
      "Processing image: 506/2062\n",
      "Processing image: 507/2062\n",
      "Processing image: 508/2062\n",
      "Processing image: 509/2062\n",
      "Processing image: 510/2062\n",
      "Processing image: 511/2062\n",
      "Processing image: 512/2062\n",
      "Processing image: 513/2062\n",
      "Processing image: 514/2062\n",
      "Processing image: 515/2062\n",
      "Processing image: 516/2062\n",
      "Processing image: 517/2062\n",
      "Processing image: 518/2062\n",
      "Processing image: 519/2062\n",
      "Processing image: 520/2062\n",
      "Processing image: 521/2062\n",
      "Processing image: 522/2062\n",
      "Processing image: 523/2062\n",
      "Processing image: 524/2062\n",
      "Processing image: 525/2062\n",
      "Processing image: 526/2062\n",
      "Processing image: 527/2062\n",
      "Processing image: 528/2062\n",
      "Processing image: 529/2062\n",
      "Processing image: 530/2062\n",
      "Processing image: 531/2062\n",
      "Processing image: 532/2062\n",
      "Processing image: 533/2062\n",
      "Processing image: 534/2062\n",
      "Processing image: 535/2062\n",
      "Processing image: 536/2062\n",
      "Processing image: 537/2062\n",
      "Processing image: 538/2062\n",
      "Processing image: 539/2062\n",
      "Processing image: 540/2062\n",
      "Processing image: 541/2062\n",
      "Processing image: 542/2062\n",
      "Processing image: 543/2062\n",
      "Processing image: 544/2062\n",
      "Processing image: 545/2062\n",
      "Processing image: 546/2062\n",
      "Processing image: 547/2062\n",
      "Processing image: 548/2062\n",
      "Processing image: 549/2062\n",
      "Processing image: 550/2062\n",
      "Processing image: 551/2062\n",
      "Processing image: 552/2062\n",
      "Processing image: 553/2062\n",
      "Processing image: 554/2062\n",
      "Processing image: 555/2062\n",
      "Processing image: 556/2062\n",
      "Processing image: 557/2062\n",
      "Processing image: 558/2062\n",
      "Processing image: 559/2062\n",
      "Processing image: 560/2062\n",
      "Processing image: 561/2062\n",
      "Processing image: 562/2062\n",
      "Processing image: 563/2062\n",
      "Processing image: 564/2062\n",
      "Processing image: 565/2062\n",
      "Processing image: 566/2062\n",
      "Processing image: 567/2062\n",
      "Processing image: 568/2062\n",
      "Processing image: 569/2062\n",
      "Processing image: 570/2062\n",
      "Processing image: 571/2062\n",
      "Processing image: 572/2062\n",
      "Processing image: 573/2062\n",
      "Processing image: 574/2062\n",
      "Processing image: 575/2062\n",
      "Processing image: 576/2062\n",
      "Processing image: 577/2062\n",
      "Processing image: 578/2062\n",
      "Processing image: 579/2062\n",
      "Processing image: 580/2062\n",
      "Processing image: 581/2062\n",
      "Processing image: 582/2062\n",
      "Processing image: 583/2062\n",
      "Processing image: 584/2062\n",
      "Processing image: 585/2062\n",
      "Processing image: 586/2062\n",
      "Processing image: 587/2062\n",
      "Processing image: 588/2062\n",
      "Processing image: 589/2062\n",
      "Processing image: 590/2062\n",
      "Processing image: 591/2062\n",
      "Processing image: 592/2062\n",
      "Processing image: 593/2062\n",
      "Processing image: 594/2062\n",
      "Processing image: 595/2062\n",
      "Processing image: 596/2062\n",
      "Processing image: 597/2062\n",
      "Processing image: 598/2062\n",
      "Processing image: 599/2062\n",
      "Processing image: 600/2062\n",
      "Processing image: 601/2062\n",
      "Processing image: 602/2062\n",
      "Processing image: 603/2062\n",
      "Processing image: 604/2062\n",
      "Processing image: 605/2062\n",
      "Processing image: 606/2062\n",
      "Processing image: 607/2062\n",
      "Processing image: 608/2062\n",
      "Processing image: 609/2062\n",
      "Processing image: 610/2062\n",
      "Processing image: 611/2062\n",
      "Processing image: 612/2062\n",
      "Processing image: 613/2062\n",
      "Processing image: 614/2062\n",
      "Processing image: 615/2062\n",
      "Processing image: 616/2062\n",
      "Processing image: 617/2062\n",
      "Processing image: 618/2062\n",
      "Processing image: 619/2062\n",
      "Processing image: 620/2062\n",
      "Processing image: 621/2062\n",
      "Processing image: 622/2062\n",
      "Processing image: 623/2062\n",
      "Processing image: 624/2062\n",
      "Processing image: 625/2062\n",
      "Processing image: 626/2062\n",
      "Processing image: 627/2062\n",
      "Processing image: 628/2062\n",
      "Processing image: 629/2062\n",
      "Processing image: 630/2062\n",
      "Processing image: 631/2062\n",
      "Processing image: 632/2062\n",
      "Processing image: 633/2062\n",
      "Processing image: 634/2062\n",
      "Processing image: 635/2062\n",
      "Processing image: 636/2062\n",
      "Processing image: 637/2062\n",
      "Processing image: 638/2062\n",
      "Processing image: 639/2062\n",
      "Processing image: 640/2062\n",
      "Processing image: 641/2062\n",
      "Processing image: 642/2062\n",
      "Processing image: 643/2062\n",
      "Processing image: 644/2062\n",
      "Processing image: 645/2062\n",
      "Processing image: 646/2062\n",
      "Processing image: 647/2062\n",
      "Processing image: 648/2062\n",
      "Processing image: 649/2062\n",
      "Processing image: 650/2062\n",
      "Processing image: 651/2062\n",
      "Processing image: 652/2062\n",
      "Processing image: 653/2062\n",
      "Processing image: 654/2062\n",
      "Processing image: 655/2062\n",
      "Processing image: 656/2062\n",
      "Processing image: 657/2062\n",
      "Processing image: 658/2062\n",
      "Processing image: 659/2062\n",
      "Processing image: 660/2062\n",
      "Processing image: 661/2062\n",
      "Processing image: 662/2062\n",
      "Processing image: 663/2062\n",
      "Processing image: 664/2062\n",
      "Processing image: 665/2062\n",
      "Processing image: 666/2062\n",
      "Processing image: 667/2062\n",
      "Processing image: 668/2062\n",
      "Processing image: 669/2062\n",
      "Processing image: 670/2062\n",
      "Processing image: 671/2062\n",
      "Processing image: 672/2062\n",
      "Processing image: 673/2062\n",
      "Processing image: 674/2062\n",
      "Processing image: 675/2062\n",
      "Processing image: 676/2062\n",
      "Processing image: 677/2062\n",
      "Processing image: 678/2062\n",
      "Processing image: 679/2062\n",
      "Processing image: 680/2062\n",
      "Processing image: 681/2062\n",
      "Processing image: 682/2062\n",
      "Processing image: 683/2062\n",
      "Processing image: 684/2062\n",
      "Processing image: 685/2062\n",
      "Processing image: 686/2062\n",
      "Processing image: 687/2062\n",
      "Processing image: 688/2062\n",
      "Processing image: 689/2062\n",
      "Processing image: 690/2062\n",
      "Processing image: 691/2062\n",
      "Processing image: 692/2062\n",
      "Processing image: 693/2062\n",
      "Processing image: 694/2062\n",
      "Processing image: 695/2062\n",
      "Processing image: 696/2062\n",
      "Processing image: 697/2062\n",
      "Processing image: 698/2062\n",
      "Processing image: 699/2062\n",
      "Processing image: 700/2062\n",
      "Processing image: 701/2062\n",
      "Processing image: 702/2062\n",
      "Processing image: 703/2062\n",
      "Processing image: 704/2062\n",
      "Processing image: 705/2062\n",
      "Processing image: 706/2062\n",
      "Processing image: 707/2062\n",
      "Processing image: 708/2062\n",
      "Processing image: 709/2062\n",
      "Processing image: 710/2062\n",
      "Processing image: 711/2062\n",
      "Processing image: 712/2062\n",
      "Processing image: 713/2062\n",
      "Processing image: 714/2062\n",
      "Processing image: 715/2062\n",
      "Processing image: 716/2062\n",
      "Processing image: 717/2062\n",
      "Processing image: 718/2062\n",
      "Processing image: 719/2062\n",
      "Processing image: 720/2062\n",
      "Processing image: 721/2062\n",
      "Processing image: 722/2062\n",
      "Processing image: 723/2062\n",
      "Processing image: 724/2062\n",
      "Processing image: 725/2062\n",
      "Processing image: 726/2062\n",
      "Processing image: 727/2062\n",
      "Processing image: 728/2062\n",
      "Processing image: 729/2062\n",
      "Processing image: 730/2062\n",
      "Processing image: 731/2062\n",
      "Processing image: 732/2062\n",
      "Processing image: 733/2062\n",
      "Processing image: 734/2062\n",
      "Processing image: 735/2062\n",
      "Processing image: 736/2062\n",
      "Processing image: 737/2062\n",
      "Processing image: 738/2062\n",
      "Processing image: 739/2062\n",
      "Processing image: 740/2062\n",
      "Processing image: 741/2062\n",
      "Processing image: 742/2062\n",
      "Processing image: 743/2062\n",
      "Processing image: 744/2062\n",
      "Processing image: 745/2062\n",
      "Processing image: 746/2062\n",
      "Processing image: 747/2062\n",
      "Processing image: 748/2062\n",
      "Processing image: 749/2062\n",
      "Processing image: 750/2062\n",
      "Processing image: 751/2062\n",
      "Processing image: 752/2062\n",
      "Processing image: 753/2062\n",
      "Processing image: 754/2062\n",
      "Processing image: 755/2062\n",
      "Processing image: 756/2062\n",
      "Processing image: 757/2062\n",
      "Processing image: 758/2062\n",
      "Processing image: 759/2062\n",
      "Processing image: 760/2062\n",
      "Processing image: 761/2062\n",
      "Processing image: 762/2062\n",
      "Processing image: 763/2062\n",
      "Processing image: 764/2062\n",
      "Processing image: 765/2062\n",
      "Processing image: 766/2062\n",
      "Processing image: 767/2062\n",
      "Processing image: 768/2062\n",
      "Processing image: 769/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 770/2062\n",
      "Processing image: 771/2062\n",
      "Processing image: 772/2062\n",
      "Processing image: 773/2062\n",
      "Processing image: 774/2062\n",
      "Processing image: 775/2062\n",
      "Processing image: 776/2062\n",
      "Processing image: 777/2062\n",
      "Processing image: 778/2062\n",
      "Processing image: 779/2062\n",
      "Processing image: 780/2062\n",
      "Processing image: 781/2062\n",
      "Processing image: 782/2062\n",
      "Processing image: 783/2062\n",
      "Processing image: 784/2062\n",
      "Processing image: 785/2062\n",
      "Processing image: 786/2062\n",
      "Processing image: 787/2062\n",
      "Processing image: 788/2062\n",
      "Processing image: 789/2062\n",
      "Processing image: 790/2062\n",
      "Processing image: 791/2062\n",
      "Processing image: 792/2062\n",
      "Processing image: 793/2062\n",
      "Processing image: 794/2062\n",
      "Processing image: 795/2062\n",
      "Processing image: 796/2062\n",
      "Processing image: 797/2062\n",
      "Processing image: 798/2062\n",
      "Processing image: 799/2062\n",
      "Processing image: 800/2062\n",
      "Processing image: 801/2062\n",
      "Processing image: 802/2062\n",
      "Processing image: 803/2062\n",
      "Processing image: 804/2062\n",
      "Processing image: 805/2062\n",
      "Processing image: 806/2062\n",
      "Processing image: 807/2062\n",
      "Processing image: 808/2062\n",
      "Processing image: 809/2062\n",
      "Processing image: 810/2062\n",
      "Processing image: 811/2062\n",
      "Processing image: 812/2062\n",
      "Processing image: 813/2062\n",
      "Processing image: 814/2062\n",
      "Processing image: 815/2062\n",
      "Processing image: 816/2062\n",
      "Processing image: 817/2062\n",
      "Processing image: 818/2062\n",
      "Processing image: 819/2062\n",
      "Processing image: 820/2062\n",
      "Processing image: 821/2062\n",
      "Processing image: 822/2062\n",
      "Processing image: 823/2062\n",
      "Processing image: 824/2062\n",
      "Processing image: 825/2062\n",
      "Processing image: 826/2062\n",
      "Processing image: 827/2062\n",
      "Processing image: 828/2062\n",
      "Processing image: 829/2062\n",
      "Processing image: 830/2062\n",
      "Processing image: 831/2062\n",
      "Processing image: 832/2062\n",
      "Processing image: 833/2062\n",
      "Processing image: 834/2062\n",
      "Processing image: 835/2062\n",
      "Processing image: 836/2062\n",
      "Processing image: 837/2062\n",
      "Processing image: 838/2062\n",
      "Processing image: 839/2062\n",
      "Processing image: 840/2062\n",
      "Processing image: 841/2062\n",
      "Processing image: 842/2062\n",
      "Processing image: 843/2062\n",
      "Processing image: 844/2062\n",
      "Processing image: 845/2062\n",
      "Processing image: 846/2062\n",
      "Processing image: 847/2062\n",
      "Processing image: 848/2062\n",
      "Processing image: 849/2062\n",
      "Processing image: 850/2062\n",
      "Processing image: 851/2062\n",
      "Processing image: 852/2062\n",
      "Processing image: 853/2062\n",
      "Processing image: 854/2062\n",
      "Processing image: 855/2062\n",
      "Processing image: 856/2062\n",
      "Processing image: 857/2062\n",
      "Processing image: 858/2062\n",
      "Processing image: 859/2062\n",
      "Processing image: 860/2062\n",
      "Processing image: 861/2062\n",
      "Processing image: 862/2062\n",
      "Processing image: 863/2062\n",
      "Processing image: 864/2062\n",
      "Processing image: 865/2062\n",
      "Processing image: 866/2062\n",
      "Processing image: 867/2062\n",
      "Processing image: 868/2062\n",
      "Processing image: 869/2062\n",
      "Processing image: 870/2062\n",
      "Processing image: 871/2062\n",
      "Processing image: 872/2062\n",
      "Processing image: 873/2062\n",
      "Processing image: 874/2062\n",
      "Processing image: 875/2062\n",
      "Processing image: 876/2062\n",
      "Processing image: 877/2062\n",
      "Processing image: 878/2062\n",
      "Processing image: 879/2062\n",
      "Processing image: 880/2062\n",
      "Processing image: 881/2062\n",
      "Processing image: 882/2062\n",
      "Processing image: 883/2062\n",
      "Processing image: 884/2062\n",
      "Processing image: 885/2062\n",
      "Processing image: 886/2062\n",
      "Processing image: 887/2062\n",
      "Processing image: 888/2062\n",
      "Processing image: 889/2062\n",
      "Processing image: 890/2062\n",
      "Processing image: 891/2062\n",
      "Processing image: 892/2062\n",
      "Processing image: 893/2062\n",
      "Processing image: 894/2062\n",
      "Processing image: 895/2062\n",
      "Processing image: 896/2062\n",
      "Processing image: 897/2062\n",
      "Processing image: 898/2062\n",
      "Processing image: 899/2062\n",
      "Processing image: 900/2062\n",
      "Processing image: 901/2062\n",
      "Processing image: 902/2062\n",
      "Processing image: 903/2062\n",
      "Processing image: 904/2062\n",
      "Processing image: 905/2062\n",
      "Processing image: 906/2062\n",
      "Processing image: 907/2062\n",
      "Processing image: 908/2062\n",
      "Processing image: 909/2062\n",
      "Processing image: 910/2062\n",
      "Processing image: 911/2062\n",
      "Processing image: 912/2062\n",
      "Processing image: 913/2062\n",
      "Processing image: 914/2062\n",
      "Processing image: 915/2062\n",
      "Processing image: 916/2062\n",
      "Processing image: 917/2062\n",
      "Processing image: 918/2062\n",
      "Processing image: 919/2062\n",
      "Processing image: 920/2062\n",
      "Processing image: 921/2062\n",
      "Processing image: 922/2062\n",
      "Processing image: 923/2062\n",
      "Processing image: 924/2062\n",
      "Processing image: 925/2062\n",
      "Processing image: 926/2062\n",
      "Processing image: 927/2062\n",
      "Processing image: 928/2062\n",
      "Processing image: 929/2062\n",
      "Processing image: 930/2062\n",
      "Processing image: 931/2062\n",
      "Processing image: 932/2062\n",
      "Processing image: 933/2062\n",
      "Processing image: 934/2062\n",
      "Processing image: 935/2062\n",
      "Processing image: 936/2062\n",
      "Processing image: 937/2062\n",
      "Processing image: 938/2062\n",
      "Processing image: 939/2062\n",
      "Processing image: 940/2062\n",
      "Processing image: 941/2062\n",
      "Processing image: 942/2062\n",
      "Processing image: 943/2062\n",
      "Processing image: 944/2062\n",
      "Processing image: 945/2062\n",
      "Processing image: 946/2062\n",
      "Processing image: 947/2062\n",
      "Processing image: 948/2062\n",
      "Processing image: 949/2062\n",
      "Processing image: 950/2062\n",
      "Processing image: 951/2062\n",
      "Processing image: 952/2062\n",
      "Processing image: 953/2062\n",
      "Processing image: 954/2062\n",
      "Processing image: 955/2062\n",
      "Processing image: 956/2062\n",
      "Processing image: 957/2062\n",
      "Processing image: 958/2062\n",
      "Processing image: 959/2062\n",
      "Processing image: 960/2062\n",
      "Processing image: 961/2062\n",
      "Processing image: 962/2062\n",
      "Processing image: 963/2062\n",
      "Processing image: 964/2062\n",
      "Processing image: 965/2062\n",
      "Processing image: 966/2062\n",
      "Processing image: 967/2062\n",
      "Processing image: 968/2062\n",
      "Processing image: 969/2062\n",
      "Processing image: 970/2062\n",
      "Processing image: 971/2062\n",
      "Processing image: 972/2062\n",
      "Processing image: 973/2062\n",
      "Processing image: 974/2062\n",
      "Processing image: 975/2062\n",
      "Processing image: 976/2062\n",
      "Processing image: 977/2062\n",
      "Processing image: 978/2062\n",
      "Processing image: 979/2062\n",
      "Processing image: 980/2062\n",
      "Processing image: 981/2062\n",
      "Processing image: 982/2062\n",
      "Processing image: 983/2062\n",
      "Processing image: 984/2062\n",
      "Processing image: 985/2062\n",
      "Processing image: 986/2062\n",
      "Processing image: 987/2062\n",
      "Processing image: 988/2062\n",
      "Processing image: 989/2062\n",
      "Processing image: 990/2062\n",
      "Processing image: 991/2062\n",
      "Processing image: 992/2062\n",
      "Processing image: 993/2062\n",
      "Processing image: 994/2062\n",
      "Processing image: 995/2062\n",
      "Processing image: 996/2062\n",
      "Processing image: 997/2062\n",
      "Processing image: 998/2062\n",
      "Processing image: 999/2062\n",
      "Processing image: 1000/2062\n",
      "Processing image: 1001/2062\n",
      "Processing image: 1002/2062\n",
      "Processing image: 1003/2062\n",
      "Processing image: 1004/2062\n",
      "Processing image: 1005/2062\n",
      "Processing image: 1006/2062\n",
      "Processing image: 1007/2062\n",
      "Processing image: 1008/2062\n",
      "Processing image: 1009/2062\n",
      "Processing image: 1010/2062\n",
      "Processing image: 1011/2062\n",
      "Processing image: 1012/2062\n",
      "Processing image: 1013/2062\n",
      "Processing image: 1014/2062\n",
      "Processing image: 1015/2062\n",
      "Processing image: 1016/2062\n",
      "Processing image: 1017/2062\n",
      "Processing image: 1018/2062\n",
      "Processing image: 1019/2062\n",
      "Processing image: 1020/2062\n",
      "Processing image: 1021/2062\n",
      "Processing image: 1022/2062\n",
      "Processing image: 1023/2062\n",
      "Processing image: 1024/2062\n",
      "Processing image: 1025/2062\n",
      "Processing image: 1026/2062\n",
      "Processing image: 1027/2062\n",
      "Processing image: 1028/2062\n",
      "Processing image: 1029/2062\n",
      "Processing image: 1030/2062\n",
      "Processing image: 1031/2062\n",
      "Processing image: 1032/2062\n",
      "Processing image: 1033/2062\n",
      "Processing image: 1034/2062\n",
      "Processing image: 1035/2062\n",
      "Processing image: 1036/2062\n",
      "Processing image: 1037/2062\n",
      "Processing image: 1038/2062\n",
      "Processing image: 1039/2062\n",
      "Processing image: 1040/2062\n",
      "Processing image: 1041/2062\n",
      "Processing image: 1042/2062\n",
      "Processing image: 1043/2062\n",
      "Processing image: 1044/2062\n",
      "Processing image: 1045/2062\n",
      "Processing image: 1046/2062\n",
      "Processing image: 1047/2062\n",
      "Processing image: 1048/2062\n",
      "Processing image: 1049/2062\n",
      "Processing image: 1050/2062\n",
      "Processing image: 1051/2062\n",
      "Processing image: 1052/2062\n",
      "Processing image: 1053/2062\n",
      "Processing image: 1054/2062\n",
      "Processing image: 1055/2062\n",
      "Processing image: 1056/2062\n",
      "Processing image: 1057/2062\n",
      "Processing image: 1058/2062\n",
      "Processing image: 1059/2062\n",
      "Processing image: 1060/2062\n",
      "Processing image: 1061/2062\n",
      "Processing image: 1062/2062\n",
      "Processing image: 1063/2062\n",
      "Processing image: 1064/2062\n",
      "Processing image: 1065/2062\n",
      "Processing image: 1066/2062\n",
      "Processing image: 1067/2062\n",
      "Processing image: 1068/2062\n",
      "Processing image: 1069/2062\n",
      "Processing image: 1070/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1071/2062\n",
      "Processing image: 1072/2062\n",
      "Processing image: 1073/2062\n",
      "Processing image: 1074/2062\n",
      "Processing image: 1075/2062\n",
      "Processing image: 1076/2062\n",
      "Processing image: 1077/2062\n",
      "Processing image: 1078/2062\n",
      "Processing image: 1079/2062\n",
      "Processing image: 1080/2062\n",
      "Processing image: 1081/2062\n",
      "Processing image: 1082/2062\n",
      "Processing image: 1083/2062\n",
      "Processing image: 1084/2062\n",
      "Processing image: 1085/2062\n",
      "Processing image: 1086/2062\n",
      "Processing image: 1087/2062\n",
      "Processing image: 1088/2062\n",
      "Processing image: 1089/2062\n",
      "Processing image: 1090/2062\n",
      "Processing image: 1091/2062\n",
      "Processing image: 1092/2062\n",
      "Processing image: 1093/2062\n",
      "Processing image: 1094/2062\n",
      "Processing image: 1095/2062\n",
      "Processing image: 1096/2062\n",
      "Processing image: 1097/2062\n",
      "Processing image: 1098/2062\n",
      "Processing image: 1099/2062\n",
      "Processing image: 1100/2062\n",
      "Processing image: 1101/2062\n",
      "Processing image: 1102/2062\n",
      "Processing image: 1103/2062\n",
      "Processing image: 1104/2062\n",
      "Processing image: 1105/2062\n",
      "Processing image: 1106/2062\n",
      "Processing image: 1107/2062\n",
      "Processing image: 1108/2062\n",
      "Processing image: 1109/2062\n",
      "Processing image: 1110/2062\n",
      "Processing image: 1111/2062\n",
      "Processing image: 1112/2062\n",
      "Processing image: 1113/2062\n",
      "Processing image: 1114/2062\n",
      "Processing image: 1115/2062\n",
      "Processing image: 1116/2062\n",
      "Processing image: 1117/2062\n",
      "Processing image: 1118/2062\n",
      "Processing image: 1119/2062\n",
      "Processing image: 1120/2062\n",
      "Processing image: 1121/2062\n",
      "Processing image: 1122/2062\n",
      "Processing image: 1123/2062\n",
      "Processing image: 1124/2062\n",
      "Processing image: 1125/2062\n",
      "Processing image: 1126/2062\n",
      "Processing image: 1127/2062\n",
      "Processing image: 1128/2062\n",
      "Processing image: 1129/2062\n",
      "Processing image: 1130/2062\n",
      "Processing image: 1131/2062\n",
      "Processing image: 1132/2062\n",
      "Processing image: 1133/2062\n",
      "Processing image: 1134/2062\n",
      "Processing image: 1135/2062\n",
      "Processing image: 1136/2062\n",
      "Processing image: 1137/2062\n",
      "Processing image: 1138/2062\n",
      "Processing image: 1139/2062\n",
      "Processing image: 1140/2062\n",
      "Processing image: 1141/2062\n",
      "Processing image: 1142/2062\n",
      "Processing image: 1143/2062\n",
      "Processing image: 1144/2062\n",
      "Processing image: 1145/2062\n",
      "Processing image: 1146/2062\n",
      "Processing image: 1147/2062\n",
      "Processing image: 1148/2062\n",
      "Processing image: 1149/2062\n",
      "Processing image: 1150/2062\n",
      "Processing image: 1151/2062\n",
      "Processing image: 1152/2062\n",
      "Processing image: 1153/2062\n",
      "Processing image: 1154/2062\n",
      "Processing image: 1155/2062\n",
      "Processing image: 1156/2062\n",
      "Processing image: 1157/2062\n",
      "Processing image: 1158/2062\n",
      "Processing image: 1159/2062\n",
      "Processing image: 1160/2062\n",
      "Processing image: 1161/2062\n",
      "Processing image: 1162/2062\n",
      "Processing image: 1163/2062\n",
      "Processing image: 1164/2062\n",
      "Processing image: 1165/2062\n",
      "Processing image: 1166/2062\n",
      "Processing image: 1167/2062\n",
      "Processing image: 1168/2062\n",
      "Processing image: 1169/2062\n",
      "Processing image: 1170/2062\n",
      "Processing image: 1171/2062\n",
      "Processing image: 1172/2062\n",
      "Processing image: 1173/2062\n",
      "Processing image: 1174/2062\n",
      "Processing image: 1175/2062\n",
      "Processing image: 1176/2062\n",
      "Processing image: 1177/2062\n",
      "Processing image: 1178/2062\n",
      "Processing image: 1179/2062\n",
      "Processing image: 1180/2062\n",
      "Processing image: 1181/2062\n",
      "Processing image: 1182/2062\n",
      "Processing image: 1183/2062\n",
      "Processing image: 1184/2062\n",
      "Processing image: 1185/2062\n",
      "Processing image: 1186/2062\n",
      "Processing image: 1187/2062\n",
      "Processing image: 1188/2062\n",
      "Processing image: 1189/2062\n",
      "Processing image: 1190/2062\n",
      "Processing image: 1191/2062\n",
      "Processing image: 1192/2062\n",
      "Processing image: 1193/2062\n",
      "Processing image: 1194/2062\n",
      "Processing image: 1195/2062\n",
      "Processing image: 1196/2062\n",
      "Processing image: 1197/2062\n",
      "Processing image: 1198/2062\n",
      "Processing image: 1199/2062\n",
      "Processing image: 1200/2062\n",
      "Processing image: 1201/2062\n",
      "Processing image: 1202/2062\n",
      "Processing image: 1203/2062\n",
      "Processing image: 1204/2062\n",
      "Processing image: 1205/2062\n",
      "Processing image: 1206/2062\n",
      "Processing image: 1207/2062\n",
      "Processing image: 1208/2062\n",
      "Processing image: 1209/2062\n",
      "Processing image: 1210/2062\n",
      "Processing image: 1211/2062\n",
      "Processing image: 1212/2062\n",
      "Processing image: 1213/2062\n",
      "Processing image: 1214/2062\n",
      "Processing image: 1215/2062\n",
      "Processing image: 1216/2062\n",
      "Processing image: 1217/2062\n",
      "Processing image: 1218/2062\n",
      "Processing image: 1219/2062\n",
      "Processing image: 1220/2062\n",
      "Processing image: 1221/2062\n",
      "Processing image: 1222/2062\n",
      "Processing image: 1223/2062\n",
      "Processing image: 1224/2062\n",
      "Processing image: 1225/2062\n",
      "Processing image: 1226/2062\n",
      "Processing image: 1227/2062\n",
      "Processing image: 1228/2062\n",
      "Processing image: 1229/2062\n",
      "Processing image: 1230/2062\n",
      "Processing image: 1231/2062\n",
      "Processing image: 1232/2062\n",
      "Processing image: 1233/2062\n",
      "Processing image: 1234/2062\n",
      "Processing image: 1235/2062\n",
      "Processing image: 1236/2062\n",
      "Processing image: 1237/2062\n",
      "Processing image: 1238/2062\n",
      "Processing image: 1239/2062\n",
      "Processing image: 1240/2062\n",
      "Processing image: 1241/2062\n",
      "Processing image: 1242/2062\n",
      "Processing image: 1243/2062\n",
      "Processing image: 1244/2062\n",
      "Processing image: 1245/2062\n",
      "Processing image: 1246/2062\n",
      "Processing image: 1247/2062\n",
      "Processing image: 1248/2062\n",
      "Processing image: 1249/2062\n",
      "Processing image: 1250/2062\n",
      "Processing image: 1251/2062\n",
      "Processing image: 1252/2062\n",
      "Processing image: 1253/2062\n",
      "Processing image: 1254/2062\n",
      "Processing image: 1255/2062\n",
      "Processing image: 1256/2062\n",
      "Processing image: 1257/2062\n",
      "Processing image: 1258/2062\n",
      "Processing image: 1259/2062\n",
      "Processing image: 1260/2062\n",
      "Processing image: 1261/2062\n",
      "Processing image: 1262/2062\n",
      "Processing image: 1263/2062\n",
      "Processing image: 1264/2062\n",
      "Processing image: 1265/2062\n",
      "Processing image: 1266/2062\n",
      "Processing image: 1267/2062\n",
      "Processing image: 1268/2062\n",
      "Processing image: 1269/2062\n",
      "Processing image: 1270/2062\n",
      "Processing image: 1271/2062\n",
      "Processing image: 1272/2062\n",
      "Processing image: 1273/2062\n",
      "Processing image: 1274/2062\n",
      "Processing image: 1275/2062\n",
      "Processing image: 1276/2062\n",
      "Processing image: 1277/2062\n",
      "Processing image: 1278/2062\n",
      "Processing image: 1279/2062\n",
      "Processing image: 1280/2062\n",
      "Processing image: 1281/2062\n",
      "Processing image: 1282/2062\n",
      "Processing image: 1283/2062\n",
      "Processing image: 1284/2062\n",
      "Processing image: 1285/2062\n",
      "Processing image: 1286/2062\n",
      "Processing image: 1287/2062\n",
      "Processing image: 1288/2062\n",
      "Processing image: 1289/2062\n",
      "Processing image: 1290/2062\n",
      "Processing image: 1291/2062\n",
      "Processing image: 1292/2062\n",
      "Processing image: 1293/2062\n",
      "Processing image: 1294/2062\n",
      "Processing image: 1295/2062\n",
      "Processing image: 1296/2062\n",
      "Processing image: 1297/2062\n",
      "Processing image: 1298/2062\n",
      "Processing image: 1299/2062\n",
      "Processing image: 1300/2062\n",
      "Processing image: 1301/2062\n",
      "Processing image: 1302/2062\n",
      "Processing image: 1303/2062\n",
      "Processing image: 1304/2062\n",
      "Processing image: 1305/2062\n",
      "Processing image: 1306/2062\n",
      "Processing image: 1307/2062\n",
      "Processing image: 1308/2062\n",
      "Processing image: 1309/2062\n",
      "Processing image: 1310/2062\n",
      "Processing image: 1311/2062\n",
      "Processing image: 1312/2062\n",
      "Processing image: 1313/2062\n",
      "Processing image: 1314/2062\n",
      "Processing image: 1315/2062\n",
      "Processing image: 1316/2062\n",
      "Processing image: 1317/2062\n",
      "Processing image: 1318/2062\n",
      "Processing image: 1319/2062\n",
      "Processing image: 1320/2062\n",
      "Processing image: 1321/2062\n",
      "Processing image: 1322/2062\n",
      "Processing image: 1323/2062\n",
      "Processing image: 1324/2062\n",
      "Processing image: 1325/2062\n",
      "Processing image: 1326/2062\n",
      "Processing image: 1327/2062\n",
      "Processing image: 1328/2062\n",
      "Processing image: 1329/2062\n",
      "Processing image: 1330/2062\n",
      "Processing image: 1331/2062\n",
      "Processing image: 1332/2062\n",
      "Processing image: 1333/2062\n",
      "Processing image: 1334/2062\n",
      "Processing image: 1335/2062\n",
      "Processing image: 1336/2062\n",
      "Processing image: 1337/2062\n",
      "Processing image: 1338/2062\n",
      "Processing image: 1339/2062\n",
      "Processing image: 1340/2062\n",
      "Processing image: 1341/2062\n",
      "Processing image: 1342/2062\n",
      "Processing image: 1343/2062\n",
      "Processing image: 1344/2062\n",
      "Processing image: 1345/2062\n",
      "Processing image: 1346/2062\n",
      "Processing image: 1347/2062\n",
      "Processing image: 1348/2062\n",
      "Processing image: 1349/2062\n",
      "Processing image: 1350/2062\n",
      "Processing image: 1351/2062\n",
      "Processing image: 1352/2062\n",
      "Processing image: 1353/2062\n",
      "Processing image: 1354/2062\n",
      "Processing image: 1355/2062\n",
      "Processing image: 1356/2062\n",
      "Processing image: 1357/2062\n",
      "Processing image: 1358/2062\n",
      "Processing image: 1359/2062\n",
      "Processing image: 1360/2062\n",
      "Processing image: 1361/2062\n",
      "Processing image: 1362/2062\n",
      "Processing image: 1363/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1364/2062\n",
      "Processing image: 1365/2062\n",
      "Processing image: 1366/2062\n",
      "Processing image: 1367/2062\n",
      "Processing image: 1368/2062\n",
      "Processing image: 1369/2062\n",
      "Processing image: 1370/2062\n",
      "Processing image: 1371/2062\n",
      "Processing image: 1372/2062\n",
      "Processing image: 1373/2062\n",
      "Processing image: 1374/2062\n",
      "Processing image: 1375/2062\n",
      "Processing image: 1376/2062\n",
      "Processing image: 1377/2062\n",
      "Processing image: 1378/2062\n",
      "Processing image: 1379/2062\n",
      "Processing image: 1380/2062\n",
      "Processing image: 1381/2062\n",
      "Processing image: 1382/2062\n",
      "Processing image: 1383/2062\n",
      "Processing image: 1384/2062\n",
      "Processing image: 1385/2062\n",
      "Processing image: 1386/2062\n",
      "Processing image: 1387/2062\n",
      "Processing image: 1388/2062\n",
      "Processing image: 1389/2062\n",
      "Processing image: 1390/2062\n",
      "Processing image: 1391/2062\n",
      "Processing image: 1392/2062\n",
      "Processing image: 1393/2062\n",
      "Processing image: 1394/2062\n",
      "Processing image: 1395/2062\n",
      "Processing image: 1396/2062\n",
      "Processing image: 1397/2062\n",
      "Processing image: 1398/2062\n",
      "Processing image: 1399/2062\n",
      "Processing image: 1400/2062\n",
      "Processing image: 1401/2062\n",
      "Processing image: 1402/2062\n",
      "Processing image: 1403/2062\n",
      "Processing image: 1404/2062\n",
      "Processing image: 1405/2062\n",
      "Processing image: 1406/2062\n",
      "Processing image: 1407/2062\n",
      "Processing image: 1408/2062\n",
      "Processing image: 1409/2062\n",
      "Processing image: 1410/2062\n",
      "Processing image: 1411/2062\n",
      "Processing image: 1412/2062\n",
      "Processing image: 1413/2062\n",
      "Processing image: 1414/2062\n",
      "Processing image: 1415/2062\n",
      "Processing image: 1416/2062\n",
      "Processing image: 1417/2062\n",
      "Processing image: 1418/2062\n",
      "Processing image: 1419/2062\n",
      "Processing image: 1420/2062\n",
      "Processing image: 1421/2062\n",
      "Processing image: 1422/2062\n",
      "Processing image: 1423/2062\n",
      "Processing image: 1424/2062\n",
      "Processing image: 1425/2062\n",
      "Processing image: 1426/2062\n",
      "Processing image: 1427/2062\n",
      "Processing image: 1428/2062\n",
      "Processing image: 1429/2062\n",
      "Processing image: 1430/2062\n",
      "Processing image: 1431/2062\n",
      "Processing image: 1432/2062\n",
      "Processing image: 1433/2062\n",
      "Processing image: 1434/2062\n",
      "Processing image: 1435/2062\n",
      "Processing image: 1436/2062\n",
      "Processing image: 1437/2062\n",
      "Processing image: 1438/2062\n",
      "Processing image: 1439/2062\n",
      "Processing image: 1440/2062\n",
      "Processing image: 1441/2062\n",
      "Processing image: 1442/2062\n",
      "Processing image: 1443/2062\n",
      "Processing image: 1444/2062\n",
      "Processing image: 1445/2062\n",
      "Processing image: 1446/2062\n",
      "Processing image: 1447/2062\n",
      "Processing image: 1448/2062\n",
      "Processing image: 1449/2062\n",
      "Processing image: 1450/2062\n",
      "Processing image: 1451/2062\n",
      "Processing image: 1452/2062\n",
      "Processing image: 1453/2062\n",
      "Processing image: 1454/2062\n",
      "Processing image: 1455/2062\n",
      "Processing image: 1456/2062\n",
      "Processing image: 1457/2062\n",
      "Processing image: 1458/2062\n",
      "Processing image: 1459/2062\n",
      "Processing image: 1460/2062\n",
      "Processing image: 1461/2062\n",
      "Processing image: 1462/2062\n",
      "Processing image: 1463/2062\n",
      "Processing image: 1464/2062\n",
      "Processing image: 1465/2062\n",
      "Processing image: 1466/2062\n",
      "Processing image: 1467/2062\n",
      "Processing image: 1468/2062\n",
      "Processing image: 1469/2062\n",
      "Processing image: 1470/2062\n",
      "Processing image: 1471/2062\n",
      "Processing image: 1472/2062\n",
      "Processing image: 1473/2062\n",
      "Processing image: 1474/2062\n",
      "Processing image: 1475/2062\n",
      "Processing image: 1476/2062\n",
      "Processing image: 1477/2062\n",
      "Processing image: 1478/2062\n",
      "Processing image: 1479/2062\n",
      "Processing image: 1480/2062\n",
      "Processing image: 1481/2062\n",
      "Processing image: 1482/2062\n",
      "Processing image: 1483/2062\n",
      "Processing image: 1484/2062\n",
      "Processing image: 1485/2062\n",
      "Processing image: 1486/2062\n",
      "Processing image: 1487/2062\n",
      "Processing image: 1488/2062\n",
      "Processing image: 1489/2062\n",
      "Processing image: 1490/2062\n",
      "Processing image: 1491/2062\n",
      "Processing image: 1492/2062\n",
      "Processing image: 1493/2062\n",
      "Processing image: 1494/2062\n",
      "Processing image: 1495/2062\n",
      "Processing image: 1496/2062\n",
      "Processing image: 1497/2062\n",
      "Processing image: 1498/2062\n",
      "Processing image: 1499/2062\n",
      "Processing image: 1500/2062\n",
      "Processing image: 1501/2062\n",
      "Processing image: 1502/2062\n",
      "Processing image: 1503/2062\n",
      "Processing image: 1504/2062\n",
      "Processing image: 1505/2062\n",
      "Processing image: 1506/2062\n",
      "Processing image: 1507/2062\n",
      "Processing image: 1508/2062\n",
      "Processing image: 1509/2062\n",
      "Processing image: 1510/2062\n",
      "Processing image: 1511/2062\n",
      "Processing image: 1512/2062\n",
      "Processing image: 1513/2062\n",
      "Processing image: 1514/2062\n",
      "Processing image: 1515/2062\n",
      "Processing image: 1516/2062\n",
      "Processing image: 1517/2062\n",
      "Processing image: 1518/2062\n",
      "Processing image: 1519/2062\n",
      "Processing image: 1520/2062\n",
      "Processing image: 1521/2062\n",
      "Processing image: 1522/2062\n",
      "Processing image: 1523/2062\n",
      "Processing image: 1524/2062\n",
      "Processing image: 1525/2062\n",
      "Processing image: 1526/2062\n",
      "Processing image: 1527/2062\n",
      "Processing image: 1528/2062\n",
      "Processing image: 1529/2062\n",
      "Processing image: 1530/2062\n",
      "Processing image: 1531/2062\n",
      "Processing image: 1532/2062\n",
      "Processing image: 1533/2062\n",
      "Processing image: 1534/2062\n",
      "Processing image: 1535/2062\n",
      "Processing image: 1536/2062\n",
      "Processing image: 1537/2062\n",
      "Processing image: 1538/2062\n",
      "Processing image: 1539/2062\n",
      "Processing image: 1540/2062\n",
      "Processing image: 1541/2062\n",
      "Processing image: 1542/2062\n",
      "Processing image: 1543/2062\n",
      "Processing image: 1544/2062\n",
      "Processing image: 1545/2062\n",
      "Processing image: 1546/2062\n",
      "Processing image: 1547/2062\n",
      "Processing image: 1548/2062\n",
      "Processing image: 1549/2062\n",
      "Processing image: 1550/2062\n",
      "Processing image: 1551/2062\n",
      "Processing image: 1552/2062\n",
      "Processing image: 1553/2062\n",
      "Processing image: 1554/2062\n",
      "Processing image: 1555/2062\n",
      "Processing image: 1556/2062\n",
      "Processing image: 1557/2062\n",
      "Processing image: 1558/2062\n",
      "Processing image: 1559/2062\n",
      "Processing image: 1560/2062\n",
      "Processing image: 1561/2062\n",
      "Processing image: 1562/2062\n",
      "Processing image: 1563/2062\n",
      "Processing image: 1564/2062\n",
      "Processing image: 1565/2062\n",
      "Processing image: 1566/2062\n",
      "Processing image: 1567/2062\n",
      "Processing image: 1568/2062\n",
      "Processing image: 1569/2062\n",
      "Processing image: 1570/2062\n",
      "Processing image: 1571/2062\n",
      "Processing image: 1572/2062\n",
      "Processing image: 1573/2062\n",
      "Processing image: 1574/2062\n",
      "Processing image: 1575/2062\n",
      "Processing image: 1576/2062\n",
      "Processing image: 1577/2062\n",
      "Processing image: 1578/2062\n",
      "Processing image: 1579/2062\n",
      "Processing image: 1580/2062\n",
      "Processing image: 1581/2062\n",
      "Processing image: 1582/2062\n",
      "Processing image: 1583/2062\n",
      "Processing image: 1584/2062\n",
      "Processing image: 1585/2062\n",
      "Processing image: 1586/2062\n",
      "Processing image: 1587/2062\n",
      "Processing image: 1588/2062\n",
      "Processing image: 1589/2062\n",
      "Processing image: 1590/2062\n",
      "Processing image: 1591/2062\n",
      "Processing image: 1592/2062\n",
      "Processing image: 1593/2062\n",
      "Processing image: 1594/2062\n",
      "Processing image: 1595/2062\n",
      "Processing image: 1596/2062\n",
      "Processing image: 1597/2062\n",
      "Processing image: 1598/2062\n",
      "Processing image: 1599/2062\n",
      "Processing image: 1600/2062\n",
      "Processing image: 1601/2062\n",
      "Processing image: 1602/2062\n",
      "Processing image: 1603/2062\n",
      "Processing image: 1604/2062\n",
      "Processing image: 1605/2062\n",
      "Processing image: 1606/2062\n",
      "Processing image: 1607/2062\n",
      "Processing image: 1608/2062\n",
      "Processing image: 1609/2062\n",
      "Processing image: 1610/2062\n",
      "Processing image: 1611/2062\n",
      "Processing image: 1612/2062\n",
      "Processing image: 1613/2062\n",
      "Processing image: 1614/2062\n",
      "Processing image: 1615/2062\n",
      "Processing image: 1616/2062\n",
      "Processing image: 1617/2062\n",
      "Processing image: 1618/2062\n",
      "Processing image: 1619/2062\n",
      "Processing image: 1620/2062\n",
      "Processing image: 1621/2062\n",
      "Processing image: 1622/2062\n",
      "Processing image: 1623/2062\n",
      "Processing image: 1624/2062\n",
      "Processing image: 1625/2062\n",
      "Processing image: 1626/2062\n",
      "Processing image: 1627/2062\n",
      "Processing image: 1628/2062\n",
      "Processing image: 1629/2062\n",
      "Processing image: 1630/2062\n",
      "Processing image: 1631/2062\n",
      "Processing image: 1632/2062\n",
      "Processing image: 1633/2062\n",
      "Processing image: 1634/2062\n",
      "Processing image: 1635/2062\n",
      "Processing image: 1636/2062\n",
      "Processing image: 1637/2062\n",
      "Processing image: 1638/2062\n",
      "Processing image: 1639/2062\n",
      "Processing image: 1640/2062\n",
      "Processing image: 1641/2062\n",
      "Processing image: 1642/2062\n",
      "Processing image: 1643/2062\n",
      "Processing image: 1644/2062\n",
      "Processing image: 1645/2062\n",
      "Processing image: 1646/2062\n",
      "Processing image: 1647/2062\n",
      "Processing image: 1648/2062\n",
      "Processing image: 1649/2062\n",
      "Processing image: 1650/2062\n",
      "Processing image: 1651/2062\n",
      "Processing image: 1652/2062\n",
      "Processing image: 1653/2062\n",
      "Processing image: 1654/2062\n",
      "Processing image: 1655/2062\n",
      "Processing image: 1656/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1657/2062\n",
      "Processing image: 1658/2062\n",
      "Processing image: 1659/2062\n",
      "Processing image: 1660/2062\n",
      "Processing image: 1661/2062\n",
      "Processing image: 1662/2062\n",
      "Processing image: 1663/2062\n",
      "Processing image: 1664/2062\n",
      "Processing image: 1665/2062\n",
      "Processing image: 1666/2062\n",
      "Processing image: 1667/2062\n",
      "Processing image: 1668/2062\n",
      "Processing image: 1669/2062\n",
      "Processing image: 1670/2062\n",
      "Processing image: 1671/2062\n",
      "Processing image: 1672/2062\n",
      "Processing image: 1673/2062\n",
      "Processing image: 1674/2062\n",
      "Processing image: 1675/2062\n",
      "Processing image: 1676/2062\n",
      "Processing image: 1677/2062\n",
      "Processing image: 1678/2062\n",
      "Processing image: 1679/2062\n",
      "Processing image: 1680/2062\n",
      "Processing image: 1681/2062\n",
      "Processing image: 1682/2062\n",
      "Processing image: 1683/2062\n",
      "Processing image: 1684/2062\n",
      "Processing image: 1685/2062\n",
      "Processing image: 1686/2062\n",
      "Processing image: 1687/2062\n",
      "Processing image: 1688/2062\n",
      "Processing image: 1689/2062\n",
      "Processing image: 1690/2062\n",
      "Processing image: 1691/2062\n",
      "Processing image: 1692/2062\n",
      "Processing image: 1693/2062\n",
      "Processing image: 1694/2062\n",
      "Processing image: 1695/2062\n",
      "Processing image: 1696/2062\n",
      "Processing image: 1697/2062\n",
      "Processing image: 1698/2062\n",
      "Processing image: 1699/2062\n",
      "Processing image: 1700/2062\n",
      "Processing image: 1701/2062\n",
      "Processing image: 1702/2062\n",
      "Processing image: 1703/2062\n",
      "Processing image: 1704/2062\n",
      "Processing image: 1705/2062\n",
      "Processing image: 1706/2062\n",
      "Processing image: 1707/2062\n",
      "Processing image: 1708/2062\n",
      "Processing image: 1709/2062\n",
      "Processing image: 1710/2062\n",
      "Processing image: 1711/2062\n",
      "Processing image: 1712/2062\n",
      "Processing image: 1713/2062\n",
      "Processing image: 1714/2062\n",
      "Processing image: 1715/2062\n",
      "Processing image: 1716/2062\n",
      "Processing image: 1717/2062\n",
      "Processing image: 1718/2062\n",
      "Processing image: 1719/2062\n",
      "Processing image: 1720/2062\n",
      "Processing image: 1721/2062\n",
      "Processing image: 1722/2062\n",
      "Processing image: 1723/2062\n",
      "Processing image: 1724/2062\n",
      "Processing image: 1725/2062\n",
      "Processing image: 1726/2062\n",
      "Processing image: 1727/2062\n",
      "Processing image: 1728/2062\n",
      "Processing image: 1729/2062\n",
      "Processing image: 1730/2062\n",
      "Processing image: 1731/2062\n",
      "Processing image: 1732/2062\n",
      "Processing image: 1733/2062\n",
      "Processing image: 1734/2062\n",
      "Processing image: 1735/2062\n",
      "Processing image: 1736/2062\n",
      "Processing image: 1737/2062\n",
      "Processing image: 1738/2062\n",
      "Processing image: 1739/2062\n",
      "Processing image: 1740/2062\n",
      "Processing image: 1741/2062\n",
      "Processing image: 1742/2062\n",
      "Processing image: 1743/2062\n",
      "Processing image: 1744/2062\n",
      "Processing image: 1745/2062\n",
      "Processing image: 1746/2062\n",
      "Processing image: 1747/2062\n",
      "Processing image: 1748/2062\n",
      "Processing image: 1749/2062\n",
      "Processing image: 1750/2062\n",
      "Processing image: 1751/2062\n",
      "Processing image: 1752/2062\n",
      "Processing image: 1753/2062\n",
      "Processing image: 1754/2062\n",
      "Processing image: 1755/2062\n",
      "Processing image: 1756/2062\n",
      "Processing image: 1757/2062\n",
      "Processing image: 1758/2062\n",
      "Processing image: 1759/2062\n",
      "Processing image: 1760/2062\n",
      "Processing image: 1761/2062\n",
      "Processing image: 1762/2062\n",
      "Processing image: 1763/2062\n",
      "Processing image: 1764/2062\n",
      "Processing image: 1765/2062\n",
      "Processing image: 1766/2062\n",
      "Processing image: 1767/2062\n",
      "Processing image: 1768/2062\n",
      "Processing image: 1769/2062\n",
      "Processing image: 1770/2062\n",
      "Processing image: 1771/2062\n",
      "Processing image: 1772/2062\n",
      "Processing image: 1773/2062\n",
      "Processing image: 1774/2062\n",
      "Processing image: 1775/2062\n",
      "Processing image: 1776/2062\n",
      "Processing image: 1777/2062\n",
      "Processing image: 1778/2062\n",
      "Processing image: 1779/2062\n",
      "Processing image: 1780/2062\n",
      "Processing image: 1781/2062\n",
      "Processing image: 1782/2062\n",
      "Processing image: 1783/2062\n",
      "Processing image: 1784/2062\n",
      "Processing image: 1785/2062\n",
      "Processing image: 1786/2062\n",
      "Processing image: 1787/2062\n",
      "Processing image: 1788/2062\n",
      "Processing image: 1789/2062\n",
      "Processing image: 1790/2062\n",
      "Processing image: 1791/2062\n",
      "Processing image: 1792/2062\n",
      "Processing image: 1793/2062\n",
      "Processing image: 1794/2062\n",
      "Processing image: 1795/2062\n",
      "Processing image: 1796/2062\n",
      "Processing image: 1797/2062\n",
      "Processing image: 1798/2062\n",
      "Processing image: 1799/2062\n",
      "Processing image: 1800/2062\n",
      "Processing image: 1801/2062\n",
      "Processing image: 1802/2062\n",
      "Processing image: 1803/2062\n",
      "Processing image: 1804/2062\n",
      "Processing image: 1805/2062\n",
      "Processing image: 1806/2062\n",
      "Processing image: 1807/2062\n",
      "Processing image: 1808/2062\n",
      "Processing image: 1809/2062\n",
      "Processing image: 1810/2062\n",
      "Processing image: 1811/2062\n",
      "Processing image: 1812/2062\n",
      "Processing image: 1813/2062\n",
      "Processing image: 1814/2062\n",
      "Processing image: 1815/2062\n",
      "Processing image: 1816/2062\n",
      "Processing image: 1817/2062\n",
      "Processing image: 1818/2062\n",
      "Processing image: 1819/2062\n",
      "Processing image: 1820/2062\n",
      "Processing image: 1821/2062\n",
      "Processing image: 1822/2062\n",
      "Processing image: 1823/2062\n",
      "Processing image: 1824/2062\n",
      "Processing image: 1825/2062\n",
      "Processing image: 1826/2062\n",
      "Processing image: 1827/2062\n",
      "Processing image: 1828/2062\n",
      "Processing image: 1829/2062\n",
      "Processing image: 1830/2062\n",
      "Processing image: 1831/2062\n",
      "Processing image: 1832/2062\n",
      "Processing image: 1833/2062\n",
      "Processing image: 1834/2062\n",
      "Processing image: 1835/2062\n",
      "Processing image: 1836/2062\n",
      "Processing image: 1837/2062\n",
      "Processing image: 1838/2062\n",
      "Processing image: 1839/2062\n",
      "Processing image: 1840/2062\n",
      "Processing image: 1841/2062\n",
      "Processing image: 1842/2062\n",
      "Processing image: 1843/2062\n",
      "Processing image: 1844/2062\n",
      "Processing image: 1845/2062\n",
      "Processing image: 1846/2062\n",
      "Processing image: 1847/2062\n",
      "Processing image: 1848/2062\n",
      "Processing image: 1849/2062\n",
      "Processing image: 1850/2062\n",
      "Processing image: 1851/2062\n",
      "Processing image: 1852/2062\n",
      "Processing image: 1853/2062\n",
      "Processing image: 1854/2062\n",
      "Processing image: 1855/2062\n",
      "Processing image: 1856/2062\n",
      "Processing image: 1857/2062\n",
      "Processing image: 1858/2062\n",
      "Processing image: 1859/2062\n",
      "Processing image: 1860/2062\n",
      "Processing image: 1861/2062\n",
      "Processing image: 1862/2062\n",
      "Processing image: 1863/2062\n",
      "Processing image: 1864/2062\n",
      "Processing image: 1865/2062\n",
      "Processing image: 1866/2062\n",
      "Processing image: 1867/2062\n",
      "Processing image: 1868/2062\n",
      "Processing image: 1869/2062\n",
      "Processing image: 1870/2062\n",
      "Processing image: 1871/2062\n",
      "Processing image: 1872/2062\n",
      "Processing image: 1873/2062\n",
      "Processing image: 1874/2062\n",
      "Processing image: 1875/2062\n",
      "Processing image: 1876/2062\n",
      "Processing image: 1877/2062\n",
      "Processing image: 1878/2062\n",
      "Processing image: 1879/2062\n",
      "Processing image: 1880/2062\n",
      "Processing image: 1881/2062\n",
      "Processing image: 1882/2062\n",
      "Processing image: 1883/2062\n",
      "Processing image: 1884/2062\n",
      "Processing image: 1885/2062\n",
      "Processing image: 1886/2062\n",
      "Processing image: 1887/2062\n",
      "Processing image: 1888/2062\n",
      "Processing image: 1889/2062\n",
      "Processing image: 1890/2062\n",
      "Processing image: 1891/2062\n",
      "Processing image: 1892/2062\n",
      "Processing image: 1893/2062\n",
      "Processing image: 1894/2062\n",
      "Processing image: 1895/2062\n",
      "Processing image: 1896/2062\n",
      "Processing image: 1897/2062\n",
      "Processing image: 1898/2062\n",
      "Processing image: 1899/2062\n",
      "Processing image: 1900/2062\n",
      "Processing image: 1901/2062\n",
      "Processing image: 1902/2062\n",
      "Processing image: 1903/2062\n",
      "Processing image: 1904/2062\n",
      "Processing image: 1905/2062\n",
      "Processing image: 1906/2062\n",
      "Processing image: 1907/2062\n",
      "Processing image: 1908/2062\n",
      "Processing image: 1909/2062\n",
      "Processing image: 1910/2062\n",
      "Processing image: 1911/2062\n",
      "Processing image: 1912/2062\n",
      "Processing image: 1913/2062\n",
      "Processing image: 1914/2062\n",
      "Processing image: 1915/2062\n",
      "Processing image: 1916/2062\n",
      "Processing image: 1917/2062\n",
      "Processing image: 1918/2062\n",
      "Processing image: 1919/2062\n",
      "Processing image: 1920/2062\n",
      "Processing image: 1921/2062\n",
      "Processing image: 1922/2062\n",
      "Processing image: 1923/2062\n",
      "Processing image: 1924/2062\n",
      "Processing image: 1925/2062\n",
      "Processing image: 1926/2062\n",
      "Processing image: 1927/2062\n",
      "Processing image: 1928/2062\n",
      "Processing image: 1929/2062\n",
      "Processing image: 1930/2062\n",
      "Processing image: 1931/2062\n",
      "Processing image: 1932/2062\n",
      "Processing image: 1933/2062\n",
      "Processing image: 1934/2062\n",
      "Processing image: 1935/2062\n",
      "Processing image: 1936/2062\n",
      "Processing image: 1937/2062\n",
      "Processing image: 1938/2062\n",
      "Processing image: 1939/2062\n",
      "Processing image: 1940/2062\n",
      "Processing image: 1941/2062\n",
      "Processing image: 1942/2062\n",
      "Processing image: 1943/2062\n",
      "Processing image: 1944/2062\n",
      "Processing image: 1945/2062\n",
      "Processing image: 1946/2062\n",
      "Processing image: 1947/2062\n",
      "Processing image: 1948/2062\n",
      "Processing image: 1949/2062\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1950/2062\n",
      "Processing image: 1951/2062\n",
      "Processing image: 1952/2062\n",
      "Processing image: 1953/2062\n",
      "Processing image: 1954/2062\n",
      "Processing image: 1955/2062\n",
      "Processing image: 1956/2062\n",
      "Processing image: 1957/2062\n",
      "Processing image: 1958/2062\n",
      "Processing image: 1959/2062\n",
      "Processing image: 1960/2062\n",
      "Processing image: 1961/2062\n",
      "Processing image: 1962/2062\n",
      "Processing image: 1963/2062\n",
      "Processing image: 1964/2062\n",
      "Processing image: 1965/2062\n",
      "Processing image: 1966/2062\n",
      "Processing image: 1967/2062\n",
      "Processing image: 1968/2062\n",
      "Processing image: 1969/2062\n",
      "Processing image: 1970/2062\n",
      "Processing image: 1971/2062\n",
      "Processing image: 1972/2062\n",
      "Processing image: 1973/2062\n",
      "Processing image: 1974/2062\n",
      "Processing image: 1975/2062\n",
      "Processing image: 1976/2062\n",
      "Processing image: 1977/2062\n",
      "Processing image: 1978/2062\n",
      "Processing image: 1979/2062\n",
      "Processing image: 1980/2062\n",
      "Processing image: 1981/2062\n",
      "Processing image: 1982/2062\n",
      "Processing image: 1983/2062\n",
      "Processing image: 1984/2062\n",
      "Processing image: 1985/2062\n",
      "Processing image: 1986/2062\n",
      "Processing image: 1987/2062\n",
      "Processing image: 1988/2062\n",
      "Processing image: 1989/2062\n",
      "Processing image: 1990/2062\n",
      "Processing image: 1991/2062\n",
      "Processing image: 1992/2062\n",
      "Processing image: 1993/2062\n",
      "Processing image: 1994/2062\n",
      "Processing image: 1995/2062\n",
      "Processing image: 1996/2062\n",
      "Processing image: 1997/2062\n",
      "Processing image: 1998/2062\n",
      "Processing image: 1999/2062\n",
      "Processing image: 2000/2062\n",
      "Processing image: 2001/2062\n",
      "Processing image: 2002/2062\n",
      "Processing image: 2003/2062\n",
      "Processing image: 2004/2062\n",
      "Processing image: 2005/2062\n",
      "Processing image: 2006/2062\n",
      "Processing image: 2007/2062\n",
      "Processing image: 2008/2062\n",
      "Processing image: 2009/2062\n",
      "Processing image: 2010/2062\n",
      "Processing image: 2011/2062\n",
      "Processing image: 2012/2062\n",
      "Processing image: 2013/2062\n",
      "Processing image: 2014/2062\n",
      "Processing image: 2015/2062\n",
      "Processing image: 2016/2062\n",
      "Processing image: 2017/2062\n",
      "Processing image: 2018/2062\n",
      "Processing image: 2019/2062\n",
      "Processing image: 2020/2062\n",
      "Processing image: 2021/2062\n",
      "Processing image: 2022/2062\n",
      "Processing image: 2023/2062\n",
      "Processing image: 2024/2062\n",
      "Processing image: 2025/2062\n",
      "Processing image: 2026/2062\n",
      "Processing image: 2027/2062\n",
      "Processing image: 2028/2062\n",
      "Processing image: 2029/2062\n",
      "Processing image: 2030/2062\n",
      "Processing image: 2031/2062\n",
      "Processing image: 2032/2062\n",
      "Processing image: 2033/2062\n",
      "Processing image: 2034/2062\n",
      "Processing image: 2035/2062\n",
      "Processing image: 2036/2062\n",
      "Processing image: 2037/2062\n",
      "Processing image: 2038/2062\n",
      "Processing image: 2039/2062\n",
      "Processing image: 2040/2062\n",
      "Processing image: 2041/2062\n",
      "Processing image: 2042/2062\n",
      "Processing image: 2043/2062\n",
      "Processing image: 2044/2062\n",
      "Processing image: 2045/2062\n",
      "Processing image: 2046/2062\n",
      "Processing image: 2047/2062\n",
      "Processing image: 2048/2062\n",
      "Processing image: 2049/2062\n",
      "Processing image: 2050/2062\n",
      "Processing image: 2051/2062\n",
      "Processing image: 2052/2062\n",
      "Processing image: 2053/2062\n",
      "Processing image: 2054/2062\n",
      "Processing image: 2055/2062\n",
      "Processing image: 2056/2062\n",
      "Processing image: 2057/2062\n",
      "Processing image: 2058/2062\n",
      "Processing image: 2059/2062\n",
      "Processing image: 2060/2062\n",
      "Processing image: 2061/2062\n",
      "Processing image: 2062/2062\n"
     ]
    }
   ],
   "source": [
    "evaluate(aw_paths, aw_res_path, model, device, img_size, con_thr, iou_thr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1/5000\n",
      "Processing image: 2/5000\n",
      "Processing image: 3/5000\n",
      "Processing image: 4/5000\n",
      "Processing image: 5/5000\n",
      "Processing image: 6/5000\n",
      "Processing image: 7/5000\n",
      "Processing image: 8/5000\n",
      "Processing image: 9/5000\n",
      "Processing image: 10/5000\n",
      "Processing image: 11/5000\n",
      "Processing image: 12/5000\n",
      "Processing image: 13/5000\n",
      "Processing image: 14/5000\n",
      "Processing image: 15/5000\n",
      "Processing image: 16/5000\n",
      "Processing image: 17/5000\n",
      "Processing image: 18/5000\n",
      "Processing image: 19/5000\n",
      "Processing image: 20/5000\n",
      "Processing image: 21/5000\n",
      "Processing image: 22/5000\n",
      "Processing image: 23/5000\n",
      "Processing image: 24/5000\n",
      "Processing image: 25/5000\n",
      "Processing image: 26/5000\n",
      "Processing image: 27/5000\n",
      "Processing image: 28/5000\n",
      "Processing image: 29/5000\n",
      "Processing image: 30/5000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:49: RuntimeWarning: invalid value encountered in true_divide\n",
      "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:51: RuntimeWarning: invalid value encountered in less\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 31/5000\n",
      "Processing image: 32/5000\n",
      "Processing image: 33/5000\n",
      "Processing image: 34/5000\n",
      "Processing image: 35/5000\n",
      "Processing image: 36/5000\n",
      "Processing image: 37/5000\n",
      "Processing image: 38/5000\n",
      "Processing image: 39/5000\n",
      "Processing image: 40/5000\n",
      "Processing image: 41/5000\n",
      "Processing image: 42/5000\n",
      "Processing image: 43/5000\n",
      "Processing image: 44/5000\n",
      "Processing image: 45/5000\n",
      "Processing image: 46/5000\n",
      "Processing image: 47/5000\n",
      "Processing image: 48/5000\n",
      "Processing image: 49/5000\n",
      "Processing image: 50/5000\n",
      "Processing image: 51/5000\n",
      "Processing image: 52/5000\n",
      "Processing image: 53/5000\n",
      "Processing image: 54/5000\n",
      "Processing image: 55/5000\n",
      "Processing image: 56/5000\n",
      "Processing image: 57/5000\n",
      "Processing image: 58/5000\n",
      "Processing image: 59/5000\n",
      "Processing image: 60/5000\n",
      "Processing image: 61/5000\n",
      "Processing image: 62/5000\n",
      "Processing image: 63/5000\n",
      "Processing image: 64/5000\n",
      "Processing image: 65/5000\n",
      "Processing image: 66/5000\n",
      "Processing image: 67/5000\n",
      "Processing image: 68/5000\n",
      "Processing image: 69/5000\n",
      "Processing image: 70/5000\n",
      "Processing image: 71/5000\n",
      "Processing image: 72/5000\n",
      "Processing image: 73/5000\n",
      "Processing image: 74/5000\n",
      "Processing image: 75/5000\n",
      "Processing image: 76/5000\n",
      "Processing image: 77/5000\n",
      "Processing image: 78/5000\n",
      "Processing image: 79/5000\n",
      "Processing image: 80/5000\n",
      "Processing image: 81/5000\n",
      "Processing image: 82/5000\n",
      "Processing image: 83/5000\n",
      "Processing image: 84/5000\n",
      "Processing image: 85/5000\n",
      "Processing image: 86/5000\n",
      "Processing image: 87/5000\n",
      "Processing image: 88/5000\n",
      "Processing image: 89/5000\n",
      "Processing image: 90/5000\n",
      "Processing image: 91/5000\n",
      "Processing image: 92/5000\n",
      "Processing image: 93/5000\n",
      "Processing image: 94/5000\n",
      "Processing image: 95/5000\n",
      "Processing image: 96/5000\n",
      "Processing image: 97/5000\n",
      "Processing image: 98/5000\n",
      "Processing image: 99/5000\n",
      "Processing image: 100/5000\n",
      "Processing image: 101/5000\n",
      "Processing image: 102/5000\n",
      "Processing image: 103/5000\n",
      "Processing image: 104/5000\n",
      "Processing image: 105/5000\n",
      "Processing image: 106/5000\n",
      "Processing image: 107/5000\n",
      "Processing image: 108/5000\n",
      "Processing image: 109/5000\n",
      "Processing image: 110/5000\n",
      "Processing image: 111/5000\n",
      "Processing image: 112/5000\n",
      "Processing image: 113/5000\n",
      "Processing image: 114/5000\n",
      "Processing image: 115/5000\n",
      "Processing image: 116/5000\n",
      "Processing image: 117/5000\n",
      "Processing image: 118/5000\n",
      "Processing image: 119/5000\n",
      "Processing image: 120/5000\n",
      "Processing image: 121/5000\n",
      "Processing image: 122/5000\n",
      "Processing image: 123/5000\n",
      "Processing image: 124/5000\n",
      "Processing image: 125/5000\n",
      "Processing image: 126/5000\n",
      "Processing image: 127/5000\n",
      "Processing image: 128/5000\n",
      "Processing image: 129/5000\n",
      "Processing image: 130/5000\n",
      "Processing image: 131/5000\n",
      "Processing image: 132/5000\n",
      "Processing image: 133/5000\n",
      "Processing image: 134/5000\n",
      "Processing image: 135/5000\n",
      "Processing image: 136/5000\n",
      "Processing image: 137/5000\n",
      "Processing image: 138/5000\n",
      "Processing image: 139/5000\n",
      "Processing image: 140/5000\n",
      "Processing image: 141/5000\n",
      "Processing image: 142/5000\n",
      "Processing image: 143/5000\n",
      "Processing image: 144/5000\n",
      "Processing image: 145/5000\n",
      "Processing image: 146/5000\n",
      "Processing image: 147/5000\n",
      "Processing image: 148/5000\n",
      "Processing image: 149/5000\n",
      "Processing image: 150/5000\n",
      "Processing image: 151/5000\n",
      "Processing image: 152/5000\n",
      "Processing image: 153/5000\n",
      "Processing image: 154/5000\n",
      "Processing image: 155/5000\n",
      "Processing image: 156/5000\n",
      "Processing image: 157/5000\n",
      "Processing image: 158/5000\n",
      "Processing image: 159/5000\n",
      "Processing image: 160/5000\n",
      "Processing image: 161/5000\n",
      "Processing image: 162/5000\n",
      "Processing image: 163/5000\n",
      "Processing image: 164/5000\n",
      "Processing image: 165/5000\n",
      "Processing image: 166/5000\n",
      "Processing image: 167/5000\n",
      "Processing image: 168/5000\n",
      "Processing image: 169/5000\n",
      "Processing image: 170/5000\n",
      "Processing image: 171/5000\n",
      "Processing image: 172/5000\n",
      "Processing image: 173/5000\n",
      "Processing image: 174/5000\n",
      "Processing image: 175/5000\n",
      "Processing image: 176/5000\n",
      "Processing image: 177/5000\n",
      "Processing image: 178/5000\n",
      "Processing image: 179/5000\n",
      "Processing image: 180/5000\n",
      "Processing image: 181/5000\n",
      "Processing image: 182/5000\n",
      "Processing image: 183/5000\n",
      "Processing image: 184/5000\n",
      "Processing image: 185/5000\n",
      "Processing image: 186/5000\n",
      "Processing image: 187/5000\n",
      "Processing image: 188/5000\n",
      "Processing image: 189/5000\n",
      "Processing image: 190/5000\n",
      "Processing image: 191/5000\n",
      "Processing image: 192/5000\n",
      "Processing image: 193/5000\n",
      "Processing image: 194/5000\n",
      "Processing image: 195/5000\n",
      "Processing image: 196/5000\n",
      "Processing image: 197/5000\n",
      "Processing image: 198/5000\n",
      "Processing image: 199/5000\n",
      "Processing image: 200/5000\n",
      "Processing image: 201/5000\n",
      "Processing image: 202/5000\n",
      "Processing image: 203/5000\n",
      "Processing image: 204/5000\n",
      "Processing image: 205/5000\n",
      "Processing image: 206/5000\n",
      "Processing image: 207/5000\n",
      "Processing image: 208/5000\n",
      "Processing image: 209/5000\n",
      "Processing image: 210/5000\n",
      "Processing image: 211/5000\n",
      "Processing image: 212/5000\n",
      "Processing image: 213/5000\n",
      "Processing image: 214/5000\n",
      "Processing image: 215/5000\n",
      "Processing image: 216/5000\n",
      "Processing image: 217/5000\n",
      "Processing image: 218/5000\n",
      "Processing image: 219/5000\n",
      "Processing image: 220/5000\n",
      "Processing image: 221/5000\n",
      "Processing image: 222/5000\n",
      "Processing image: 223/5000\n",
      "Processing image: 224/5000\n",
      "Processing image: 225/5000\n",
      "Processing image: 226/5000\n",
      "Processing image: 227/5000\n",
      "Processing image: 228/5000\n",
      "Processing image: 229/5000\n",
      "Processing image: 230/5000\n",
      "Processing image: 231/5000\n",
      "Processing image: 232/5000\n",
      "Processing image: 233/5000\n",
      "Processing image: 234/5000\n",
      "Processing image: 235/5000\n",
      "Processing image: 236/5000\n",
      "Processing image: 237/5000\n",
      "Processing image: 238/5000\n",
      "Processing image: 239/5000\n",
      "Processing image: 240/5000\n",
      "Processing image: 241/5000\n",
      "Processing image: 242/5000\n",
      "Processing image: 243/5000\n",
      "Processing image: 244/5000\n",
      "Processing image: 245/5000\n",
      "Processing image: 246/5000\n",
      "Processing image: 247/5000\n",
      "Processing image: 248/5000\n",
      "Processing image: 249/5000\n",
      "Processing image: 250/5000\n",
      "Processing image: 251/5000\n",
      "Processing image: 252/5000\n",
      "Processing image: 253/5000\n",
      "Processing image: 254/5000\n",
      "Processing image: 255/5000\n",
      "Processing image: 256/5000\n",
      "Processing image: 257/5000\n",
      "Processing image: 258/5000\n",
      "Processing image: 259/5000\n",
      "Processing image: 260/5000\n",
      "Processing image: 261/5000\n",
      "Processing image: 262/5000\n",
      "Processing image: 263/5000\n",
      "Processing image: 264/5000\n",
      "Processing image: 265/5000\n",
      "Processing image: 266/5000\n",
      "Processing image: 267/5000\n",
      "Processing image: 268/5000\n",
      "Processing image: 269/5000\n",
      "Processing image: 270/5000\n",
      "Processing image: 271/5000\n",
      "Processing image: 272/5000\n",
      "Processing image: 273/5000\n",
      "Processing image: 274/5000\n",
      "Processing image: 275/5000\n",
      "Processing image: 276/5000\n",
      "Processing image: 277/5000\n",
      "Processing image: 278/5000\n",
      "Processing image: 279/5000\n",
      "Processing image: 280/5000\n",
      "Processing image: 281/5000\n",
      "Processing image: 282/5000\n",
      "Processing image: 283/5000\n",
      "Processing image: 284/5000\n",
      "Processing image: 285/5000\n",
      "Processing image: 286/5000\n",
      "Processing image: 287/5000\n",
      "Processing image: 288/5000\n",
      "Processing image: 289/5000\n",
      "Processing image: 290/5000\n",
      "Processing image: 291/5000\n",
      "Processing image: 292/5000\n",
      "Processing image: 293/5000\n",
      "Processing image: 294/5000\n",
      "Processing image: 295/5000\n",
      "Processing image: 296/5000\n",
      "Processing image: 297/5000\n",
      "Processing image: 298/5000\n",
      "Processing image: 299/5000\n",
      "Processing image: 300/5000\n",
      "Processing image: 301/5000\n",
      "Processing image: 302/5000\n",
      "Processing image: 303/5000\n",
      "Processing image: 304/5000\n",
      "Processing image: 305/5000\n",
      "Processing image: 306/5000\n",
      "Processing image: 307/5000\n",
      "Processing image: 308/5000\n",
      "Processing image: 309/5000\n",
      "Processing image: 310/5000\n",
      "Processing image: 311/5000\n",
      "Processing image: 312/5000\n",
      "Processing image: 313/5000\n",
      "Processing image: 314/5000\n",
      "Processing image: 315/5000\n",
      "Processing image: 316/5000\n",
      "Processing image: 317/5000\n",
      "Processing image: 318/5000\n",
      "Processing image: 319/5000\n",
      "Processing image: 320/5000\n",
      "Processing image: 321/5000\n",
      "Processing image: 322/5000\n",
      "Processing image: 323/5000\n",
      "Processing image: 324/5000\n",
      "Processing image: 325/5000\n",
      "Processing image: 326/5000\n",
      "Processing image: 327/5000\n",
      "Processing image: 328/5000\n",
      "Processing image: 329/5000\n",
      "Processing image: 330/5000\n",
      "Processing image: 331/5000\n",
      "Processing image: 332/5000\n",
      "Processing image: 333/5000\n",
      "Processing image: 334/5000\n",
      "Processing image: 335/5000\n",
      "Processing image: 336/5000\n",
      "Processing image: 337/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 338/5000\n",
      "Processing image: 339/5000\n",
      "Processing image: 340/5000\n",
      "Processing image: 341/5000\n",
      "Processing image: 342/5000\n",
      "Processing image: 343/5000\n",
      "Processing image: 344/5000\n",
      "Processing image: 345/5000\n",
      "Processing image: 346/5000\n",
      "Processing image: 347/5000\n",
      "Processing image: 348/5000\n",
      "Processing image: 349/5000\n",
      "Processing image: 350/5000\n",
      "Processing image: 351/5000\n",
      "Processing image: 352/5000\n",
      "Processing image: 353/5000\n",
      "Processing image: 354/5000\n",
      "Processing image: 355/5000\n",
      "Processing image: 356/5000\n",
      "Processing image: 357/5000\n",
      "Processing image: 358/5000\n",
      "Processing image: 359/5000\n",
      "Processing image: 360/5000\n",
      "Processing image: 361/5000\n",
      "Processing image: 362/5000\n",
      "Processing image: 363/5000\n",
      "Processing image: 364/5000\n",
      "Processing image: 365/5000\n",
      "Processing image: 366/5000\n",
      "Processing image: 367/5000\n",
      "Processing image: 368/5000\n",
      "Processing image: 369/5000\n",
      "Processing image: 370/5000\n",
      "Processing image: 371/5000\n",
      "Processing image: 372/5000\n",
      "Processing image: 373/5000\n",
      "Processing image: 374/5000\n",
      "Processing image: 375/5000\n",
      "Processing image: 376/5000\n",
      "Processing image: 377/5000\n",
      "Processing image: 378/5000\n",
      "Processing image: 379/5000\n",
      "Processing image: 380/5000\n",
      "Processing image: 381/5000\n",
      "Processing image: 382/5000\n",
      "Processing image: 383/5000\n",
      "Processing image: 384/5000\n",
      "Processing image: 385/5000\n",
      "Processing image: 386/5000\n",
      "Processing image: 387/5000\n",
      "Processing image: 388/5000\n",
      "Processing image: 389/5000\n",
      "Processing image: 390/5000\n",
      "Processing image: 391/5000\n",
      "Processing image: 392/5000\n",
      "Processing image: 393/5000\n",
      "Processing image: 394/5000\n",
      "Processing image: 395/5000\n",
      "Processing image: 396/5000\n",
      "Processing image: 397/5000\n",
      "Processing image: 398/5000\n",
      "Processing image: 399/5000\n",
      "Processing image: 400/5000\n",
      "Processing image: 401/5000\n",
      "Processing image: 402/5000\n",
      "Processing image: 403/5000\n",
      "Processing image: 404/5000\n",
      "Processing image: 405/5000\n",
      "Processing image: 406/5000\n",
      "Processing image: 407/5000\n",
      "Processing image: 408/5000\n",
      "Processing image: 409/5000\n",
      "Processing image: 410/5000\n",
      "Processing image: 411/5000\n",
      "Processing image: 412/5000\n",
      "Processing image: 413/5000\n",
      "Processing image: 414/5000\n",
      "Processing image: 415/5000\n",
      "Processing image: 416/5000\n",
      "Processing image: 417/5000\n",
      "Processing image: 418/5000\n",
      "Processing image: 419/5000\n",
      "Processing image: 420/5000\n",
      "Processing image: 421/5000\n",
      "Processing image: 422/5000\n",
      "Processing image: 423/5000\n",
      "Processing image: 424/5000\n",
      "Processing image: 425/5000\n",
      "Processing image: 426/5000\n",
      "Processing image: 427/5000\n",
      "Processing image: 428/5000\n",
      "Processing image: 429/5000\n",
      "Processing image: 430/5000\n",
      "Processing image: 431/5000\n",
      "Processing image: 432/5000\n",
      "Processing image: 433/5000\n",
      "Processing image: 434/5000\n",
      "Processing image: 435/5000\n",
      "Processing image: 436/5000\n",
      "Processing image: 437/5000\n",
      "Processing image: 438/5000\n",
      "Processing image: 439/5000\n",
      "Processing image: 440/5000\n",
      "Processing image: 441/5000\n",
      "Processing image: 442/5000\n",
      "Processing image: 443/5000\n",
      "Processing image: 444/5000\n",
      "Processing image: 445/5000\n",
      "Processing image: 446/5000\n",
      "Processing image: 447/5000\n",
      "Processing image: 448/5000\n",
      "Processing image: 449/5000\n",
      "Processing image: 450/5000\n",
      "Processing image: 451/5000\n",
      "Processing image: 452/5000\n",
      "Processing image: 453/5000\n",
      "Processing image: 454/5000\n",
      "Processing image: 455/5000\n",
      "Processing image: 456/5000\n",
      "Processing image: 457/5000\n",
      "Processing image: 458/5000\n",
      "Processing image: 459/5000\n",
      "Processing image: 460/5000\n",
      "Processing image: 461/5000\n",
      "Processing image: 462/5000\n",
      "Processing image: 463/5000\n",
      "Processing image: 464/5000\n",
      "Processing image: 465/5000\n",
      "Processing image: 466/5000\n",
      "Processing image: 467/5000\n",
      "Processing image: 468/5000\n",
      "Processing image: 469/5000\n",
      "Processing image: 470/5000\n",
      "Processing image: 471/5000\n",
      "Processing image: 472/5000\n",
      "Processing image: 473/5000\n",
      "Processing image: 474/5000\n",
      "Processing image: 475/5000\n",
      "Processing image: 476/5000\n",
      "Processing image: 477/5000\n",
      "Processing image: 478/5000\n",
      "Processing image: 479/5000\n",
      "Processing image: 480/5000\n",
      "Processing image: 481/5000\n",
      "Processing image: 482/5000\n",
      "Processing image: 483/5000\n",
      "Processing image: 484/5000\n",
      "Processing image: 485/5000\n",
      "Processing image: 486/5000\n",
      "Processing image: 487/5000\n",
      "Processing image: 488/5000\n",
      "Processing image: 489/5000\n",
      "Processing image: 490/5000\n",
      "Processing image: 491/5000\n",
      "Processing image: 492/5000\n",
      "Processing image: 493/5000\n",
      "Processing image: 494/5000\n",
      "Processing image: 495/5000\n",
      "Processing image: 496/5000\n",
      "Processing image: 497/5000\n",
      "Processing image: 498/5000\n",
      "Processing image: 499/5000\n",
      "Processing image: 500/5000\n",
      "Processing image: 501/5000\n",
      "Processing image: 502/5000\n",
      "Processing image: 503/5000\n",
      "Processing image: 504/5000\n",
      "Processing image: 505/5000\n",
      "Processing image: 506/5000\n",
      "Processing image: 507/5000\n",
      "Processing image: 508/5000\n",
      "Processing image: 509/5000\n",
      "Processing image: 510/5000\n",
      "Processing image: 511/5000\n",
      "Processing image: 512/5000\n",
      "Processing image: 513/5000\n",
      "Processing image: 514/5000\n",
      "Processing image: 515/5000\n",
      "Processing image: 516/5000\n",
      "Processing image: 517/5000\n",
      "Processing image: 518/5000\n",
      "Processing image: 519/5000\n",
      "Processing image: 520/5000\n",
      "Processing image: 521/5000\n",
      "Processing image: 522/5000\n",
      "Processing image: 523/5000\n",
      "Processing image: 524/5000\n",
      "Processing image: 525/5000\n",
      "Processing image: 526/5000\n",
      "Processing image: 527/5000\n",
      "Processing image: 528/5000\n",
      "Processing image: 529/5000\n",
      "Processing image: 530/5000\n",
      "Processing image: 531/5000\n",
      "Processing image: 532/5000\n",
      "Processing image: 533/5000\n",
      "Processing image: 534/5000\n",
      "Processing image: 535/5000\n",
      "Processing image: 536/5000\n",
      "Processing image: 537/5000\n",
      "Processing image: 538/5000\n",
      "Processing image: 539/5000\n",
      "Processing image: 540/5000\n",
      "Processing image: 541/5000\n",
      "Processing image: 542/5000\n",
      "Processing image: 543/5000\n",
      "Processing image: 544/5000\n",
      "Processing image: 545/5000\n",
      "Processing image: 546/5000\n",
      "Processing image: 547/5000\n",
      "Processing image: 548/5000\n",
      "Processing image: 549/5000\n",
      "Processing image: 550/5000\n",
      "Processing image: 551/5000\n",
      "Processing image: 552/5000\n",
      "Processing image: 553/5000\n",
      "Processing image: 554/5000\n",
      "Processing image: 555/5000\n",
      "Processing image: 556/5000\n",
      "Processing image: 557/5000\n",
      "Processing image: 558/5000\n",
      "Processing image: 559/5000\n",
      "Processing image: 560/5000\n",
      "Processing image: 561/5000\n",
      "Processing image: 562/5000\n",
      "Processing image: 563/5000\n",
      "Processing image: 564/5000\n",
      "Processing image: 565/5000\n",
      "Processing image: 566/5000\n",
      "Processing image: 567/5000\n",
      "Processing image: 568/5000\n",
      "Processing image: 569/5000\n",
      "Processing image: 570/5000\n",
      "Processing image: 571/5000\n",
      "Processing image: 572/5000\n",
      "Processing image: 573/5000\n",
      "Processing image: 574/5000\n",
      "Processing image: 575/5000\n",
      "Processing image: 576/5000\n",
      "Processing image: 577/5000\n",
      "Processing image: 578/5000\n",
      "Processing image: 579/5000\n",
      "Processing image: 580/5000\n",
      "Processing image: 581/5000\n",
      "Processing image: 582/5000\n",
      "Processing image: 583/5000\n",
      "Processing image: 584/5000\n",
      "Processing image: 585/5000\n",
      "Processing image: 586/5000\n",
      "Processing image: 587/5000\n",
      "Processing image: 588/5000\n",
      "Processing image: 589/5000\n",
      "Processing image: 590/5000\n",
      "Processing image: 591/5000\n",
      "Processing image: 592/5000\n",
      "Processing image: 593/5000\n",
      "Processing image: 594/5000\n",
      "Processing image: 595/5000\n",
      "Processing image: 596/5000\n",
      "Processing image: 597/5000\n",
      "Processing image: 598/5000\n",
      "Processing image: 599/5000\n",
      "Processing image: 600/5000\n",
      "Processing image: 601/5000\n",
      "Processing image: 602/5000\n",
      "Processing image: 603/5000\n",
      "Processing image: 604/5000\n",
      "Processing image: 605/5000\n",
      "Processing image: 606/5000\n",
      "Processing image: 607/5000\n",
      "Processing image: 608/5000\n",
      "Processing image: 609/5000\n",
      "Processing image: 610/5000\n",
      "Processing image: 611/5000\n",
      "Processing image: 612/5000\n",
      "Processing image: 613/5000\n",
      "Processing image: 614/5000\n",
      "Processing image: 615/5000\n",
      "Processing image: 616/5000\n",
      "Processing image: 617/5000\n",
      "Processing image: 618/5000\n",
      "Processing image: 619/5000\n",
      "Processing image: 620/5000\n",
      "Processing image: 621/5000\n",
      "Processing image: 622/5000\n",
      "Processing image: 623/5000\n",
      "Processing image: 624/5000\n",
      "Processing image: 625/5000\n",
      "Processing image: 626/5000\n",
      "Processing image: 627/5000\n",
      "Processing image: 628/5000\n",
      "Processing image: 629/5000\n",
      "Processing image: 630/5000\n",
      "Processing image: 631/5000\n",
      "Processing image: 632/5000\n",
      "Processing image: 633/5000\n",
      "Processing image: 634/5000\n",
      "Processing image: 635/5000\n",
      "Processing image: 636/5000\n",
      "Processing image: 637/5000\n",
      "Processing image: 638/5000\n",
      "Processing image: 639/5000\n",
      "Processing image: 640/5000\n",
      "Processing image: 641/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 642/5000\n",
      "Processing image: 643/5000\n",
      "Processing image: 644/5000\n",
      "Processing image: 645/5000\n",
      "Processing image: 646/5000\n",
      "Processing image: 647/5000\n",
      "Processing image: 648/5000\n",
      "Processing image: 649/5000\n",
      "Processing image: 650/5000\n",
      "Processing image: 651/5000\n",
      "Processing image: 652/5000\n",
      "Processing image: 653/5000\n",
      "Processing image: 654/5000\n",
      "Processing image: 655/5000\n",
      "Processing image: 656/5000\n",
      "Processing image: 657/5000\n",
      "Processing image: 658/5000\n",
      "Processing image: 659/5000\n",
      "Processing image: 660/5000\n",
      "Processing image: 661/5000\n",
      "Processing image: 662/5000\n",
      "Processing image: 663/5000\n",
      "Processing image: 664/5000\n",
      "Processing image: 665/5000\n",
      "Processing image: 666/5000\n",
      "Processing image: 667/5000\n",
      "Processing image: 668/5000\n",
      "Processing image: 669/5000\n",
      "Processing image: 670/5000\n",
      "Processing image: 671/5000\n",
      "Processing image: 672/5000\n",
      "Processing image: 673/5000\n",
      "Processing image: 674/5000\n",
      "Processing image: 675/5000\n",
      "Processing image: 676/5000\n",
      "Processing image: 677/5000\n",
      "Processing image: 678/5000\n",
      "Processing image: 679/5000\n",
      "Processing image: 680/5000\n",
      "Processing image: 681/5000\n",
      "Processing image: 682/5000\n",
      "Processing image: 683/5000\n",
      "Processing image: 684/5000\n",
      "Processing image: 685/5000\n",
      "Processing image: 686/5000\n",
      "Processing image: 687/5000\n",
      "Processing image: 688/5000\n",
      "Processing image: 689/5000\n",
      "Processing image: 690/5000\n",
      "Processing image: 691/5000\n",
      "Processing image: 692/5000\n",
      "Processing image: 693/5000\n",
      "Processing image: 694/5000\n",
      "Processing image: 695/5000\n",
      "Processing image: 696/5000\n",
      "Processing image: 697/5000\n",
      "Processing image: 698/5000\n",
      "Processing image: 699/5000\n",
      "Processing image: 700/5000\n",
      "Processing image: 701/5000\n",
      "Processing image: 702/5000\n",
      "Processing image: 703/5000\n",
      "Processing image: 704/5000\n",
      "Processing image: 705/5000\n",
      "Processing image: 706/5000\n",
      "Processing image: 707/5000\n",
      "Processing image: 708/5000\n",
      "Processing image: 709/5000\n",
      "Processing image: 710/5000\n",
      "Processing image: 711/5000\n",
      "Processing image: 712/5000\n",
      "Processing image: 713/5000\n",
      "Processing image: 714/5000\n",
      "Processing image: 715/5000\n",
      "Processing image: 716/5000\n",
      "Processing image: 717/5000\n",
      "Processing image: 718/5000\n",
      "Processing image: 719/5000\n",
      "Processing image: 720/5000\n",
      "Processing image: 721/5000\n",
      "Processing image: 722/5000\n",
      "Processing image: 723/5000\n",
      "Processing image: 724/5000\n",
      "Processing image: 725/5000\n",
      "Processing image: 726/5000\n",
      "Processing image: 727/5000\n",
      "Processing image: 728/5000\n",
      "Processing image: 729/5000\n",
      "Processing image: 730/5000\n",
      "Processing image: 731/5000\n",
      "Processing image: 732/5000\n",
      "Processing image: 733/5000\n",
      "Processing image: 734/5000\n",
      "Processing image: 735/5000\n",
      "Processing image: 736/5000\n",
      "Processing image: 737/5000\n",
      "Processing image: 738/5000\n",
      "Processing image: 739/5000\n",
      "Processing image: 740/5000\n",
      "Processing image: 741/5000\n",
      "Processing image: 742/5000\n",
      "Processing image: 743/5000\n",
      "Processing image: 744/5000\n",
      "Processing image: 745/5000\n",
      "Processing image: 746/5000\n",
      "Processing image: 747/5000\n",
      "Processing image: 748/5000\n",
      "Processing image: 749/5000\n",
      "Processing image: 750/5000\n",
      "Processing image: 751/5000\n",
      "Processing image: 752/5000\n",
      "Processing image: 753/5000\n",
      "Processing image: 754/5000\n",
      "Processing image: 755/5000\n",
      "Processing image: 756/5000\n",
      "Processing image: 757/5000\n",
      "Processing image: 758/5000\n",
      "Processing image: 759/5000\n",
      "Processing image: 760/5000\n",
      "Processing image: 761/5000\n",
      "Processing image: 762/5000\n",
      "Processing image: 763/5000\n",
      "Processing image: 764/5000\n",
      "Processing image: 765/5000\n",
      "Processing image: 766/5000\n",
      "Processing image: 767/5000\n",
      "Processing image: 768/5000\n",
      "Processing image: 769/5000\n",
      "Processing image: 770/5000\n",
      "Processing image: 771/5000\n",
      "Processing image: 772/5000\n",
      "Processing image: 773/5000\n",
      "Processing image: 774/5000\n",
      "Processing image: 775/5000\n",
      "Processing image: 776/5000\n",
      "Processing image: 777/5000\n",
      "Processing image: 778/5000\n",
      "Processing image: 779/5000\n",
      "Processing image: 780/5000\n",
      "Processing image: 781/5000\n",
      "Processing image: 782/5000\n",
      "Processing image: 783/5000\n",
      "Processing image: 784/5000\n",
      "Processing image: 785/5000\n",
      "Processing image: 786/5000\n",
      "Processing image: 787/5000\n",
      "Processing image: 788/5000\n",
      "Processing image: 789/5000\n",
      "Processing image: 790/5000\n",
      "Processing image: 791/5000\n",
      "Processing image: 792/5000\n",
      "Processing image: 793/5000\n",
      "Processing image: 794/5000\n",
      "Processing image: 795/5000\n",
      "Processing image: 796/5000\n",
      "Processing image: 797/5000\n",
      "Processing image: 798/5000\n",
      "Processing image: 799/5000\n",
      "Processing image: 800/5000\n",
      "Processing image: 801/5000\n",
      "Processing image: 802/5000\n",
      "Processing image: 803/5000\n",
      "Processing image: 804/5000\n",
      "Processing image: 805/5000\n",
      "Processing image: 806/5000\n",
      "Processing image: 807/5000\n",
      "Processing image: 808/5000\n",
      "Processing image: 809/5000\n",
      "Processing image: 810/5000\n",
      "Processing image: 811/5000\n",
      "Processing image: 812/5000\n",
      "Processing image: 813/5000\n",
      "Processing image: 814/5000\n",
      "Processing image: 815/5000\n",
      "Processing image: 816/5000\n",
      "Processing image: 817/5000\n",
      "Processing image: 818/5000\n",
      "Processing image: 819/5000\n",
      "Processing image: 820/5000\n",
      "Processing image: 821/5000\n",
      "Processing image: 822/5000\n",
      "Processing image: 823/5000\n",
      "Processing image: 824/5000\n",
      "Processing image: 825/5000\n",
      "Processing image: 826/5000\n",
      "Processing image: 827/5000\n",
      "Processing image: 828/5000\n",
      "Processing image: 829/5000\n",
      "Processing image: 830/5000\n",
      "Processing image: 831/5000\n",
      "Processing image: 832/5000\n",
      "Processing image: 833/5000\n",
      "Processing image: 834/5000\n",
      "Processing image: 835/5000\n",
      "Processing image: 836/5000\n",
      "Processing image: 837/5000\n",
      "Processing image: 838/5000\n",
      "Processing image: 839/5000\n",
      "Processing image: 840/5000\n",
      "Processing image: 841/5000\n",
      "Processing image: 842/5000\n",
      "Processing image: 843/5000\n",
      "Processing image: 844/5000\n",
      "Processing image: 845/5000\n",
      "Processing image: 846/5000\n",
      "Processing image: 847/5000\n",
      "Processing image: 848/5000\n",
      "Processing image: 849/5000\n",
      "Processing image: 850/5000\n",
      "Processing image: 851/5000\n",
      "Processing image: 852/5000\n",
      "Processing image: 853/5000\n",
      "Processing image: 854/5000\n",
      "Processing image: 855/5000\n",
      "Processing image: 856/5000\n",
      "Processing image: 857/5000\n",
      "Processing image: 858/5000\n",
      "Processing image: 859/5000\n",
      "Processing image: 860/5000\n",
      "Processing image: 861/5000\n",
      "Processing image: 862/5000\n",
      "Processing image: 863/5000\n",
      "Processing image: 864/5000\n",
      "Processing image: 865/5000\n",
      "Processing image: 866/5000\n",
      "Processing image: 867/5000\n",
      "Processing image: 868/5000\n",
      "Processing image: 869/5000\n",
      "Processing image: 870/5000\n",
      "Processing image: 871/5000\n",
      "Processing image: 872/5000\n",
      "Processing image: 873/5000\n",
      "Processing image: 874/5000\n",
      "Processing image: 875/5000\n",
      "Processing image: 876/5000\n",
      "Processing image: 877/5000\n",
      "Processing image: 878/5000\n",
      "Processing image: 879/5000\n",
      "Processing image: 880/5000\n",
      "Processing image: 881/5000\n",
      "Processing image: 882/5000\n",
      "Processing image: 883/5000\n",
      "Processing image: 884/5000\n",
      "Processing image: 885/5000\n",
      "Processing image: 886/5000\n",
      "Processing image: 887/5000\n",
      "Processing image: 888/5000\n",
      "Processing image: 889/5000\n",
      "Processing image: 890/5000\n",
      "Processing image: 891/5000\n",
      "Processing image: 892/5000\n",
      "Processing image: 893/5000\n",
      "Processing image: 894/5000\n",
      "Processing image: 895/5000\n",
      "Processing image: 896/5000\n",
      "Processing image: 897/5000\n",
      "Processing image: 898/5000\n",
      "Processing image: 899/5000\n",
      "Processing image: 900/5000\n",
      "Processing image: 901/5000\n",
      "Processing image: 902/5000\n",
      "Processing image: 903/5000\n",
      "Processing image: 904/5000\n",
      "Processing image: 905/5000\n",
      "Processing image: 906/5000\n",
      "Processing image: 907/5000\n",
      "Processing image: 908/5000\n",
      "Processing image: 909/5000\n",
      "Processing image: 910/5000\n",
      "Processing image: 911/5000\n",
      "Processing image: 912/5000\n",
      "Processing image: 913/5000\n",
      "Processing image: 914/5000\n",
      "Processing image: 915/5000\n",
      "Processing image: 916/5000\n",
      "Processing image: 917/5000\n",
      "Processing image: 918/5000\n",
      "Processing image: 919/5000\n",
      "Processing image: 920/5000\n",
      "Processing image: 921/5000\n",
      "Processing image: 922/5000\n",
      "Processing image: 923/5000\n",
      "Processing image: 924/5000\n",
      "Processing image: 925/5000\n",
      "Processing image: 926/5000\n",
      "Processing image: 927/5000\n",
      "Processing image: 928/5000\n",
      "Processing image: 929/5000\n",
      "Processing image: 930/5000\n",
      "Processing image: 931/5000\n",
      "Processing image: 932/5000\n",
      "Processing image: 933/5000\n",
      "Processing image: 934/5000\n",
      "Processing image: 935/5000\n",
      "Processing image: 936/5000\n",
      "Processing image: 937/5000\n",
      "Processing image: 938/5000\n",
      "Processing image: 939/5000\n",
      "Processing image: 940/5000\n",
      "Processing image: 941/5000\n",
      "Processing image: 942/5000\n",
      "Processing image: 943/5000\n",
      "Processing image: 944/5000\n",
      "Processing image: 945/5000\n",
      "Processing image: 946/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 947/5000\n",
      "Processing image: 948/5000\n",
      "Processing image: 949/5000\n",
      "Processing image: 950/5000\n",
      "Processing image: 951/5000\n",
      "Processing image: 952/5000\n",
      "Processing image: 953/5000\n",
      "Processing image: 954/5000\n",
      "Processing image: 955/5000\n",
      "Processing image: 956/5000\n",
      "Processing image: 957/5000\n",
      "Processing image: 958/5000\n",
      "Processing image: 959/5000\n",
      "Processing image: 960/5000\n",
      "Processing image: 961/5000\n",
      "Processing image: 962/5000\n",
      "Processing image: 963/5000\n",
      "Processing image: 964/5000\n",
      "Processing image: 965/5000\n",
      "Processing image: 966/5000\n",
      "Processing image: 967/5000\n",
      "Processing image: 968/5000\n",
      "Processing image: 969/5000\n",
      "Processing image: 970/5000\n",
      "Processing image: 971/5000\n",
      "Processing image: 972/5000\n",
      "Processing image: 973/5000\n",
      "Processing image: 974/5000\n",
      "Processing image: 975/5000\n",
      "Processing image: 976/5000\n",
      "Processing image: 977/5000\n",
      "Processing image: 978/5000\n",
      "Processing image: 979/5000\n",
      "Processing image: 980/5000\n",
      "Processing image: 981/5000\n",
      "Processing image: 982/5000\n",
      "Processing image: 983/5000\n",
      "Processing image: 984/5000\n",
      "Processing image: 985/5000\n",
      "Processing image: 986/5000\n",
      "Processing image: 987/5000\n",
      "Processing image: 988/5000\n",
      "Processing image: 989/5000\n",
      "Processing image: 990/5000\n",
      "Processing image: 991/5000\n",
      "Processing image: 992/5000\n",
      "Processing image: 993/5000\n",
      "Processing image: 994/5000\n",
      "Processing image: 995/5000\n",
      "Processing image: 996/5000\n",
      "Processing image: 997/5000\n",
      "Processing image: 998/5000\n",
      "Processing image: 999/5000\n",
      "Processing image: 1000/5000\n",
      "Processing image: 1001/5000\n",
      "Processing image: 1002/5000\n",
      "Processing image: 1003/5000\n",
      "Processing image: 1004/5000\n",
      "Processing image: 1005/5000\n",
      "Processing image: 1006/5000\n",
      "Processing image: 1007/5000\n",
      "Processing image: 1008/5000\n",
      "Processing image: 1009/5000\n",
      "Processing image: 1010/5000\n",
      "Processing image: 1011/5000\n",
      "Processing image: 1012/5000\n",
      "Processing image: 1013/5000\n",
      "Processing image: 1014/5000\n",
      "Processing image: 1015/5000\n",
      "Processing image: 1016/5000\n",
      "Processing image: 1017/5000\n",
      "Processing image: 1018/5000\n",
      "Processing image: 1019/5000\n",
      "Processing image: 1020/5000\n",
      "Processing image: 1021/5000\n",
      "Processing image: 1022/5000\n",
      "Processing image: 1023/5000\n",
      "Processing image: 1024/5000\n",
      "Processing image: 1025/5000\n",
      "Processing image: 1026/5000\n",
      "Processing image: 1027/5000\n",
      "Processing image: 1028/5000\n",
      "Processing image: 1029/5000\n",
      "Processing image: 1030/5000\n",
      "Processing image: 1031/5000\n",
      "Processing image: 1032/5000\n",
      "Processing image: 1033/5000\n",
      "Processing image: 1034/5000\n",
      "Processing image: 1035/5000\n",
      "Processing image: 1036/5000\n",
      "Processing image: 1037/5000\n",
      "Processing image: 1038/5000\n",
      "Processing image: 1039/5000\n",
      "Processing image: 1040/5000\n",
      "Processing image: 1041/5000\n",
      "Processing image: 1042/5000\n",
      "Processing image: 1043/5000\n",
      "Processing image: 1044/5000\n",
      "Processing image: 1045/5000\n",
      "Processing image: 1046/5000\n",
      "Processing image: 1047/5000\n",
      "Processing image: 1048/5000\n",
      "Processing image: 1049/5000\n",
      "Processing image: 1050/5000\n",
      "Processing image: 1051/5000\n",
      "Processing image: 1052/5000\n",
      "Processing image: 1053/5000\n",
      "Processing image: 1054/5000\n",
      "Processing image: 1055/5000\n",
      "Processing image: 1056/5000\n",
      "Processing image: 1057/5000\n",
      "Processing image: 1058/5000\n",
      "Processing image: 1059/5000\n",
      "Processing image: 1060/5000\n",
      "Processing image: 1061/5000\n",
      "Processing image: 1062/5000\n",
      "Processing image: 1063/5000\n",
      "Processing image: 1064/5000\n",
      "Processing image: 1065/5000\n",
      "Processing image: 1066/5000\n",
      "Processing image: 1067/5000\n",
      "Processing image: 1068/5000\n",
      "Processing image: 1069/5000\n",
      "Processing image: 1070/5000\n",
      "Processing image: 1071/5000\n",
      "Processing image: 1072/5000\n",
      "Processing image: 1073/5000\n",
      "Processing image: 1074/5000\n",
      "Processing image: 1075/5000\n",
      "Processing image: 1076/5000\n",
      "Processing image: 1077/5000\n",
      "Processing image: 1078/5000\n",
      "Processing image: 1079/5000\n",
      "Processing image: 1080/5000\n",
      "Processing image: 1081/5000\n",
      "Processing image: 1082/5000\n",
      "Processing image: 1083/5000\n",
      "Processing image: 1084/5000\n",
      "Processing image: 1085/5000\n",
      "Processing image: 1086/5000\n",
      "Processing image: 1087/5000\n",
      "Processing image: 1088/5000\n",
      "Processing image: 1089/5000\n",
      "Processing image: 1090/5000\n",
      "Processing image: 1091/5000\n",
      "Processing image: 1092/5000\n",
      "Processing image: 1093/5000\n",
      "Processing image: 1094/5000\n",
      "Processing image: 1095/5000\n",
      "Processing image: 1096/5000\n",
      "Processing image: 1097/5000\n",
      "Processing image: 1098/5000\n",
      "Processing image: 1099/5000\n",
      "Processing image: 1100/5000\n",
      "Processing image: 1101/5000\n",
      "Processing image: 1102/5000\n",
      "Processing image: 1103/5000\n",
      "Processing image: 1104/5000\n",
      "Processing image: 1105/5000\n",
      "Processing image: 1106/5000\n",
      "Processing image: 1107/5000\n",
      "Processing image: 1108/5000\n",
      "Processing image: 1109/5000\n",
      "Processing image: 1110/5000\n",
      "Processing image: 1111/5000\n",
      "Processing image: 1112/5000\n",
      "Processing image: 1113/5000\n",
      "Processing image: 1114/5000\n",
      "Processing image: 1115/5000\n",
      "Processing image: 1116/5000\n",
      "Processing image: 1117/5000\n",
      "Processing image: 1118/5000\n",
      "Processing image: 1119/5000\n",
      "Processing image: 1120/5000\n",
      "Processing image: 1121/5000\n",
      "Processing image: 1122/5000\n",
      "Processing image: 1123/5000\n",
      "Processing image: 1124/5000\n",
      "Processing image: 1125/5000\n",
      "Processing image: 1126/5000\n",
      "Processing image: 1127/5000\n",
      "Processing image: 1128/5000\n",
      "Processing image: 1129/5000\n",
      "Processing image: 1130/5000\n",
      "Processing image: 1131/5000\n",
      "Processing image: 1132/5000\n",
      "Processing image: 1133/5000\n",
      "Processing image: 1134/5000\n",
      "Processing image: 1135/5000\n",
      "Processing image: 1136/5000\n",
      "Processing image: 1137/5000\n",
      "Processing image: 1138/5000\n",
      "Processing image: 1139/5000\n",
      "Processing image: 1140/5000\n",
      "Processing image: 1141/5000\n",
      "Processing image: 1142/5000\n",
      "Processing image: 1143/5000\n",
      "Processing image: 1144/5000\n",
      "Processing image: 1145/5000\n",
      "Processing image: 1146/5000\n",
      "Processing image: 1147/5000\n",
      "Processing image: 1148/5000\n",
      "Processing image: 1149/5000\n",
      "Processing image: 1150/5000\n",
      "Processing image: 1151/5000\n",
      "Processing image: 1152/5000\n",
      "Processing image: 1153/5000\n",
      "Processing image: 1154/5000\n",
      "Processing image: 1155/5000\n",
      "Processing image: 1156/5000\n",
      "Processing image: 1157/5000\n",
      "Processing image: 1158/5000\n",
      "Processing image: 1159/5000\n",
      "Processing image: 1160/5000\n",
      "Processing image: 1161/5000\n",
      "Processing image: 1162/5000\n",
      "Processing image: 1163/5000\n",
      "Processing image: 1164/5000\n",
      "Processing image: 1165/5000\n",
      "Processing image: 1166/5000\n",
      "Processing image: 1167/5000\n",
      "Processing image: 1168/5000\n",
      "Processing image: 1169/5000\n",
      "Processing image: 1170/5000\n",
      "Processing image: 1171/5000\n",
      "Processing image: 1172/5000\n",
      "Processing image: 1173/5000\n",
      "Processing image: 1174/5000\n",
      "Processing image: 1175/5000\n",
      "Processing image: 1176/5000\n",
      "Processing image: 1177/5000\n",
      "Processing image: 1178/5000\n",
      "Processing image: 1179/5000\n",
      "Processing image: 1180/5000\n",
      "Processing image: 1181/5000\n",
      "Processing image: 1182/5000\n",
      "Processing image: 1183/5000\n",
      "Processing image: 1184/5000\n",
      "Processing image: 1185/5000\n",
      "Processing image: 1186/5000\n",
      "Processing image: 1187/5000\n",
      "Processing image: 1188/5000\n",
      "Processing image: 1189/5000\n",
      "Processing image: 1190/5000\n",
      "Processing image: 1191/5000\n",
      "Processing image: 1192/5000\n",
      "Processing image: 1193/5000\n",
      "Processing image: 1194/5000\n",
      "Processing image: 1195/5000\n",
      "Processing image: 1196/5000\n",
      "Processing image: 1197/5000\n",
      "Processing image: 1198/5000\n",
      "Processing image: 1199/5000\n",
      "Processing image: 1200/5000\n",
      "Processing image: 1201/5000\n",
      "Processing image: 1202/5000\n",
      "Processing image: 1203/5000\n",
      "Processing image: 1204/5000\n",
      "Processing image: 1205/5000\n",
      "Processing image: 1206/5000\n",
      "Processing image: 1207/5000\n",
      "Processing image: 1208/5000\n",
      "Processing image: 1209/5000\n",
      "Processing image: 1210/5000\n",
      "Processing image: 1211/5000\n",
      "Processing image: 1212/5000\n",
      "Processing image: 1213/5000\n",
      "Processing image: 1214/5000\n",
      "Processing image: 1215/5000\n",
      "Processing image: 1216/5000\n",
      "Processing image: 1217/5000\n",
      "Processing image: 1218/5000\n",
      "Processing image: 1219/5000\n",
      "Processing image: 1220/5000\n",
      "Processing image: 1221/5000\n",
      "Processing image: 1222/5000\n",
      "Processing image: 1223/5000\n",
      "Processing image: 1224/5000\n",
      "Processing image: 1225/5000\n",
      "Processing image: 1226/5000\n",
      "Processing image: 1227/5000\n",
      "Processing image: 1228/5000\n",
      "Processing image: 1229/5000\n",
      "Processing image: 1230/5000\n",
      "Processing image: 1231/5000\n",
      "Processing image: 1232/5000\n",
      "Processing image: 1233/5000\n",
      "Processing image: 1234/5000\n",
      "Processing image: 1235/5000\n",
      "Processing image: 1236/5000\n",
      "Processing image: 1237/5000\n",
      "Processing image: 1238/5000\n",
      "Processing image: 1239/5000\n",
      "Processing image: 1240/5000\n",
      "Processing image: 1241/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1242/5000\n",
      "Processing image: 1243/5000\n",
      "Processing image: 1244/5000\n",
      "Processing image: 1245/5000\n",
      "Processing image: 1246/5000\n",
      "Processing image: 1247/5000\n",
      "Processing image: 1248/5000\n",
      "Processing image: 1249/5000\n",
      "Processing image: 1250/5000\n",
      "Processing image: 1251/5000\n",
      "Processing image: 1252/5000\n",
      "Processing image: 1253/5000\n",
      "Processing image: 1254/5000\n",
      "Processing image: 1255/5000\n",
      "Processing image: 1256/5000\n",
      "Processing image: 1257/5000\n",
      "Processing image: 1258/5000\n",
      "Processing image: 1259/5000\n",
      "Processing image: 1260/5000\n",
      "Processing image: 1261/5000\n",
      "Processing image: 1262/5000\n",
      "Processing image: 1263/5000\n",
      "Processing image: 1264/5000\n",
      "Processing image: 1265/5000\n",
      "Processing image: 1266/5000\n",
      "Processing image: 1267/5000\n",
      "Processing image: 1268/5000\n",
      "Processing image: 1269/5000\n",
      "Processing image: 1270/5000\n",
      "Processing image: 1271/5000\n",
      "Processing image: 1272/5000\n",
      "Processing image: 1273/5000\n",
      "Processing image: 1274/5000\n",
      "Processing image: 1275/5000\n",
      "Processing image: 1276/5000\n",
      "Processing image: 1277/5000\n",
      "Processing image: 1278/5000\n",
      "Processing image: 1279/5000\n",
      "Processing image: 1280/5000\n",
      "Processing image: 1281/5000\n",
      "Processing image: 1282/5000\n",
      "Processing image: 1283/5000\n",
      "Processing image: 1284/5000\n",
      "Processing image: 1285/5000\n",
      "Processing image: 1286/5000\n",
      "Processing image: 1287/5000\n",
      "Processing image: 1288/5000\n",
      "Processing image: 1289/5000\n",
      "Processing image: 1290/5000\n",
      "Processing image: 1291/5000\n",
      "Processing image: 1292/5000\n",
      "Processing image: 1293/5000\n",
      "Processing image: 1294/5000\n",
      "Processing image: 1295/5000\n",
      "Processing image: 1296/5000\n",
      "Processing image: 1297/5000\n",
      "Processing image: 1298/5000\n",
      "Processing image: 1299/5000\n",
      "Processing image: 1300/5000\n",
      "Processing image: 1301/5000\n",
      "Processing image: 1302/5000\n",
      "Processing image: 1303/5000\n",
      "Processing image: 1304/5000\n",
      "Processing image: 1305/5000\n",
      "Processing image: 1306/5000\n",
      "Processing image: 1307/5000\n",
      "Processing image: 1308/5000\n",
      "Processing image: 1309/5000\n",
      "Processing image: 1310/5000\n",
      "Processing image: 1311/5000\n",
      "Processing image: 1312/5000\n",
      "Processing image: 1313/5000\n",
      "Processing image: 1314/5000\n",
      "Processing image: 1315/5000\n",
      "Processing image: 1316/5000\n",
      "Processing image: 1317/5000\n",
      "Processing image: 1318/5000\n",
      "Processing image: 1319/5000\n",
      "Processing image: 1320/5000\n",
      "Processing image: 1321/5000\n",
      "Processing image: 1322/5000\n",
      "Processing image: 1323/5000\n",
      "Processing image: 1324/5000\n",
      "Processing image: 1325/5000\n",
      "Processing image: 1326/5000\n",
      "Processing image: 1327/5000\n",
      "Processing image: 1328/5000\n",
      "Processing image: 1329/5000\n",
      "Processing image: 1330/5000\n",
      "Processing image: 1331/5000\n",
      "Processing image: 1332/5000\n",
      "Processing image: 1333/5000\n",
      "Processing image: 1334/5000\n",
      "Processing image: 1335/5000\n",
      "Processing image: 1336/5000\n",
      "Processing image: 1337/5000\n",
      "Processing image: 1338/5000\n",
      "Processing image: 1339/5000\n",
      "Processing image: 1340/5000\n",
      "Processing image: 1341/5000\n",
      "Processing image: 1342/5000\n",
      "Processing image: 1343/5000\n",
      "Processing image: 1344/5000\n",
      "Processing image: 1345/5000\n",
      "Processing image: 1346/5000\n",
      "Processing image: 1347/5000\n",
      "Processing image: 1348/5000\n",
      "Processing image: 1349/5000\n",
      "Processing image: 1350/5000\n",
      "Processing image: 1351/5000\n",
      "Processing image: 1352/5000\n",
      "Processing image: 1353/5000\n",
      "Processing image: 1354/5000\n",
      "Processing image: 1355/5000\n",
      "Processing image: 1356/5000\n",
      "Processing image: 1357/5000\n",
      "Processing image: 1358/5000\n",
      "Processing image: 1359/5000\n",
      "Processing image: 1360/5000\n",
      "Processing image: 1361/5000\n",
      "Processing image: 1362/5000\n",
      "Processing image: 1363/5000\n",
      "Processing image: 1364/5000\n",
      "Processing image: 1365/5000\n",
      "Processing image: 1366/5000\n",
      "Processing image: 1367/5000\n",
      "Processing image: 1368/5000\n",
      "Processing image: 1369/5000\n",
      "Processing image: 1370/5000\n",
      "Processing image: 1371/5000\n",
      "Processing image: 1372/5000\n",
      "Processing image: 1373/5000\n",
      "Processing image: 1374/5000\n",
      "Processing image: 1375/5000\n",
      "Processing image: 1376/5000\n",
      "Processing image: 1377/5000\n",
      "Processing image: 1378/5000\n",
      "Processing image: 1379/5000\n",
      "Processing image: 1380/5000\n",
      "Processing image: 1381/5000\n",
      "Processing image: 1382/5000\n",
      "Processing image: 1383/5000\n",
      "Processing image: 1384/5000\n",
      "Processing image: 1385/5000\n",
      "Processing image: 1386/5000\n",
      "Processing image: 1387/5000\n",
      "Processing image: 1388/5000\n",
      "Processing image: 1389/5000\n",
      "Processing image: 1390/5000\n",
      "Processing image: 1391/5000\n",
      "Processing image: 1392/5000\n",
      "Processing image: 1393/5000\n",
      "Processing image: 1394/5000\n",
      "Processing image: 1395/5000\n",
      "Processing image: 1396/5000\n",
      "Processing image: 1397/5000\n",
      "Processing image: 1398/5000\n",
      "Processing image: 1399/5000\n",
      "Processing image: 1400/5000\n",
      "Processing image: 1401/5000\n",
      "Processing image: 1402/5000\n",
      "Processing image: 1403/5000\n",
      "Processing image: 1404/5000\n",
      "Processing image: 1405/5000\n",
      "Processing image: 1406/5000\n",
      "Processing image: 1407/5000\n",
      "Processing image: 1408/5000\n",
      "Processing image: 1409/5000\n",
      "Processing image: 1410/5000\n",
      "Processing image: 1411/5000\n",
      "Processing image: 1412/5000\n",
      "Processing image: 1413/5000\n",
      "Processing image: 1414/5000\n",
      "Processing image: 1415/5000\n",
      "Processing image: 1416/5000\n",
      "Processing image: 1417/5000\n",
      "Processing image: 1418/5000\n",
      "Processing image: 1419/5000\n",
      "Processing image: 1420/5000\n",
      "Processing image: 1421/5000\n",
      "Processing image: 1422/5000\n",
      "Processing image: 1423/5000\n",
      "Processing image: 1424/5000\n",
      "Processing image: 1425/5000\n",
      "Processing image: 1426/5000\n",
      "Processing image: 1427/5000\n",
      "Processing image: 1428/5000\n",
      "Processing image: 1429/5000\n",
      "Processing image: 1430/5000\n",
      "Processing image: 1431/5000\n",
      "Processing image: 1432/5000\n",
      "Processing image: 1433/5000\n",
      "Processing image: 1434/5000\n",
      "Processing image: 1435/5000\n",
      "Processing image: 1436/5000\n",
      "Processing image: 1437/5000\n",
      "Processing image: 1438/5000\n",
      "Processing image: 1439/5000\n",
      "Processing image: 1440/5000\n",
      "Processing image: 1441/5000\n",
      "Processing image: 1442/5000\n",
      "Processing image: 1443/5000\n",
      "Processing image: 1444/5000\n",
      "Processing image: 1445/5000\n",
      "Processing image: 1446/5000\n",
      "Processing image: 1447/5000\n",
      "Processing image: 1448/5000\n",
      "Processing image: 1449/5000\n",
      "Processing image: 1450/5000\n",
      "Processing image: 1451/5000\n",
      "Processing image: 1452/5000\n",
      "Processing image: 1453/5000\n",
      "Processing image: 1454/5000\n",
      "Processing image: 1455/5000\n",
      "Processing image: 1456/5000\n",
      "Processing image: 1457/5000\n",
      "Processing image: 1458/5000\n",
      "Processing image: 1459/5000\n",
      "Processing image: 1460/5000\n",
      "Processing image: 1461/5000\n",
      "Processing image: 1462/5000\n",
      "Processing image: 1463/5000\n",
      "Processing image: 1464/5000\n",
      "Processing image: 1465/5000\n",
      "Processing image: 1466/5000\n",
      "Processing image: 1467/5000\n",
      "Processing image: 1468/5000\n",
      "Processing image: 1469/5000\n",
      "Processing image: 1470/5000\n",
      "Processing image: 1471/5000\n",
      "Processing image: 1472/5000\n",
      "Processing image: 1473/5000\n",
      "Processing image: 1474/5000\n",
      "Processing image: 1475/5000\n",
      "Processing image: 1476/5000\n",
      "Processing image: 1477/5000\n",
      "Processing image: 1478/5000\n",
      "Processing image: 1479/5000\n",
      "Processing image: 1480/5000\n",
      "Processing image: 1481/5000\n",
      "Processing image: 1482/5000\n",
      "Processing image: 1483/5000\n",
      "Processing image: 1484/5000\n",
      "Processing image: 1485/5000\n",
      "Processing image: 1486/5000\n",
      "Processing image: 1487/5000\n",
      "Processing image: 1488/5000\n",
      "Processing image: 1489/5000\n",
      "Processing image: 1490/5000\n",
      "Processing image: 1491/5000\n",
      "Processing image: 1492/5000\n",
      "Processing image: 1493/5000\n",
      "Processing image: 1494/5000\n",
      "Processing image: 1495/5000\n",
      "Processing image: 1496/5000\n",
      "Processing image: 1497/5000\n",
      "Processing image: 1498/5000\n",
      "Processing image: 1499/5000\n",
      "Processing image: 1500/5000\n",
      "Processing image: 1501/5000\n",
      "Processing image: 1502/5000\n",
      "Processing image: 1503/5000\n",
      "Processing image: 1504/5000\n",
      "Processing image: 1505/5000\n",
      "Processing image: 1506/5000\n",
      "Processing image: 1507/5000\n",
      "Processing image: 1508/5000\n",
      "Processing image: 1509/5000\n",
      "Processing image: 1510/5000\n",
      "Processing image: 1511/5000\n",
      "Processing image: 1512/5000\n",
      "Processing image: 1513/5000\n",
      "Processing image: 1514/5000\n",
      "Processing image: 1515/5000\n",
      "Processing image: 1516/5000\n",
      "Processing image: 1517/5000\n",
      "Processing image: 1518/5000\n",
      "Processing image: 1519/5000\n",
      "Processing image: 1520/5000\n",
      "Processing image: 1521/5000\n",
      "Processing image: 1522/5000\n",
      "Processing image: 1523/5000\n",
      "Processing image: 1524/5000\n",
      "Processing image: 1525/5000\n",
      "Processing image: 1526/5000\n",
      "Processing image: 1527/5000\n",
      "Processing image: 1528/5000\n",
      "Processing image: 1529/5000\n",
      "Processing image: 1530/5000\n",
      "Processing image: 1531/5000\n",
      "Processing image: 1532/5000\n",
      "Processing image: 1533/5000\n",
      "Processing image: 1534/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1535/5000\n",
      "Processing image: 1536/5000\n",
      "Processing image: 1537/5000\n",
      "Processing image: 1538/5000\n",
      "Processing image: 1539/5000\n",
      "Processing image: 1540/5000\n",
      "Processing image: 1541/5000\n",
      "Processing image: 1542/5000\n",
      "Processing image: 1543/5000\n",
      "Processing image: 1544/5000\n",
      "Processing image: 1545/5000\n",
      "Processing image: 1546/5000\n",
      "Processing image: 1547/5000\n",
      "Processing image: 1548/5000\n",
      "Processing image: 1549/5000\n",
      "Processing image: 1550/5000\n",
      "Processing image: 1551/5000\n",
      "Processing image: 1552/5000\n",
      "Processing image: 1553/5000\n",
      "Processing image: 1554/5000\n",
      "Processing image: 1555/5000\n",
      "Processing image: 1556/5000\n",
      "Processing image: 1557/5000\n",
      "Processing image: 1558/5000\n",
      "Processing image: 1559/5000\n",
      "Processing image: 1560/5000\n",
      "Processing image: 1561/5000\n",
      "Processing image: 1562/5000\n",
      "Processing image: 1563/5000\n",
      "Processing image: 1564/5000\n",
      "Processing image: 1565/5000\n",
      "Processing image: 1566/5000\n",
      "Processing image: 1567/5000\n",
      "Processing image: 1568/5000\n",
      "Processing image: 1569/5000\n",
      "Processing image: 1570/5000\n",
      "Processing image: 1571/5000\n",
      "Processing image: 1572/5000\n",
      "Processing image: 1573/5000\n",
      "Processing image: 1574/5000\n",
      "Processing image: 1575/5000\n",
      "Processing image: 1576/5000\n",
      "Processing image: 1577/5000\n",
      "Processing image: 1578/5000\n",
      "Processing image: 1579/5000\n",
      "Processing image: 1580/5000\n",
      "Processing image: 1581/5000\n",
      "Processing image: 1582/5000\n",
      "Processing image: 1583/5000\n",
      "Processing image: 1584/5000\n",
      "Processing image: 1585/5000\n",
      "Processing image: 1586/5000\n",
      "Processing image: 1587/5000\n",
      "Processing image: 1588/5000\n",
      "Processing image: 1589/5000\n",
      "Processing image: 1590/5000\n",
      "Processing image: 1591/5000\n",
      "Processing image: 1592/5000\n",
      "Processing image: 1593/5000\n",
      "Processing image: 1594/5000\n",
      "Processing image: 1595/5000\n",
      "Processing image: 1596/5000\n",
      "Processing image: 1597/5000\n",
      "Processing image: 1598/5000\n",
      "Processing image: 1599/5000\n",
      "Processing image: 1600/5000\n",
      "Processing image: 1601/5000\n",
      "Processing image: 1602/5000\n",
      "Processing image: 1603/5000\n",
      "Processing image: 1604/5000\n",
      "Processing image: 1605/5000\n",
      "Processing image: 1606/5000\n",
      "Processing image: 1607/5000\n",
      "Processing image: 1608/5000\n",
      "Processing image: 1609/5000\n",
      "Processing image: 1610/5000\n",
      "Processing image: 1611/5000\n",
      "Processing image: 1612/5000\n",
      "Processing image: 1613/5000\n",
      "Processing image: 1614/5000\n",
      "Processing image: 1615/5000\n",
      "Processing image: 1616/5000\n",
      "Processing image: 1617/5000\n",
      "Processing image: 1618/5000\n",
      "Processing image: 1619/5000\n",
      "Processing image: 1620/5000\n",
      "Processing image: 1621/5000\n",
      "Processing image: 1622/5000\n",
      "Processing image: 1623/5000\n",
      "Processing image: 1624/5000\n",
      "Processing image: 1625/5000\n",
      "Processing image: 1626/5000\n",
      "Processing image: 1627/5000\n",
      "Processing image: 1628/5000\n",
      "Processing image: 1629/5000\n",
      "Processing image: 1630/5000\n",
      "Processing image: 1631/5000\n",
      "Processing image: 1632/5000\n",
      "Processing image: 1633/5000\n",
      "Processing image: 1634/5000\n",
      "Processing image: 1635/5000\n",
      "Processing image: 1636/5000\n",
      "Processing image: 1637/5000\n",
      "Processing image: 1638/5000\n",
      "Processing image: 1639/5000\n",
      "Processing image: 1640/5000\n",
      "Processing image: 1641/5000\n",
      "Processing image: 1642/5000\n",
      "Processing image: 1643/5000\n",
      "Processing image: 1644/5000\n",
      "Processing image: 1645/5000\n",
      "Processing image: 1646/5000\n",
      "Processing image: 1647/5000\n",
      "Processing image: 1648/5000\n",
      "Processing image: 1649/5000\n",
      "Processing image: 1650/5000\n",
      "Processing image: 1651/5000\n",
      "Processing image: 1652/5000\n",
      "Processing image: 1653/5000\n",
      "Processing image: 1654/5000\n",
      "Processing image: 1655/5000\n",
      "Processing image: 1656/5000\n",
      "Processing image: 1657/5000\n",
      "Processing image: 1658/5000\n",
      "Processing image: 1659/5000\n",
      "Processing image: 1660/5000\n",
      "Processing image: 1661/5000\n",
      "Processing image: 1662/5000\n",
      "Processing image: 1663/5000\n",
      "Processing image: 1664/5000\n",
      "Processing image: 1665/5000\n",
      "Processing image: 1666/5000\n",
      "Processing image: 1667/5000\n",
      "Processing image: 1668/5000\n",
      "Processing image: 1669/5000\n",
      "Processing image: 1670/5000\n",
      "Processing image: 1671/5000\n",
      "Processing image: 1672/5000\n",
      "Processing image: 1673/5000\n",
      "Processing image: 1674/5000\n",
      "Processing image: 1675/5000\n",
      "Processing image: 1676/5000\n",
      "Processing image: 1677/5000\n",
      "Processing image: 1678/5000\n",
      "Processing image: 1679/5000\n",
      "Processing image: 1680/5000\n",
      "Processing image: 1681/5000\n",
      "Processing image: 1682/5000\n",
      "Processing image: 1683/5000\n",
      "Processing image: 1684/5000\n",
      "Processing image: 1685/5000\n",
      "Processing image: 1686/5000\n",
      "Processing image: 1687/5000\n",
      "Processing image: 1688/5000\n",
      "Processing image: 1689/5000\n",
      "Processing image: 1690/5000\n",
      "Processing image: 1691/5000\n",
      "Processing image: 1692/5000\n",
      "Processing image: 1693/5000\n",
      "Processing image: 1694/5000\n",
      "Processing image: 1695/5000\n",
      "Processing image: 1696/5000\n",
      "Processing image: 1697/5000\n",
      "Processing image: 1698/5000\n",
      "Processing image: 1699/5000\n",
      "Processing image: 1700/5000\n",
      "Processing image: 1701/5000\n",
      "Processing image: 1702/5000\n",
      "Processing image: 1703/5000\n",
      "Processing image: 1704/5000\n",
      "Processing image: 1705/5000\n",
      "Processing image: 1706/5000\n",
      "Processing image: 1707/5000\n",
      "Processing image: 1708/5000\n",
      "Processing image: 1709/5000\n",
      "Processing image: 1710/5000\n",
      "Processing image: 1711/5000\n",
      "Processing image: 1712/5000\n",
      "Processing image: 1713/5000\n",
      "Processing image: 1714/5000\n",
      "Processing image: 1715/5000\n",
      "Processing image: 1716/5000\n",
      "Processing image: 1717/5000\n",
      "Processing image: 1718/5000\n",
      "Processing image: 1719/5000\n",
      "Processing image: 1720/5000\n",
      "Processing image: 1721/5000\n",
      "Processing image: 1722/5000\n",
      "Processing image: 1723/5000\n",
      "Processing image: 1724/5000\n",
      "Processing image: 1725/5000\n",
      "Processing image: 1726/5000\n",
      "Processing image: 1727/5000\n",
      "Processing image: 1728/5000\n",
      "Processing image: 1729/5000\n",
      "Processing image: 1730/5000\n",
      "Processing image: 1731/5000\n",
      "Processing image: 1732/5000\n",
      "Processing image: 1733/5000\n",
      "Processing image: 1734/5000\n",
      "Processing image: 1735/5000\n",
      "Processing image: 1736/5000\n",
      "Processing image: 1737/5000\n",
      "Processing image: 1738/5000\n",
      "Processing image: 1739/5000\n",
      "Processing image: 1740/5000\n",
      "Processing image: 1741/5000\n",
      "Processing image: 1742/5000\n",
      "Processing image: 1743/5000\n",
      "Processing image: 1744/5000\n",
      "Processing image: 1745/5000\n",
      "Processing image: 1746/5000\n",
      "Processing image: 1747/5000\n",
      "Processing image: 1748/5000\n",
      "Processing image: 1749/5000\n",
      "Processing image: 1750/5000\n",
      "Processing image: 1751/5000\n",
      "Processing image: 1752/5000\n",
      "Processing image: 1753/5000\n",
      "Processing image: 1754/5000\n",
      "Processing image: 1755/5000\n",
      "Processing image: 1756/5000\n",
      "Processing image: 1757/5000\n",
      "Processing image: 1758/5000\n",
      "Processing image: 1759/5000\n",
      "Processing image: 1760/5000\n",
      "Processing image: 1761/5000\n",
      "Processing image: 1762/5000\n",
      "Processing image: 1763/5000\n",
      "Processing image: 1764/5000\n",
      "Processing image: 1765/5000\n",
      "Processing image: 1766/5000\n",
      "Processing image: 1767/5000\n",
      "Processing image: 1768/5000\n",
      "Processing image: 1769/5000\n",
      "Processing image: 1770/5000\n",
      "Processing image: 1771/5000\n",
      "Processing image: 1772/5000\n",
      "Processing image: 1773/5000\n",
      "Processing image: 1774/5000\n",
      "Processing image: 1775/5000\n",
      "Processing image: 1776/5000\n",
      "Processing image: 1777/5000\n",
      "Processing image: 1778/5000\n",
      "Processing image: 1779/5000\n",
      "Processing image: 1780/5000\n",
      "Processing image: 1781/5000\n",
      "Processing image: 1782/5000\n",
      "Processing image: 1783/5000\n",
      "Processing image: 1784/5000\n",
      "Processing image: 1785/5000\n",
      "Processing image: 1786/5000\n",
      "Processing image: 1787/5000\n",
      "Processing image: 1788/5000\n",
      "Processing image: 1789/5000\n",
      "Processing image: 1790/5000\n",
      "Processing image: 1791/5000\n",
      "Processing image: 1792/5000\n",
      "Processing image: 1793/5000\n",
      "Processing image: 1794/5000\n",
      "Processing image: 1795/5000\n",
      "Processing image: 1796/5000\n",
      "Processing image: 1797/5000\n",
      "Processing image: 1798/5000\n",
      "Processing image: 1799/5000\n",
      "Processing image: 1800/5000\n",
      "Processing image: 1801/5000\n",
      "Processing image: 1802/5000\n",
      "Processing image: 1803/5000\n",
      "Processing image: 1804/5000\n",
      "Processing image: 1805/5000\n",
      "Processing image: 1806/5000\n",
      "Processing image: 1807/5000\n",
      "Processing image: 1808/5000\n",
      "Processing image: 1809/5000\n",
      "Processing image: 1810/5000\n",
      "Processing image: 1811/5000\n",
      "Processing image: 1812/5000\n",
      "Processing image: 1813/5000\n",
      "Processing image: 1814/5000\n",
      "Processing image: 1815/5000\n",
      "Processing image: 1816/5000\n",
      "Processing image: 1817/5000\n",
      "Processing image: 1818/5000\n",
      "Processing image: 1819/5000\n",
      "Processing image: 1820/5000\n",
      "Processing image: 1821/5000\n",
      "Processing image: 1822/5000\n",
      "Processing image: 1823/5000\n",
      "Processing image: 1824/5000\n",
      "Processing image: 1825/5000\n",
      "Processing image: 1826/5000\n",
      "Processing image: 1827/5000\n",
      "Processing image: 1828/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 1829/5000\n",
      "Processing image: 1830/5000\n",
      "Processing image: 1831/5000\n",
      "Processing image: 1832/5000\n",
      "Processing image: 1833/5000\n",
      "Processing image: 1834/5000\n",
      "Processing image: 1835/5000\n",
      "Processing image: 1836/5000\n",
      "Processing image: 1837/5000\n",
      "Processing image: 1838/5000\n",
      "Processing image: 1839/5000\n",
      "Processing image: 1840/5000\n",
      "Processing image: 1841/5000\n",
      "Processing image: 1842/5000\n",
      "Processing image: 1843/5000\n",
      "Processing image: 1844/5000\n",
      "Processing image: 1845/5000\n",
      "Processing image: 1846/5000\n",
      "Processing image: 1847/5000\n",
      "Processing image: 1848/5000\n",
      "Processing image: 1849/5000\n",
      "Processing image: 1850/5000\n",
      "Processing image: 1851/5000\n",
      "Processing image: 1852/5000\n",
      "Processing image: 1853/5000\n",
      "Processing image: 1854/5000\n",
      "Processing image: 1855/5000\n",
      "Processing image: 1856/5000\n",
      "Processing image: 1857/5000\n",
      "Processing image: 1858/5000\n",
      "Processing image: 1859/5000\n",
      "Processing image: 1860/5000\n",
      "Processing image: 1861/5000\n",
      "Processing image: 1862/5000\n",
      "Processing image: 1863/5000\n",
      "Processing image: 1864/5000\n",
      "Processing image: 1865/5000\n",
      "Processing image: 1866/5000\n",
      "Processing image: 1867/5000\n",
      "Processing image: 1868/5000\n",
      "Processing image: 1869/5000\n",
      "Processing image: 1870/5000\n",
      "Processing image: 1871/5000\n",
      "Processing image: 1872/5000\n",
      "Processing image: 1873/5000\n",
      "Processing image: 1874/5000\n",
      "Processing image: 1875/5000\n",
      "Processing image: 1876/5000\n",
      "Processing image: 1877/5000\n",
      "Processing image: 1878/5000\n",
      "Processing image: 1879/5000\n",
      "Processing image: 1880/5000\n",
      "Processing image: 1881/5000\n",
      "Processing image: 1882/5000\n",
      "Processing image: 1883/5000\n",
      "Processing image: 1884/5000\n",
      "Processing image: 1885/5000\n",
      "Processing image: 1886/5000\n",
      "Processing image: 1887/5000\n",
      "Processing image: 1888/5000\n",
      "Processing image: 1889/5000\n",
      "Processing image: 1890/5000\n",
      "Processing image: 1891/5000\n",
      "Processing image: 1892/5000\n",
      "Processing image: 1893/5000\n",
      "Processing image: 1894/5000\n",
      "Processing image: 1895/5000\n",
      "Processing image: 1896/5000\n",
      "Processing image: 1897/5000\n",
      "Processing image: 1898/5000\n",
      "Processing image: 1899/5000\n",
      "Processing image: 1900/5000\n",
      "Processing image: 1901/5000\n",
      "Processing image: 1902/5000\n",
      "Processing image: 1903/5000\n",
      "Processing image: 1904/5000\n",
      "Processing image: 1905/5000\n",
      "Processing image: 1906/5000\n",
      "Processing image: 1907/5000\n",
      "Processing image: 1908/5000\n",
      "Processing image: 1909/5000\n",
      "Processing image: 1910/5000\n",
      "Processing image: 1911/5000\n",
      "Processing image: 1912/5000\n",
      "Processing image: 1913/5000\n",
      "Processing image: 1914/5000\n",
      "Processing image: 1915/5000\n",
      "Processing image: 1916/5000\n",
      "Processing image: 1917/5000\n",
      "Processing image: 1918/5000\n",
      "Processing image: 1919/5000\n",
      "Processing image: 1920/5000\n",
      "Processing image: 1921/5000\n",
      "Processing image: 1922/5000\n",
      "Processing image: 1923/5000\n",
      "Processing image: 1924/5000\n",
      "Processing image: 1925/5000\n",
      "Processing image: 1926/5000\n",
      "Processing image: 1927/5000\n",
      "Processing image: 1928/5000\n",
      "Processing image: 1929/5000\n",
      "Processing image: 1930/5000\n",
      "Processing image: 1931/5000\n",
      "Processing image: 1932/5000\n",
      "Processing image: 1933/5000\n",
      "Processing image: 1934/5000\n",
      "Processing image: 1935/5000\n",
      "Processing image: 1936/5000\n",
      "Processing image: 1937/5000\n",
      "Processing image: 1938/5000\n",
      "Processing image: 1939/5000\n",
      "Processing image: 1940/5000\n",
      "Processing image: 1941/5000\n",
      "Processing image: 1942/5000\n",
      "Processing image: 1943/5000\n",
      "Processing image: 1944/5000\n",
      "Processing image: 1945/5000\n",
      "Processing image: 1946/5000\n",
      "Processing image: 1947/5000\n",
      "Processing image: 1948/5000\n",
      "Processing image: 1949/5000\n",
      "Processing image: 1950/5000\n",
      "Processing image: 1951/5000\n",
      "Processing image: 1952/5000\n",
      "Processing image: 1953/5000\n",
      "Processing image: 1954/5000\n",
      "Processing image: 1955/5000\n",
      "Processing image: 1956/5000\n",
      "Processing image: 1957/5000\n",
      "Processing image: 1958/5000\n",
      "Processing image: 1959/5000\n",
      "Processing image: 1960/5000\n",
      "Processing image: 1961/5000\n",
      "Processing image: 1962/5000\n",
      "Processing image: 1963/5000\n",
      "Processing image: 1964/5000\n",
      "Processing image: 1965/5000\n",
      "Processing image: 1966/5000\n",
      "Processing image: 1967/5000\n",
      "Processing image: 1968/5000\n",
      "Processing image: 1969/5000\n",
      "Processing image: 1970/5000\n",
      "Processing image: 1971/5000\n",
      "Processing image: 1972/5000\n",
      "Processing image: 1973/5000\n",
      "Processing image: 1974/5000\n",
      "Processing image: 1975/5000\n",
      "Processing image: 1976/5000\n",
      "Processing image: 1977/5000\n",
      "Processing image: 1978/5000\n",
      "Processing image: 1979/5000\n",
      "Processing image: 1980/5000\n",
      "Processing image: 1981/5000\n",
      "Processing image: 1982/5000\n",
      "Processing image: 1983/5000\n",
      "Processing image: 1984/5000\n",
      "Processing image: 1985/5000\n",
      "Processing image: 1986/5000\n",
      "Processing image: 1987/5000\n",
      "Processing image: 1988/5000\n",
      "Processing image: 1989/5000\n",
      "Processing image: 1990/5000\n",
      "Processing image: 1991/5000\n",
      "Processing image: 1992/5000\n",
      "Processing image: 1993/5000\n",
      "Processing image: 1994/5000\n",
      "Processing image: 1995/5000\n",
      "Processing image: 1996/5000\n",
      "Processing image: 1997/5000\n",
      "Processing image: 1998/5000\n",
      "Processing image: 1999/5000\n",
      "Processing image: 2000/5000\n",
      "Processing image: 2001/5000\n",
      "Processing image: 2002/5000\n",
      "Processing image: 2003/5000\n",
      "Processing image: 2004/5000\n",
      "Processing image: 2005/5000\n",
      "Processing image: 2006/5000\n",
      "Processing image: 2007/5000\n",
      "Processing image: 2008/5000\n",
      "Processing image: 2009/5000\n",
      "Processing image: 2010/5000\n",
      "Processing image: 2011/5000\n",
      "Processing image: 2012/5000\n",
      "Processing image: 2013/5000\n",
      "Processing image: 2014/5000\n",
      "Processing image: 2015/5000\n",
      "Processing image: 2016/5000\n",
      "Processing image: 2017/5000\n",
      "Processing image: 2018/5000\n",
      "Processing image: 2019/5000\n",
      "Processing image: 2020/5000\n",
      "Processing image: 2021/5000\n",
      "Processing image: 2022/5000\n",
      "Processing image: 2023/5000\n",
      "Processing image: 2024/5000\n",
      "Processing image: 2025/5000\n",
      "Processing image: 2026/5000\n",
      "Processing image: 2027/5000\n",
      "Processing image: 2028/5000\n",
      "Processing image: 2029/5000\n",
      "Processing image: 2030/5000\n",
      "Processing image: 2031/5000\n",
      "Processing image: 2032/5000\n",
      "Processing image: 2033/5000\n",
      "Processing image: 2034/5000\n",
      "Processing image: 2035/5000\n",
      "Processing image: 2036/5000\n",
      "Processing image: 2037/5000\n",
      "Processing image: 2038/5000\n",
      "Processing image: 2039/5000\n",
      "Processing image: 2040/5000\n",
      "Processing image: 2041/5000\n",
      "Processing image: 2042/5000\n",
      "Processing image: 2043/5000\n",
      "Processing image: 2044/5000\n",
      "Processing image: 2045/5000\n",
      "Processing image: 2046/5000\n",
      "Processing image: 2047/5000\n",
      "Processing image: 2048/5000\n",
      "Processing image: 2049/5000\n",
      "Processing image: 2050/5000\n",
      "Processing image: 2051/5000\n",
      "Processing image: 2052/5000\n",
      "Processing image: 2053/5000\n",
      "Processing image: 2054/5000\n",
      "Processing image: 2055/5000\n",
      "Processing image: 2056/5000\n",
      "Processing image: 2057/5000\n",
      "Processing image: 2058/5000\n",
      "Processing image: 2059/5000\n",
      "Processing image: 2060/5000\n",
      "Processing image: 2061/5000\n",
      "Processing image: 2062/5000\n",
      "Processing image: 2063/5000\n",
      "Processing image: 2064/5000\n",
      "Processing image: 2065/5000\n",
      "Processing image: 2066/5000\n",
      "Processing image: 2067/5000\n",
      "Processing image: 2068/5000\n",
      "Processing image: 2069/5000\n",
      "Processing image: 2070/5000\n",
      "Processing image: 2071/5000\n",
      "Processing image: 2072/5000\n",
      "Processing image: 2073/5000\n",
      "Processing image: 2074/5000\n",
      "Processing image: 2075/5000\n",
      "Processing image: 2076/5000\n",
      "Processing image: 2077/5000\n",
      "Processing image: 2078/5000\n",
      "Processing image: 2079/5000\n",
      "Processing image: 2080/5000\n",
      "Processing image: 2081/5000\n",
      "Processing image: 2082/5000\n",
      "Processing image: 2083/5000\n",
      "Processing image: 2084/5000\n",
      "Processing image: 2085/5000\n",
      "Processing image: 2086/5000\n",
      "Processing image: 2087/5000\n",
      "Processing image: 2088/5000\n",
      "Processing image: 2089/5000\n",
      "Processing image: 2090/5000\n",
      "Processing image: 2091/5000\n",
      "Processing image: 2092/5000\n",
      "Processing image: 2093/5000\n",
      "Processing image: 2094/5000\n",
      "Processing image: 2095/5000\n",
      "Processing image: 2096/5000\n",
      "Processing image: 2097/5000\n",
      "Processing image: 2098/5000\n",
      "Processing image: 2099/5000\n",
      "Processing image: 2100/5000\n",
      "Processing image: 2101/5000\n",
      "Processing image: 2102/5000\n",
      "Processing image: 2103/5000\n",
      "Processing image: 2104/5000\n",
      "Processing image: 2105/5000\n",
      "Processing image: 2106/5000\n",
      "Processing image: 2107/5000\n",
      "Processing image: 2108/5000\n",
      "Processing image: 2109/5000\n",
      "Processing image: 2110/5000\n",
      "Processing image: 2111/5000\n",
      "Processing image: 2112/5000\n",
      "Processing image: 2113/5000\n",
      "Processing image: 2114/5000\n",
      "Processing image: 2115/5000\n",
      "Processing image: 2116/5000\n",
      "Processing image: 2117/5000\n",
      "Processing image: 2118/5000\n",
      "Processing image: 2119/5000\n",
      "Processing image: 2120/5000\n",
      "Processing image: 2121/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 2122/5000\n",
      "Processing image: 2123/5000\n",
      "Processing image: 2124/5000\n",
      "Processing image: 2125/5000\n",
      "Processing image: 2126/5000\n",
      "Processing image: 2127/5000\n",
      "Processing image: 2128/5000\n",
      "Processing image: 2129/5000\n",
      "Processing image: 2130/5000\n",
      "Processing image: 2131/5000\n",
      "Processing image: 2132/5000\n",
      "Processing image: 2133/5000\n",
      "Processing image: 2134/5000\n",
      "Processing image: 2135/5000\n",
      "Processing image: 2136/5000\n",
      "Processing image: 2137/5000\n",
      "Processing image: 2138/5000\n",
      "Processing image: 2139/5000\n",
      "Processing image: 2140/5000\n",
      "Processing image: 2141/5000\n",
      "Processing image: 2142/5000\n",
      "Processing image: 2143/5000\n",
      "Processing image: 2144/5000\n",
      "Processing image: 2145/5000\n",
      "Processing image: 2146/5000\n",
      "Processing image: 2147/5000\n",
      "Processing image: 2148/5000\n",
      "Processing image: 2149/5000\n",
      "Processing image: 2150/5000\n",
      "Processing image: 2151/5000\n",
      "Processing image: 2152/5000\n",
      "Processing image: 2153/5000\n",
      "Processing image: 2154/5000\n",
      "Processing image: 2155/5000\n",
      "Processing image: 2156/5000\n",
      "Processing image: 2157/5000\n",
      "Processing image: 2158/5000\n",
      "Processing image: 2159/5000\n",
      "Processing image: 2160/5000\n",
      "Processing image: 2161/5000\n",
      "Processing image: 2162/5000\n",
      "Processing image: 2163/5000\n",
      "Processing image: 2164/5000\n",
      "Processing image: 2165/5000\n",
      "Processing image: 2166/5000\n",
      "Processing image: 2167/5000\n",
      "Processing image: 2168/5000\n",
      "Processing image: 2169/5000\n",
      "Processing image: 2170/5000\n",
      "Processing image: 2171/5000\n",
      "Processing image: 2172/5000\n",
      "Processing image: 2173/5000\n",
      "Processing image: 2174/5000\n",
      "Processing image: 2175/5000\n",
      "Processing image: 2176/5000\n",
      "Processing image: 2177/5000\n",
      "Processing image: 2178/5000\n",
      "Processing image: 2179/5000\n",
      "Processing image: 2180/5000\n",
      "Processing image: 2181/5000\n",
      "Processing image: 2182/5000\n",
      "Processing image: 2183/5000\n",
      "Processing image: 2184/5000\n",
      "Processing image: 2185/5000\n",
      "Processing image: 2186/5000\n",
      "Processing image: 2187/5000\n",
      "Processing image: 2188/5000\n",
      "Processing image: 2189/5000\n",
      "Processing image: 2190/5000\n",
      "Processing image: 2191/5000\n",
      "Processing image: 2192/5000\n",
      "Processing image: 2193/5000\n",
      "Processing image: 2194/5000\n",
      "Processing image: 2195/5000\n",
      "Processing image: 2196/5000\n",
      "Processing image: 2197/5000\n",
      "Processing image: 2198/5000\n",
      "Processing image: 2199/5000\n",
      "Processing image: 2200/5000\n",
      "Processing image: 2201/5000\n",
      "Processing image: 2202/5000\n",
      "Processing image: 2203/5000\n",
      "Processing image: 2204/5000\n",
      "Processing image: 2205/5000\n",
      "Processing image: 2206/5000\n",
      "Processing image: 2207/5000\n",
      "Processing image: 2208/5000\n",
      "Processing image: 2209/5000\n",
      "Processing image: 2210/5000\n",
      "Processing image: 2211/5000\n",
      "Processing image: 2212/5000\n",
      "Processing image: 2213/5000\n",
      "Processing image: 2214/5000\n",
      "Processing image: 2215/5000\n",
      "Processing image: 2216/5000\n",
      "Processing image: 2217/5000\n",
      "Processing image: 2218/5000\n",
      "Processing image: 2219/5000\n",
      "Processing image: 2220/5000\n",
      "Processing image: 2221/5000\n",
      "Processing image: 2222/5000\n",
      "Processing image: 2223/5000\n",
      "Processing image: 2224/5000\n",
      "Processing image: 2225/5000\n",
      "Processing image: 2226/5000\n",
      "Processing image: 2227/5000\n",
      "Processing image: 2228/5000\n",
      "Processing image: 2229/5000\n",
      "Processing image: 2230/5000\n",
      "Processing image: 2231/5000\n",
      "Processing image: 2232/5000\n",
      "Processing image: 2233/5000\n",
      "Processing image: 2234/5000\n",
      "Processing image: 2235/5000\n",
      "Processing image: 2236/5000\n",
      "Processing image: 2237/5000\n",
      "Processing image: 2238/5000\n",
      "Processing image: 2239/5000\n",
      "Processing image: 2240/5000\n",
      "Processing image: 2241/5000\n",
      "Processing image: 2242/5000\n",
      "Processing image: 2243/5000\n",
      "Processing image: 2244/5000\n",
      "Processing image: 2245/5000\n",
      "Processing image: 2246/5000\n",
      "Processing image: 2247/5000\n",
      "Processing image: 2248/5000\n",
      "Processing image: 2249/5000\n",
      "Processing image: 2250/5000\n",
      "Processing image: 2251/5000\n",
      "Processing image: 2252/5000\n",
      "Processing image: 2253/5000\n",
      "Processing image: 2254/5000\n",
      "Processing image: 2255/5000\n",
      "Processing image: 2256/5000\n",
      "Processing image: 2257/5000\n",
      "Processing image: 2258/5000\n",
      "Processing image: 2259/5000\n",
      "Processing image: 2260/5000\n",
      "Processing image: 2261/5000\n",
      "Processing image: 2262/5000\n",
      "Processing image: 2263/5000\n",
      "Processing image: 2264/5000\n",
      "Processing image: 2265/5000\n",
      "Processing image: 2266/5000\n",
      "Processing image: 2267/5000\n",
      "Processing image: 2268/5000\n",
      "Processing image: 2269/5000\n",
      "Processing image: 2270/5000\n",
      "Processing image: 2271/5000\n",
      "Processing image: 2272/5000\n",
      "Processing image: 2273/5000\n",
      "Processing image: 2274/5000\n",
      "Processing image: 2275/5000\n",
      "Processing image: 2276/5000\n",
      "Processing image: 2277/5000\n",
      "Processing image: 2278/5000\n",
      "Processing image: 2279/5000\n",
      "Processing image: 2280/5000\n",
      "Processing image: 2281/5000\n",
      "Processing image: 2282/5000\n",
      "Processing image: 2283/5000\n",
      "Processing image: 2284/5000\n",
      "Processing image: 2285/5000\n",
      "Processing image: 2286/5000\n",
      "Processing image: 2287/5000\n",
      "Processing image: 2288/5000\n",
      "Processing image: 2289/5000\n",
      "Processing image: 2290/5000\n",
      "Processing image: 2291/5000\n",
      "Processing image: 2292/5000\n",
      "Processing image: 2293/5000\n",
      "Processing image: 2294/5000\n",
      "Processing image: 2295/5000\n",
      "Processing image: 2296/5000\n",
      "Processing image: 2297/5000\n",
      "Processing image: 2298/5000\n",
      "Processing image: 2299/5000\n",
      "Processing image: 2300/5000\n",
      "Processing image: 2301/5000\n",
      "Processing image: 2302/5000\n",
      "Processing image: 2303/5000\n",
      "Processing image: 2304/5000\n",
      "Processing image: 2305/5000\n",
      "Processing image: 2306/5000\n",
      "Processing image: 2307/5000\n",
      "Processing image: 2308/5000\n",
      "Processing image: 2309/5000\n",
      "Processing image: 2310/5000\n",
      "Processing image: 2311/5000\n",
      "Processing image: 2312/5000\n",
      "Processing image: 2313/5000\n",
      "Processing image: 2314/5000\n",
      "Processing image: 2315/5000\n",
      "Processing image: 2316/5000\n",
      "Processing image: 2317/5000\n",
      "Processing image: 2318/5000\n",
      "Processing image: 2319/5000\n",
      "Processing image: 2320/5000\n",
      "Processing image: 2321/5000\n",
      "Processing image: 2322/5000\n",
      "Processing image: 2323/5000\n",
      "Processing image: 2324/5000\n",
      "Processing image: 2325/5000\n",
      "Processing image: 2326/5000\n",
      "Processing image: 2327/5000\n",
      "Processing image: 2328/5000\n",
      "Processing image: 2329/5000\n",
      "Processing image: 2330/5000\n",
      "Processing image: 2331/5000\n",
      "Processing image: 2332/5000\n",
      "Processing image: 2333/5000\n",
      "Processing image: 2334/5000\n",
      "Processing image: 2335/5000\n",
      "Processing image: 2336/5000\n",
      "Processing image: 2337/5000\n",
      "Processing image: 2338/5000\n",
      "Processing image: 2339/5000\n",
      "Processing image: 2340/5000\n",
      "Processing image: 2341/5000\n",
      "Processing image: 2342/5000\n",
      "Processing image: 2343/5000\n",
      "Processing image: 2344/5000\n",
      "Processing image: 2345/5000\n",
      "Processing image: 2346/5000\n",
      "Processing image: 2347/5000\n",
      "Processing image: 2348/5000\n",
      "Processing image: 2349/5000\n",
      "Processing image: 2350/5000\n",
      "Processing image: 2351/5000\n",
      "Processing image: 2352/5000\n",
      "Processing image: 2353/5000\n",
      "Processing image: 2354/5000\n",
      "Processing image: 2355/5000\n",
      "Processing image: 2356/5000\n",
      "Processing image: 2357/5000\n",
      "Processing image: 2358/5000\n",
      "Processing image: 2359/5000\n",
      "Processing image: 2360/5000\n",
      "Processing image: 2361/5000\n",
      "Processing image: 2362/5000\n",
      "Processing image: 2363/5000\n",
      "Processing image: 2364/5000\n",
      "Processing image: 2365/5000\n",
      "Processing image: 2366/5000\n",
      "Processing image: 2367/5000\n",
      "Processing image: 2368/5000\n",
      "Processing image: 2369/5000\n",
      "Processing image: 2370/5000\n",
      "Processing image: 2371/5000\n",
      "Processing image: 2372/5000\n",
      "Processing image: 2373/5000\n",
      "Processing image: 2374/5000\n",
      "Processing image: 2375/5000\n",
      "Processing image: 2376/5000\n",
      "Processing image: 2377/5000\n",
      "Processing image: 2378/5000\n",
      "Processing image: 2379/5000\n",
      "Processing image: 2380/5000\n",
      "Processing image: 2381/5000\n",
      "Processing image: 2382/5000\n",
      "Processing image: 2383/5000\n",
      "Processing image: 2384/5000\n",
      "Processing image: 2385/5000\n",
      "Processing image: 2386/5000\n",
      "Processing image: 2387/5000\n",
      "Processing image: 2388/5000\n",
      "Processing image: 2389/5000\n",
      "Processing image: 2390/5000\n",
      "Processing image: 2391/5000\n",
      "Processing image: 2392/5000\n",
      "Processing image: 2393/5000\n",
      "Processing image: 2394/5000\n",
      "Processing image: 2395/5000\n",
      "Processing image: 2396/5000\n",
      "Processing image: 2397/5000\n",
      "Processing image: 2398/5000\n",
      "Processing image: 2399/5000\n",
      "Processing image: 2400/5000\n",
      "Processing image: 2401/5000\n",
      "Processing image: 2402/5000\n",
      "Processing image: 2403/5000\n",
      "Processing image: 2404/5000\n",
      "Processing image: 2405/5000\n",
      "Processing image: 2406/5000\n",
      "Processing image: 2407/5000\n",
      "Processing image: 2408/5000\n",
      "Processing image: 2409/5000\n",
      "Processing image: 2410/5000\n",
      "Processing image: 2411/5000\n",
      "Processing image: 2412/5000\n",
      "Processing image: 2413/5000\n",
      "Processing image: 2414/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 2415/5000\n",
      "Processing image: 2416/5000\n",
      "Processing image: 2417/5000\n",
      "Processing image: 2418/5000\n",
      "Processing image: 2419/5000\n",
      "Processing image: 2420/5000\n",
      "Processing image: 2421/5000\n",
      "Processing image: 2422/5000\n",
      "Processing image: 2423/5000\n",
      "Processing image: 2424/5000\n",
      "Processing image: 2425/5000\n",
      "Processing image: 2426/5000\n",
      "Processing image: 2427/5000\n",
      "Processing image: 2428/5000\n",
      "Processing image: 2429/5000\n",
      "Processing image: 2430/5000\n",
      "Processing image: 2431/5000\n",
      "Processing image: 2432/5000\n",
      "Processing image: 2433/5000\n",
      "Processing image: 2434/5000\n",
      "Processing image: 2435/5000\n",
      "Processing image: 2436/5000\n",
      "Processing image: 2437/5000\n",
      "Processing image: 2438/5000\n",
      "Processing image: 2439/5000\n",
      "Processing image: 2440/5000\n",
      "Processing image: 2441/5000\n",
      "Processing image: 2442/5000\n",
      "Processing image: 2443/5000\n",
      "Processing image: 2444/5000\n",
      "Processing image: 2445/5000\n",
      "Processing image: 2446/5000\n",
      "Processing image: 2447/5000\n",
      "Processing image: 2448/5000\n",
      "Processing image: 2449/5000\n",
      "Processing image: 2450/5000\n",
      "Processing image: 2451/5000\n",
      "Processing image: 2452/5000\n",
      "Processing image: 2453/5000\n",
      "Processing image: 2454/5000\n",
      "Processing image: 2455/5000\n",
      "Processing image: 2456/5000\n",
      "Processing image: 2457/5000\n",
      "Processing image: 2458/5000\n",
      "Processing image: 2459/5000\n",
      "Processing image: 2460/5000\n",
      "Processing image: 2461/5000\n",
      "Processing image: 2462/5000\n",
      "Processing image: 2463/5000\n",
      "Processing image: 2464/5000\n",
      "Processing image: 2465/5000\n",
      "Processing image: 2466/5000\n",
      "Processing image: 2467/5000\n",
      "Processing image: 2468/5000\n",
      "Processing image: 2469/5000\n",
      "Processing image: 2470/5000\n",
      "Processing image: 2471/5000\n",
      "Processing image: 2472/5000\n",
      "Processing image: 2473/5000\n",
      "Processing image: 2474/5000\n",
      "Processing image: 2475/5000\n",
      "Processing image: 2476/5000\n",
      "Processing image: 2477/5000\n",
      "Processing image: 2478/5000\n",
      "Processing image: 2479/5000\n",
      "Processing image: 2480/5000\n",
      "Processing image: 2481/5000\n",
      "Processing image: 2482/5000\n",
      "Processing image: 2483/5000\n",
      "Processing image: 2484/5000\n",
      "Processing image: 2485/5000\n",
      "Processing image: 2486/5000\n",
      "Processing image: 2487/5000\n",
      "Processing image: 2488/5000\n",
      "Processing image: 2489/5000\n",
      "Processing image: 2490/5000\n",
      "Processing image: 2491/5000\n",
      "Processing image: 2492/5000\n",
      "Processing image: 2493/5000\n",
      "Processing image: 2494/5000\n",
      "Processing image: 2495/5000\n",
      "Processing image: 2496/5000\n",
      "Processing image: 2497/5000\n",
      "Processing image: 2498/5000\n",
      "Processing image: 2499/5000\n",
      "Processing image: 2500/5000\n",
      "Processing image: 2501/5000\n",
      "Processing image: 2502/5000\n",
      "Processing image: 2503/5000\n",
      "Processing image: 2504/5000\n",
      "Processing image: 2505/5000\n",
      "Processing image: 2506/5000\n",
      "Processing image: 2507/5000\n",
      "Processing image: 2508/5000\n",
      "Processing image: 2509/5000\n",
      "Processing image: 2510/5000\n",
      "Processing image: 2511/5000\n",
      "Processing image: 2512/5000\n",
      "Processing image: 2513/5000\n",
      "Processing image: 2514/5000\n",
      "Processing image: 2515/5000\n",
      "Processing image: 2516/5000\n",
      "Processing image: 2517/5000\n",
      "Processing image: 2518/5000\n",
      "Processing image: 2519/5000\n",
      "Processing image: 2520/5000\n",
      "Processing image: 2521/5000\n",
      "Processing image: 2522/5000\n",
      "Processing image: 2523/5000\n",
      "Processing image: 2524/5000\n",
      "Processing image: 2525/5000\n",
      "Processing image: 2526/5000\n",
      "Processing image: 2527/5000\n",
      "Processing image: 2528/5000\n",
      "Processing image: 2529/5000\n",
      "Processing image: 2530/5000\n",
      "Processing image: 2531/5000\n",
      "Processing image: 2532/5000\n",
      "Processing image: 2533/5000\n",
      "Processing image: 2534/5000\n",
      "Processing image: 2535/5000\n",
      "Processing image: 2536/5000\n",
      "Processing image: 2537/5000\n",
      "Processing image: 2538/5000\n",
      "Processing image: 2539/5000\n",
      "Processing image: 2540/5000\n",
      "Processing image: 2541/5000\n",
      "Processing image: 2542/5000\n",
      "Processing image: 2543/5000\n",
      "Processing image: 2544/5000\n",
      "Processing image: 2545/5000\n",
      "Processing image: 2546/5000\n",
      "Processing image: 2547/5000\n",
      "Processing image: 2548/5000\n",
      "Processing image: 2549/5000\n",
      "Processing image: 2550/5000\n",
      "Processing image: 2551/5000\n",
      "Processing image: 2552/5000\n",
      "Processing image: 2553/5000\n",
      "Processing image: 2554/5000\n",
      "Processing image: 2555/5000\n",
      "Processing image: 2556/5000\n",
      "Processing image: 2557/5000\n",
      "Processing image: 2558/5000\n",
      "Processing image: 2559/5000\n",
      "Processing image: 2560/5000\n",
      "Processing image: 2561/5000\n",
      "Processing image: 2562/5000\n",
      "Processing image: 2563/5000\n",
      "Processing image: 2564/5000\n",
      "Processing image: 2565/5000\n",
      "Processing image: 2566/5000\n",
      "Processing image: 2567/5000\n",
      "Processing image: 2568/5000\n",
      "Processing image: 2569/5000\n",
      "Processing image: 2570/5000\n",
      "Processing image: 2571/5000\n",
      "Processing image: 2572/5000\n",
      "Processing image: 2573/5000\n",
      "Processing image: 2574/5000\n",
      "Processing image: 2575/5000\n",
      "Processing image: 2576/5000\n",
      "Processing image: 2577/5000\n",
      "Processing image: 2578/5000\n",
      "Processing image: 2579/5000\n",
      "Processing image: 2580/5000\n",
      "Processing image: 2581/5000\n",
      "Processing image: 2582/5000\n",
      "Processing image: 2583/5000\n",
      "Processing image: 2584/5000\n",
      "Processing image: 2585/5000\n",
      "Processing image: 2586/5000\n",
      "Processing image: 2587/5000\n",
      "Processing image: 2588/5000\n",
      "Processing image: 2589/5000\n",
      "Processing image: 2590/5000\n",
      "Processing image: 2591/5000\n",
      "Processing image: 2592/5000\n",
      "Processing image: 2593/5000\n",
      "Processing image: 2594/5000\n",
      "Processing image: 2595/5000\n",
      "Processing image: 2596/5000\n",
      "Processing image: 2597/5000\n",
      "Processing image: 2598/5000\n",
      "Processing image: 2599/5000\n",
      "Processing image: 2600/5000\n",
      "Processing image: 2601/5000\n",
      "Processing image: 2602/5000\n",
      "Processing image: 2603/5000\n",
      "Processing image: 2604/5000\n",
      "Processing image: 2605/5000\n",
      "Processing image: 2606/5000\n",
      "Processing image: 2607/5000\n",
      "Processing image: 2608/5000\n",
      "Processing image: 2609/5000\n",
      "Processing image: 2610/5000\n",
      "Processing image: 2611/5000\n",
      "Processing image: 2612/5000\n",
      "Processing image: 2613/5000\n",
      "Processing image: 2614/5000\n",
      "Processing image: 2615/5000\n",
      "Processing image: 2616/5000\n",
      "Processing image: 2617/5000\n",
      "Processing image: 2618/5000\n",
      "Processing image: 2619/5000\n",
      "Processing image: 2620/5000\n",
      "Processing image: 2621/5000\n",
      "Processing image: 2622/5000\n",
      "Processing image: 2623/5000\n",
      "Processing image: 2624/5000\n",
      "Processing image: 2625/5000\n",
      "Processing image: 2626/5000\n",
      "Processing image: 2627/5000\n",
      "Processing image: 2628/5000\n",
      "Processing image: 2629/5000\n",
      "Processing image: 2630/5000\n",
      "Processing image: 2631/5000\n",
      "Processing image: 2632/5000\n",
      "Processing image: 2633/5000\n",
      "Processing image: 2634/5000\n",
      "Processing image: 2635/5000\n",
      "Processing image: 2636/5000\n",
      "Processing image: 2637/5000\n",
      "Processing image: 2638/5000\n",
      "Processing image: 2639/5000\n",
      "Processing image: 2640/5000\n",
      "Processing image: 2641/5000\n",
      "Processing image: 2642/5000\n",
      "Processing image: 2643/5000\n",
      "Processing image: 2644/5000\n",
      "Processing image: 2645/5000\n",
      "Processing image: 2646/5000\n",
      "Processing image: 2647/5000\n",
      "Processing image: 2648/5000\n",
      "Processing image: 2649/5000\n",
      "Processing image: 2650/5000\n",
      "Processing image: 2651/5000\n",
      "Processing image: 2652/5000\n",
      "Processing image: 2653/5000\n",
      "Processing image: 2654/5000\n",
      "Processing image: 2655/5000\n",
      "Processing image: 2656/5000\n",
      "Processing image: 2657/5000\n",
      "Processing image: 2658/5000\n",
      "Processing image: 2659/5000\n",
      "Processing image: 2660/5000\n",
      "Processing image: 2661/5000\n",
      "Processing image: 2662/5000\n",
      "Processing image: 2663/5000\n",
      "Processing image: 2664/5000\n",
      "Processing image: 2665/5000\n",
      "Processing image: 2666/5000\n",
      "Processing image: 2667/5000\n",
      "Processing image: 2668/5000\n",
      "Processing image: 2669/5000\n",
      "Processing image: 2670/5000\n",
      "Processing image: 2671/5000\n",
      "Processing image: 2672/5000\n",
      "Processing image: 2673/5000\n",
      "Processing image: 2674/5000\n",
      "Processing image: 2675/5000\n",
      "Processing image: 2676/5000\n",
      "Processing image: 2677/5000\n",
      "Processing image: 2678/5000\n",
      "Processing image: 2679/5000\n",
      "Processing image: 2680/5000\n",
      "Processing image: 2681/5000\n",
      "Processing image: 2682/5000\n",
      "Processing image: 2683/5000\n",
      "Processing image: 2684/5000\n",
      "Processing image: 2685/5000\n",
      "Processing image: 2686/5000\n",
      "Processing image: 2687/5000\n",
      "Processing image: 2688/5000\n",
      "Processing image: 2689/5000\n",
      "Processing image: 2690/5000\n",
      "Processing image: 2691/5000\n",
      "Processing image: 2692/5000\n",
      "Processing image: 2693/5000\n",
      "Processing image: 2694/5000\n",
      "Processing image: 2695/5000\n",
      "Processing image: 2696/5000\n",
      "Processing image: 2697/5000\n",
      "Processing image: 2698/5000\n",
      "Processing image: 2699/5000\n",
      "Processing image: 2700/5000\n",
      "Processing image: 2701/5000\n",
      "Processing image: 2702/5000\n",
      "Processing image: 2703/5000\n",
      "Processing image: 2704/5000\n",
      "Processing image: 2705/5000\n",
      "Processing image: 2706/5000\n",
      "Processing image: 2707/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 2708/5000\n",
      "Processing image: 2709/5000\n",
      "Processing image: 2710/5000\n",
      "Processing image: 2711/5000\n",
      "Processing image: 2712/5000\n",
      "Processing image: 2713/5000\n",
      "Processing image: 2714/5000\n",
      "Processing image: 2715/5000\n",
      "Processing image: 2716/5000\n",
      "Processing image: 2717/5000\n",
      "Processing image: 2718/5000\n",
      "Processing image: 2719/5000\n",
      "Processing image: 2720/5000\n",
      "Processing image: 2721/5000\n",
      "Processing image: 2722/5000\n",
      "Processing image: 2723/5000\n",
      "Processing image: 2724/5000\n",
      "Processing image: 2725/5000\n",
      "Processing image: 2726/5000\n",
      "Processing image: 2727/5000\n",
      "Processing image: 2728/5000\n",
      "Processing image: 2729/5000\n",
      "Processing image: 2730/5000\n",
      "Processing image: 2731/5000\n",
      "Processing image: 2732/5000\n",
      "Processing image: 2733/5000\n",
      "Processing image: 2734/5000\n",
      "Processing image: 2735/5000\n",
      "Processing image: 2736/5000\n",
      "Processing image: 2737/5000\n",
      "Processing image: 2738/5000\n",
      "Processing image: 2739/5000\n",
      "Processing image: 2740/5000\n",
      "Processing image: 2741/5000\n",
      "Processing image: 2742/5000\n",
      "Processing image: 2743/5000\n",
      "Processing image: 2744/5000\n",
      "Processing image: 2745/5000\n",
      "Processing image: 2746/5000\n",
      "Processing image: 2747/5000\n",
      "Processing image: 2748/5000\n",
      "Processing image: 2749/5000\n",
      "Processing image: 2750/5000\n",
      "Processing image: 2751/5000\n",
      "Processing image: 2752/5000\n",
      "Processing image: 2753/5000\n",
      "Processing image: 2754/5000\n",
      "Processing image: 2755/5000\n",
      "Processing image: 2756/5000\n",
      "Processing image: 2757/5000\n",
      "Processing image: 2758/5000\n",
      "Processing image: 2759/5000\n",
      "Processing image: 2760/5000\n",
      "Processing image: 2761/5000\n",
      "Processing image: 2762/5000\n",
      "Processing image: 2763/5000\n",
      "Processing image: 2764/5000\n",
      "Processing image: 2765/5000\n",
      "Processing image: 2766/5000\n",
      "Processing image: 2767/5000\n",
      "Processing image: 2768/5000\n",
      "Processing image: 2769/5000\n",
      "Processing image: 2770/5000\n",
      "Processing image: 2771/5000\n",
      "Processing image: 2772/5000\n",
      "Processing image: 2773/5000\n",
      "Processing image: 2774/5000\n",
      "Processing image: 2775/5000\n",
      "Processing image: 2776/5000\n",
      "Processing image: 2777/5000\n",
      "Processing image: 2778/5000\n",
      "Processing image: 2779/5000\n",
      "Processing image: 2780/5000\n",
      "Processing image: 2781/5000\n",
      "Processing image: 2782/5000\n",
      "Processing image: 2783/5000\n",
      "Processing image: 2784/5000\n",
      "Processing image: 2785/5000\n",
      "Processing image: 2786/5000\n",
      "Processing image: 2787/5000\n",
      "Processing image: 2788/5000\n",
      "Processing image: 2789/5000\n",
      "Processing image: 2790/5000\n",
      "Processing image: 2791/5000\n",
      "Processing image: 2792/5000\n",
      "Processing image: 2793/5000\n",
      "Processing image: 2794/5000\n",
      "Processing image: 2795/5000\n",
      "Processing image: 2796/5000\n",
      "Processing image: 2797/5000\n",
      "Processing image: 2798/5000\n",
      "Processing image: 2799/5000\n",
      "Processing image: 2800/5000\n",
      "Processing image: 2801/5000\n",
      "Processing image: 2802/5000\n",
      "Processing image: 2803/5000\n",
      "Processing image: 2804/5000\n",
      "Processing image: 2805/5000\n",
      "Processing image: 2806/5000\n",
      "Processing image: 2807/5000\n",
      "Processing image: 2808/5000\n",
      "Processing image: 2809/5000\n",
      "Processing image: 2810/5000\n",
      "Processing image: 2811/5000\n",
      "Processing image: 2812/5000\n",
      "Processing image: 2813/5000\n",
      "Processing image: 2814/5000\n",
      "Processing image: 2815/5000\n",
      "Processing image: 2816/5000\n",
      "Processing image: 2817/5000\n",
      "Processing image: 2818/5000\n",
      "Processing image: 2819/5000\n",
      "Processing image: 2820/5000\n",
      "Processing image: 2821/5000\n",
      "Processing image: 2822/5000\n",
      "Processing image: 2823/5000\n",
      "Processing image: 2824/5000\n",
      "Processing image: 2825/5000\n",
      "Processing image: 2826/5000\n",
      "Processing image: 2827/5000\n",
      "Processing image: 2828/5000\n",
      "Processing image: 2829/5000\n",
      "Processing image: 2830/5000\n",
      "Processing image: 2831/5000\n",
      "Processing image: 2832/5000\n",
      "Processing image: 2833/5000\n",
      "Processing image: 2834/5000\n",
      "Processing image: 2835/5000\n",
      "Processing image: 2836/5000\n",
      "Processing image: 2837/5000\n",
      "Processing image: 2838/5000\n",
      "Processing image: 2839/5000\n",
      "Processing image: 2840/5000\n",
      "Processing image: 2841/5000\n",
      "Processing image: 2842/5000\n",
      "Processing image: 2843/5000\n",
      "Processing image: 2844/5000\n",
      "Processing image: 2845/5000\n",
      "Processing image: 2846/5000\n",
      "Processing image: 2847/5000\n",
      "Processing image: 2848/5000\n",
      "Processing image: 2849/5000\n",
      "Processing image: 2850/5000\n",
      "Processing image: 2851/5000\n",
      "Processing image: 2852/5000\n",
      "Processing image: 2853/5000\n",
      "Processing image: 2854/5000\n",
      "Processing image: 2855/5000\n",
      "Processing image: 2856/5000\n",
      "Processing image: 2857/5000\n",
      "Processing image: 2858/5000\n",
      "Processing image: 2859/5000\n",
      "Processing image: 2860/5000\n",
      "Processing image: 2861/5000\n",
      "Processing image: 2862/5000\n",
      "Processing image: 2863/5000\n",
      "Processing image: 2864/5000\n",
      "Processing image: 2865/5000\n",
      "Processing image: 2866/5000\n",
      "Processing image: 2867/5000\n",
      "Processing image: 2868/5000\n",
      "Processing image: 2869/5000\n",
      "Processing image: 2870/5000\n",
      "Processing image: 2871/5000\n",
      "Processing image: 2872/5000\n",
      "Processing image: 2873/5000\n",
      "Processing image: 2874/5000\n",
      "Processing image: 2875/5000\n",
      "Processing image: 2876/5000\n",
      "Processing image: 2877/5000\n",
      "Processing image: 2878/5000\n",
      "Processing image: 2879/5000\n",
      "Processing image: 2880/5000\n",
      "Processing image: 2881/5000\n",
      "Processing image: 2882/5000\n",
      "Processing image: 2883/5000\n",
      "Processing image: 2884/5000\n",
      "Processing image: 2885/5000\n",
      "Processing image: 2886/5000\n",
      "Processing image: 2887/5000\n",
      "Processing image: 2888/5000\n",
      "Processing image: 2889/5000\n",
      "Processing image: 2890/5000\n",
      "Processing image: 2891/5000\n",
      "Processing image: 2892/5000\n",
      "Processing image: 2893/5000\n",
      "Processing image: 2894/5000\n",
      "Processing image: 2895/5000\n",
      "Processing image: 2896/5000\n",
      "Processing image: 2897/5000\n",
      "Processing image: 2898/5000\n",
      "Processing image: 2899/5000\n",
      "Processing image: 2900/5000\n",
      "Processing image: 2901/5000\n",
      "Processing image: 2902/5000\n",
      "Processing image: 2903/5000\n",
      "Processing image: 2904/5000\n",
      "Processing image: 2905/5000\n",
      "Processing image: 2906/5000\n",
      "Processing image: 2907/5000\n",
      "Processing image: 2908/5000\n",
      "Processing image: 2909/5000\n",
      "Processing image: 2910/5000\n",
      "Processing image: 2911/5000\n",
      "Processing image: 2912/5000\n",
      "Processing image: 2913/5000\n",
      "Processing image: 2914/5000\n",
      "Processing image: 2915/5000\n",
      "Processing image: 2916/5000\n",
      "Processing image: 2917/5000\n",
      "Processing image: 2918/5000\n",
      "Processing image: 2919/5000\n",
      "Processing image: 2920/5000\n",
      "Processing image: 2921/5000\n",
      "Processing image: 2922/5000\n",
      "Processing image: 2923/5000\n",
      "Processing image: 2924/5000\n",
      "Processing image: 2925/5000\n",
      "Processing image: 2926/5000\n",
      "Processing image: 2927/5000\n",
      "Processing image: 2928/5000\n",
      "Processing image: 2929/5000\n",
      "Processing image: 2930/5000\n",
      "Processing image: 2931/5000\n",
      "Processing image: 2932/5000\n",
      "Processing image: 2933/5000\n",
      "Processing image: 2934/5000\n",
      "Processing image: 2935/5000\n",
      "Processing image: 2936/5000\n",
      "Processing image: 2937/5000\n",
      "Processing image: 2938/5000\n",
      "Processing image: 2939/5000\n",
      "Processing image: 2940/5000\n",
      "Processing image: 2941/5000\n",
      "Processing image: 2942/5000\n",
      "Processing image: 2943/5000\n",
      "Processing image: 2944/5000\n",
      "Processing image: 2945/5000\n",
      "Processing image: 2946/5000\n",
      "Processing image: 2947/5000\n",
      "Processing image: 2948/5000\n",
      "Processing image: 2949/5000\n",
      "Processing image: 2950/5000\n",
      "Processing image: 2951/5000\n",
      "Processing image: 2952/5000\n",
      "Processing image: 2953/5000\n",
      "Processing image: 2954/5000\n",
      "Processing image: 2955/5000\n",
      "Processing image: 2956/5000\n",
      "Processing image: 2957/5000\n",
      "Processing image: 2958/5000\n",
      "Processing image: 2959/5000\n",
      "Processing image: 2960/5000\n",
      "Processing image: 2961/5000\n",
      "Processing image: 2962/5000\n",
      "Processing image: 2963/5000\n",
      "Processing image: 2964/5000\n",
      "Processing image: 2965/5000\n",
      "Processing image: 2966/5000\n",
      "Processing image: 2967/5000\n",
      "Processing image: 2968/5000\n",
      "Processing image: 2969/5000\n",
      "Processing image: 2970/5000\n",
      "Processing image: 2971/5000\n",
      "Processing image: 2972/5000\n",
      "Processing image: 2973/5000\n",
      "Processing image: 2974/5000\n",
      "Processing image: 2975/5000\n",
      "Processing image: 2976/5000\n",
      "Processing image: 2977/5000\n",
      "Processing image: 2978/5000\n",
      "Processing image: 2979/5000\n",
      "Processing image: 2980/5000\n",
      "Processing image: 2981/5000\n",
      "Processing image: 2982/5000\n",
      "Processing image: 2983/5000\n",
      "Processing image: 2984/5000\n",
      "Processing image: 2985/5000\n",
      "Processing image: 2986/5000\n",
      "Processing image: 2987/5000\n",
      "Processing image: 2988/5000\n",
      "Processing image: 2989/5000\n",
      "Processing image: 2990/5000\n",
      "Processing image: 2991/5000\n",
      "Processing image: 2992/5000\n",
      "Processing image: 2993/5000\n",
      "Processing image: 2994/5000\n",
      "Processing image: 2995/5000\n",
      "Processing image: 2996/5000\n",
      "Processing image: 2997/5000\n",
      "Processing image: 2998/5000\n",
      "Processing image: 2999/5000\n",
      "Processing image: 3000/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 3001/5000\n",
      "Processing image: 3002/5000\n",
      "Processing image: 3003/5000\n",
      "Processing image: 3004/5000\n",
      "Processing image: 3005/5000\n",
      "Processing image: 3006/5000\n",
      "Processing image: 3007/5000\n",
      "Processing image: 3008/5000\n",
      "Processing image: 3009/5000\n",
      "Processing image: 3010/5000\n",
      "Processing image: 3011/5000\n",
      "Processing image: 3012/5000\n",
      "Processing image: 3013/5000\n",
      "Processing image: 3014/5000\n",
      "Processing image: 3015/5000\n",
      "Processing image: 3016/5000\n",
      "Processing image: 3017/5000\n",
      "Processing image: 3018/5000\n",
      "Processing image: 3019/5000\n",
      "Processing image: 3020/5000\n",
      "Processing image: 3021/5000\n",
      "Processing image: 3022/5000\n",
      "Processing image: 3023/5000\n",
      "Processing image: 3024/5000\n",
      "Processing image: 3025/5000\n",
      "Processing image: 3026/5000\n",
      "Processing image: 3027/5000\n",
      "Processing image: 3028/5000\n",
      "Processing image: 3029/5000\n",
      "Processing image: 3030/5000\n",
      "Processing image: 3031/5000\n",
      "Processing image: 3032/5000\n",
      "Processing image: 3033/5000\n",
      "Processing image: 3034/5000\n",
      "Processing image: 3035/5000\n",
      "Processing image: 3036/5000\n",
      "Processing image: 3037/5000\n",
      "Processing image: 3038/5000\n",
      "Processing image: 3039/5000\n",
      "Processing image: 3040/5000\n",
      "Processing image: 3041/5000\n",
      "Processing image: 3042/5000\n",
      "Processing image: 3043/5000\n",
      "Processing image: 3044/5000\n",
      "Processing image: 3045/5000\n",
      "Processing image: 3046/5000\n",
      "Processing image: 3047/5000\n",
      "Processing image: 3048/5000\n",
      "Processing image: 3049/5000\n",
      "Processing image: 3050/5000\n",
      "Processing image: 3051/5000\n",
      "Processing image: 3052/5000\n",
      "Processing image: 3053/5000\n",
      "Processing image: 3054/5000\n",
      "Processing image: 3055/5000\n",
      "Processing image: 3056/5000\n",
      "Processing image: 3057/5000\n",
      "Processing image: 3058/5000\n",
      "Processing image: 3059/5000\n",
      "Processing image: 3060/5000\n",
      "Processing image: 3061/5000\n",
      "Processing image: 3062/5000\n",
      "Processing image: 3063/5000\n",
      "Processing image: 3064/5000\n",
      "Processing image: 3065/5000\n",
      "Processing image: 3066/5000\n",
      "Processing image: 3067/5000\n",
      "Processing image: 3068/5000\n",
      "Processing image: 3069/5000\n",
      "Processing image: 3070/5000\n",
      "Processing image: 3071/5000\n",
      "Processing image: 3072/5000\n",
      "Processing image: 3073/5000\n",
      "Processing image: 3074/5000\n",
      "Processing image: 3075/5000\n",
      "Processing image: 3076/5000\n",
      "Processing image: 3077/5000\n",
      "Processing image: 3078/5000\n",
      "Processing image: 3079/5000\n",
      "Processing image: 3080/5000\n",
      "Processing image: 3081/5000\n",
      "Processing image: 3082/5000\n",
      "Processing image: 3083/5000\n",
      "Processing image: 3084/5000\n",
      "Processing image: 3085/5000\n",
      "Processing image: 3086/5000\n",
      "Processing image: 3087/5000\n",
      "Processing image: 3088/5000\n",
      "Processing image: 3089/5000\n",
      "Processing image: 3090/5000\n",
      "Processing image: 3091/5000\n",
      "Processing image: 3092/5000\n",
      "Processing image: 3093/5000\n",
      "Processing image: 3094/5000\n",
      "Processing image: 3095/5000\n",
      "Processing image: 3096/5000\n",
      "Processing image: 3097/5000\n",
      "Processing image: 3098/5000\n",
      "Processing image: 3099/5000\n",
      "Processing image: 3100/5000\n",
      "Processing image: 3101/5000\n",
      "Processing image: 3102/5000\n",
      "Processing image: 3103/5000\n",
      "Processing image: 3104/5000\n",
      "Processing image: 3105/5000\n",
      "Processing image: 3106/5000\n",
      "Processing image: 3107/5000\n",
      "Processing image: 3108/5000\n",
      "Processing image: 3109/5000\n",
      "Processing image: 3110/5000\n",
      "Processing image: 3111/5000\n",
      "Processing image: 3112/5000\n",
      "Processing image: 3113/5000\n",
      "Processing image: 3114/5000\n",
      "Processing image: 3115/5000\n",
      "Processing image: 3116/5000\n",
      "Processing image: 3117/5000\n",
      "Processing image: 3118/5000\n",
      "Processing image: 3119/5000\n",
      "Processing image: 3120/5000\n",
      "Processing image: 3121/5000\n",
      "Processing image: 3122/5000\n",
      "Processing image: 3123/5000\n",
      "Processing image: 3124/5000\n",
      "Processing image: 3125/5000\n",
      "Processing image: 3126/5000\n",
      "Processing image: 3127/5000\n",
      "Processing image: 3128/5000\n",
      "Processing image: 3129/5000\n",
      "Processing image: 3130/5000\n",
      "Processing image: 3131/5000\n",
      "Processing image: 3132/5000\n",
      "Processing image: 3133/5000\n",
      "Processing image: 3134/5000\n",
      "Processing image: 3135/5000\n",
      "Processing image: 3136/5000\n",
      "Processing image: 3137/5000\n",
      "Processing image: 3138/5000\n",
      "Processing image: 3139/5000\n",
      "Processing image: 3140/5000\n",
      "Processing image: 3141/5000\n",
      "Processing image: 3142/5000\n",
      "Processing image: 3143/5000\n",
      "Processing image: 3144/5000\n",
      "Processing image: 3145/5000\n",
      "Processing image: 3146/5000\n",
      "Processing image: 3147/5000\n",
      "Processing image: 3148/5000\n",
      "Processing image: 3149/5000\n",
      "Processing image: 3150/5000\n",
      "Processing image: 3151/5000\n",
      "Processing image: 3152/5000\n",
      "Processing image: 3153/5000\n",
      "Processing image: 3154/5000\n",
      "Processing image: 3155/5000\n",
      "Processing image: 3156/5000\n",
      "Processing image: 3157/5000\n",
      "Processing image: 3158/5000\n",
      "Processing image: 3159/5000\n",
      "Processing image: 3160/5000\n",
      "Processing image: 3161/5000\n",
      "Processing image: 3162/5000\n",
      "Processing image: 3163/5000\n",
      "Processing image: 3164/5000\n",
      "Processing image: 3165/5000\n",
      "Processing image: 3166/5000\n",
      "Processing image: 3167/5000\n",
      "Processing image: 3168/5000\n",
      "Processing image: 3169/5000\n",
      "Processing image: 3170/5000\n",
      "Processing image: 3171/5000\n",
      "Processing image: 3172/5000\n",
      "Processing image: 3173/5000\n",
      "Processing image: 3174/5000\n",
      "Processing image: 3175/5000\n",
      "Processing image: 3176/5000\n",
      "Processing image: 3177/5000\n",
      "Processing image: 3178/5000\n",
      "Processing image: 3179/5000\n",
      "Processing image: 3180/5000\n",
      "Processing image: 3181/5000\n",
      "Processing image: 3182/5000\n",
      "Processing image: 3183/5000\n",
      "Processing image: 3184/5000\n",
      "Processing image: 3185/5000\n",
      "Processing image: 3186/5000\n",
      "Processing image: 3187/5000\n",
      "Processing image: 3188/5000\n",
      "Processing image: 3189/5000\n",
      "Processing image: 3190/5000\n",
      "Processing image: 3191/5000\n",
      "Processing image: 3192/5000\n",
      "Processing image: 3193/5000\n",
      "Processing image: 3194/5000\n",
      "Processing image: 3195/5000\n",
      "Processing image: 3196/5000\n",
      "Processing image: 3197/5000\n",
      "Processing image: 3198/5000\n",
      "Processing image: 3199/5000\n",
      "Processing image: 3200/5000\n",
      "Processing image: 3201/5000\n",
      "Processing image: 3202/5000\n",
      "Processing image: 3203/5000\n",
      "Processing image: 3204/5000\n",
      "Processing image: 3205/5000\n",
      "Processing image: 3206/5000\n",
      "Processing image: 3207/5000\n",
      "Processing image: 3208/5000\n",
      "Processing image: 3209/5000\n",
      "Processing image: 3210/5000\n",
      "Processing image: 3211/5000\n",
      "Processing image: 3212/5000\n",
      "Processing image: 3213/5000\n",
      "Processing image: 3214/5000\n",
      "Processing image: 3215/5000\n",
      "Processing image: 3216/5000\n",
      "Processing image: 3217/5000\n",
      "Processing image: 3218/5000\n",
      "Processing image: 3219/5000\n",
      "Processing image: 3220/5000\n",
      "Processing image: 3221/5000\n",
      "Processing image: 3222/5000\n",
      "Processing image: 3223/5000\n",
      "Processing image: 3224/5000\n",
      "Processing image: 3225/5000\n",
      "Processing image: 3226/5000\n",
      "Processing image: 3227/5000\n",
      "Processing image: 3228/5000\n",
      "Processing image: 3229/5000\n",
      "Processing image: 3230/5000\n",
      "Processing image: 3231/5000\n",
      "Processing image: 3232/5000\n",
      "Processing image: 3233/5000\n",
      "Processing image: 3234/5000\n",
      "Processing image: 3235/5000\n",
      "Processing image: 3236/5000\n",
      "Processing image: 3237/5000\n",
      "Processing image: 3238/5000\n",
      "Processing image: 3239/5000\n",
      "Processing image: 3240/5000\n",
      "Processing image: 3241/5000\n",
      "Processing image: 3242/5000\n",
      "Processing image: 3243/5000\n",
      "Processing image: 3244/5000\n",
      "Processing image: 3245/5000\n",
      "Processing image: 3246/5000\n",
      "Processing image: 3247/5000\n",
      "Processing image: 3248/5000\n",
      "Processing image: 3249/5000\n",
      "Processing image: 3250/5000\n",
      "Processing image: 3251/5000\n",
      "Processing image: 3252/5000\n",
      "Processing image: 3253/5000\n",
      "Processing image: 3254/5000\n",
      "Processing image: 3255/5000\n",
      "Processing image: 3256/5000\n",
      "Processing image: 3257/5000\n",
      "Processing image: 3258/5000\n",
      "Processing image: 3259/5000\n",
      "Processing image: 3260/5000\n",
      "Processing image: 3261/5000\n",
      "Processing image: 3262/5000\n",
      "Processing image: 3263/5000\n",
      "Processing image: 3264/5000\n",
      "Processing image: 3265/5000\n",
      "Processing image: 3266/5000\n",
      "Processing image: 3267/5000\n",
      "Processing image: 3268/5000\n",
      "Processing image: 3269/5000\n",
      "Processing image: 3270/5000\n",
      "Processing image: 3271/5000\n",
      "Processing image: 3272/5000\n",
      "Processing image: 3273/5000\n",
      "Processing image: 3274/5000\n",
      "Processing image: 3275/5000\n",
      "Processing image: 3276/5000\n",
      "Processing image: 3277/5000\n",
      "Processing image: 3278/5000\n",
      "Processing image: 3279/5000\n",
      "Processing image: 3280/5000\n",
      "Processing image: 3281/5000\n",
      "Processing image: 3282/5000\n",
      "Processing image: 3283/5000\n",
      "Processing image: 3284/5000\n",
      "Processing image: 3285/5000\n",
      "Processing image: 3286/5000\n",
      "Processing image: 3287/5000\n",
      "Processing image: 3288/5000\n",
      "Processing image: 3289/5000\n",
      "Processing image: 3290/5000\n",
      "Processing image: 3291/5000\n",
      "Processing image: 3292/5000\n",
      "Processing image: 3293/5000\n",
      "Processing image: 3294/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 3295/5000\n",
      "Processing image: 3296/5000\n",
      "Processing image: 3297/5000\n",
      "Processing image: 3298/5000\n",
      "Processing image: 3299/5000\n",
      "Processing image: 3300/5000\n",
      "Processing image: 3301/5000\n",
      "Processing image: 3302/5000\n",
      "Processing image: 3303/5000\n",
      "Processing image: 3304/5000\n",
      "Processing image: 3305/5000\n",
      "Processing image: 3306/5000\n",
      "Processing image: 3307/5000\n",
      "Processing image: 3308/5000\n",
      "Processing image: 3309/5000\n",
      "Processing image: 3310/5000\n",
      "Processing image: 3311/5000\n",
      "Processing image: 3312/5000\n",
      "Processing image: 3313/5000\n",
      "Processing image: 3314/5000\n",
      "Processing image: 3315/5000\n",
      "Processing image: 3316/5000\n",
      "Processing image: 3317/5000\n",
      "Processing image: 3318/5000\n",
      "Processing image: 3319/5000\n",
      "Processing image: 3320/5000\n",
      "Processing image: 3321/5000\n",
      "Processing image: 3322/5000\n",
      "Processing image: 3323/5000\n",
      "Processing image: 3324/5000\n",
      "Processing image: 3325/5000\n",
      "Processing image: 3326/5000\n",
      "Processing image: 3327/5000\n",
      "Processing image: 3328/5000\n",
      "Processing image: 3329/5000\n",
      "Processing image: 3330/5000\n",
      "Processing image: 3331/5000\n",
      "Processing image: 3332/5000\n",
      "Processing image: 3333/5000\n",
      "Processing image: 3334/5000\n",
      "Processing image: 3335/5000\n",
      "Processing image: 3336/5000\n",
      "Processing image: 3337/5000\n",
      "Processing image: 3338/5000\n",
      "Processing image: 3339/5000\n",
      "Processing image: 3340/5000\n",
      "Processing image: 3341/5000\n",
      "Processing image: 3342/5000\n",
      "Processing image: 3343/5000\n",
      "Processing image: 3344/5000\n",
      "Processing image: 3345/5000\n",
      "Processing image: 3346/5000\n",
      "Processing image: 3347/5000\n",
      "Processing image: 3348/5000\n",
      "Processing image: 3349/5000\n",
      "Processing image: 3350/5000\n",
      "Processing image: 3351/5000\n",
      "Processing image: 3352/5000\n",
      "Processing image: 3353/5000\n",
      "Processing image: 3354/5000\n",
      "Processing image: 3355/5000\n",
      "Processing image: 3356/5000\n",
      "Processing image: 3357/5000\n",
      "Processing image: 3358/5000\n",
      "Processing image: 3359/5000\n",
      "Processing image: 3360/5000\n",
      "Processing image: 3361/5000\n",
      "Processing image: 3362/5000\n",
      "Processing image: 3363/5000\n",
      "Processing image: 3364/5000\n",
      "Processing image: 3365/5000\n",
      "Processing image: 3366/5000\n",
      "Processing image: 3367/5000\n",
      "Processing image: 3368/5000\n",
      "Processing image: 3369/5000\n",
      "Processing image: 3370/5000\n",
      "Processing image: 3371/5000\n",
      "Processing image: 3372/5000\n",
      "Processing image: 3373/5000\n",
      "Processing image: 3374/5000\n",
      "Processing image: 3375/5000\n",
      "Processing image: 3376/5000\n",
      "Processing image: 3377/5000\n",
      "Processing image: 3378/5000\n",
      "Processing image: 3379/5000\n",
      "Processing image: 3380/5000\n",
      "Processing image: 3381/5000\n",
      "Processing image: 3382/5000\n",
      "Processing image: 3383/5000\n",
      "Processing image: 3384/5000\n",
      "Processing image: 3385/5000\n",
      "Processing image: 3386/5000\n",
      "Processing image: 3387/5000\n",
      "Processing image: 3388/5000\n",
      "Processing image: 3389/5000\n",
      "Processing image: 3390/5000\n",
      "Processing image: 3391/5000\n",
      "Processing image: 3392/5000\n",
      "Processing image: 3393/5000\n",
      "Processing image: 3394/5000\n",
      "Processing image: 3395/5000\n",
      "Processing image: 3396/5000\n",
      "Processing image: 3397/5000\n",
      "Processing image: 3398/5000\n",
      "Processing image: 3399/5000\n",
      "Processing image: 3400/5000\n",
      "Processing image: 3401/5000\n",
      "Processing image: 3402/5000\n",
      "Processing image: 3403/5000\n",
      "Processing image: 3404/5000\n",
      "Processing image: 3405/5000\n",
      "Processing image: 3406/5000\n",
      "Processing image: 3407/5000\n",
      "Processing image: 3408/5000\n",
      "Processing image: 3409/5000\n",
      "Processing image: 3410/5000\n",
      "Processing image: 3411/5000\n",
      "Processing image: 3412/5000\n",
      "Processing image: 3413/5000\n",
      "Processing image: 3414/5000\n",
      "Processing image: 3415/5000\n",
      "Processing image: 3416/5000\n",
      "Processing image: 3417/5000\n",
      "Processing image: 3418/5000\n",
      "Processing image: 3419/5000\n",
      "Processing image: 3420/5000\n",
      "Processing image: 3421/5000\n",
      "Processing image: 3422/5000\n",
      "Processing image: 3423/5000\n",
      "Processing image: 3424/5000\n",
      "Processing image: 3425/5000\n",
      "Processing image: 3426/5000\n",
      "Processing image: 3427/5000\n",
      "Processing image: 3428/5000\n",
      "Processing image: 3429/5000\n",
      "Processing image: 3430/5000\n",
      "Processing image: 3431/5000\n",
      "Processing image: 3432/5000\n",
      "Processing image: 3433/5000\n",
      "Processing image: 3434/5000\n",
      "Processing image: 3435/5000\n",
      "Processing image: 3436/5000\n",
      "Processing image: 3437/5000\n",
      "Processing image: 3438/5000\n",
      "Processing image: 3439/5000\n",
      "Processing image: 3440/5000\n",
      "Processing image: 3441/5000\n",
      "Processing image: 3442/5000\n",
      "Processing image: 3443/5000\n",
      "Processing image: 3444/5000\n",
      "Processing image: 3445/5000\n",
      "Processing image: 3446/5000\n",
      "Processing image: 3447/5000\n",
      "Processing image: 3448/5000\n",
      "Processing image: 3449/5000\n",
      "Processing image: 3450/5000\n",
      "Processing image: 3451/5000\n",
      "Processing image: 3452/5000\n",
      "Processing image: 3453/5000\n",
      "Processing image: 3454/5000\n",
      "Processing image: 3455/5000\n",
      "Processing image: 3456/5000\n",
      "Processing image: 3457/5000\n",
      "Processing image: 3458/5000\n",
      "Processing image: 3459/5000\n",
      "Processing image: 3460/5000\n",
      "Processing image: 3461/5000\n",
      "Processing image: 3462/5000\n",
      "Processing image: 3463/5000\n",
      "Processing image: 3464/5000\n",
      "Processing image: 3465/5000\n",
      "Processing image: 3466/5000\n",
      "Processing image: 3467/5000\n",
      "Processing image: 3468/5000\n",
      "Processing image: 3469/5000\n",
      "Processing image: 3470/5000\n",
      "Processing image: 3471/5000\n",
      "Processing image: 3472/5000\n",
      "Processing image: 3473/5000\n",
      "Processing image: 3474/5000\n",
      "Processing image: 3475/5000\n",
      "Processing image: 3476/5000\n",
      "Processing image: 3477/5000\n",
      "Processing image: 3478/5000\n",
      "Processing image: 3479/5000\n",
      "Processing image: 3480/5000\n",
      "Processing image: 3481/5000\n",
      "Processing image: 3482/5000\n",
      "Processing image: 3483/5000\n",
      "Processing image: 3484/5000\n",
      "Processing image: 3485/5000\n",
      "Processing image: 3486/5000\n",
      "Processing image: 3487/5000\n",
      "Processing image: 3488/5000\n",
      "Processing image: 3489/5000\n",
      "Processing image: 3490/5000\n",
      "Processing image: 3491/5000\n",
      "Processing image: 3492/5000\n",
      "Processing image: 3493/5000\n",
      "Processing image: 3494/5000\n",
      "Processing image: 3495/5000\n",
      "Processing image: 3496/5000\n",
      "Processing image: 3497/5000\n",
      "Processing image: 3498/5000\n",
      "Processing image: 3499/5000\n",
      "Processing image: 3500/5000\n",
      "Processing image: 3501/5000\n",
      "Processing image: 3502/5000\n",
      "Processing image: 3503/5000\n",
      "Processing image: 3504/5000\n",
      "Processing image: 3505/5000\n",
      "Processing image: 3506/5000\n",
      "Processing image: 3507/5000\n",
      "Processing image: 3508/5000\n",
      "Processing image: 3509/5000\n",
      "Processing image: 3510/5000\n",
      "Processing image: 3511/5000\n",
      "Processing image: 3512/5000\n",
      "Processing image: 3513/5000\n",
      "Processing image: 3514/5000\n",
      "Processing image: 3515/5000\n",
      "Processing image: 3516/5000\n",
      "Processing image: 3517/5000\n",
      "Processing image: 3518/5000\n",
      "Processing image: 3519/5000\n",
      "Processing image: 3520/5000\n",
      "Processing image: 3521/5000\n",
      "Processing image: 3522/5000\n",
      "Processing image: 3523/5000\n",
      "Processing image: 3524/5000\n",
      "Processing image: 3525/5000\n",
      "Processing image: 3526/5000\n",
      "Processing image: 3527/5000\n",
      "Processing image: 3528/5000\n",
      "Processing image: 3529/5000\n",
      "Processing image: 3530/5000\n",
      "Processing image: 3531/5000\n",
      "Processing image: 3532/5000\n",
      "Processing image: 3533/5000\n",
      "Processing image: 3534/5000\n",
      "Processing image: 3535/5000\n",
      "Processing image: 3536/5000\n",
      "Processing image: 3537/5000\n",
      "Processing image: 3538/5000\n",
      "Processing image: 3539/5000\n",
      "Processing image: 3540/5000\n",
      "Processing image: 3541/5000\n",
      "Processing image: 3542/5000\n",
      "Processing image: 3543/5000\n",
      "Processing image: 3544/5000\n",
      "Processing image: 3545/5000\n",
      "Processing image: 3546/5000\n",
      "Processing image: 3547/5000\n",
      "Processing image: 3548/5000\n",
      "Processing image: 3549/5000\n",
      "Processing image: 3550/5000\n",
      "Processing image: 3551/5000\n",
      "Processing image: 3552/5000\n",
      "Processing image: 3553/5000\n",
      "Processing image: 3554/5000\n",
      "Processing image: 3555/5000\n",
      "Processing image: 3556/5000\n",
      "Processing image: 3557/5000\n",
      "Processing image: 3558/5000\n",
      "Processing image: 3559/5000\n",
      "Processing image: 3560/5000\n",
      "Processing image: 3561/5000\n",
      "Processing image: 3562/5000\n",
      "Processing image: 3563/5000\n",
      "Processing image: 3564/5000\n",
      "Processing image: 3565/5000\n",
      "Processing image: 3566/5000\n",
      "Processing image: 3567/5000\n",
      "Processing image: 3568/5000\n",
      "Processing image: 3569/5000\n",
      "Processing image: 3570/5000\n",
      "Processing image: 3571/5000\n",
      "Processing image: 3572/5000\n",
      "Processing image: 3573/5000\n",
      "Processing image: 3574/5000\n",
      "Processing image: 3575/5000\n",
      "Processing image: 3576/5000\n",
      "Processing image: 3577/5000\n",
      "Processing image: 3578/5000\n",
      "Processing image: 3579/5000\n",
      "Processing image: 3580/5000\n",
      "Processing image: 3581/5000\n",
      "Processing image: 3582/5000\n",
      "Processing image: 3583/5000\n",
      "Processing image: 3584/5000\n",
      "Processing image: 3585/5000\n",
      "Processing image: 3586/5000\n",
      "Processing image: 3587/5000\n",
      "Processing image: 3588/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 3589/5000\n",
      "Processing image: 3590/5000\n",
      "Processing image: 3591/5000\n",
      "Processing image: 3592/5000\n",
      "Processing image: 3593/5000\n",
      "Processing image: 3594/5000\n",
      "Processing image: 3595/5000\n",
      "Processing image: 3596/5000\n",
      "Processing image: 3597/5000\n",
      "Processing image: 3598/5000\n",
      "Processing image: 3599/5000\n",
      "Processing image: 3600/5000\n",
      "Processing image: 3601/5000\n",
      "Processing image: 3602/5000\n",
      "Processing image: 3603/5000\n",
      "Processing image: 3604/5000\n",
      "Processing image: 3605/5000\n",
      "Processing image: 3606/5000\n",
      "Processing image: 3607/5000\n",
      "Processing image: 3608/5000\n",
      "Processing image: 3609/5000\n",
      "Processing image: 3610/5000\n",
      "Processing image: 3611/5000\n",
      "Processing image: 3612/5000\n",
      "Processing image: 3613/5000\n",
      "Processing image: 3614/5000\n",
      "Processing image: 3615/5000\n",
      "Processing image: 3616/5000\n",
      "Processing image: 3617/5000\n",
      "Processing image: 3618/5000\n",
      "Processing image: 3619/5000\n",
      "Processing image: 3620/5000\n",
      "Processing image: 3621/5000\n",
      "Processing image: 3622/5000\n",
      "Processing image: 3623/5000\n",
      "Processing image: 3624/5000\n",
      "Processing image: 3625/5000\n",
      "Processing image: 3626/5000\n",
      "Processing image: 3627/5000\n",
      "Processing image: 3628/5000\n",
      "Processing image: 3629/5000\n",
      "Processing image: 3630/5000\n",
      "Processing image: 3631/5000\n",
      "Processing image: 3632/5000\n",
      "Processing image: 3633/5000\n",
      "Processing image: 3634/5000\n",
      "Processing image: 3635/5000\n",
      "Processing image: 3636/5000\n",
      "Processing image: 3637/5000\n",
      "Processing image: 3638/5000\n",
      "Processing image: 3639/5000\n",
      "Processing image: 3640/5000\n",
      "Processing image: 3641/5000\n",
      "Processing image: 3642/5000\n",
      "Processing image: 3643/5000\n",
      "Processing image: 3644/5000\n",
      "Processing image: 3645/5000\n",
      "Processing image: 3646/5000\n",
      "Processing image: 3647/5000\n",
      "Processing image: 3648/5000\n",
      "Processing image: 3649/5000\n",
      "Processing image: 3650/5000\n",
      "Processing image: 3651/5000\n",
      "Processing image: 3652/5000\n",
      "Processing image: 3653/5000\n",
      "Processing image: 3654/5000\n",
      "Processing image: 3655/5000\n",
      "Processing image: 3656/5000\n",
      "Processing image: 3657/5000\n",
      "Processing image: 3658/5000\n",
      "Processing image: 3659/5000\n",
      "Processing image: 3660/5000\n",
      "Processing image: 3661/5000\n",
      "Processing image: 3662/5000\n",
      "Processing image: 3663/5000\n",
      "Processing image: 3664/5000\n",
      "Processing image: 3665/5000\n",
      "Processing image: 3666/5000\n",
      "Processing image: 3667/5000\n",
      "Processing image: 3668/5000\n",
      "Processing image: 3669/5000\n",
      "Processing image: 3670/5000\n",
      "Processing image: 3671/5000\n",
      "Processing image: 3672/5000\n",
      "Processing image: 3673/5000\n",
      "Processing image: 3674/5000\n",
      "Processing image: 3675/5000\n",
      "Processing image: 3676/5000\n",
      "Processing image: 3677/5000\n",
      "Processing image: 3678/5000\n",
      "Processing image: 3679/5000\n",
      "Processing image: 3680/5000\n",
      "Processing image: 3681/5000\n",
      "Processing image: 3682/5000\n",
      "Processing image: 3683/5000\n",
      "Processing image: 3684/5000\n",
      "Processing image: 3685/5000\n",
      "Processing image: 3686/5000\n",
      "Processing image: 3687/5000\n",
      "Processing image: 3688/5000\n",
      "Processing image: 3689/5000\n",
      "Processing image: 3690/5000\n",
      "Processing image: 3691/5000\n",
      "Processing image: 3692/5000\n",
      "Processing image: 3693/5000\n",
      "Processing image: 3694/5000\n",
      "Processing image: 3695/5000\n",
      "Processing image: 3696/5000\n",
      "Processing image: 3697/5000\n",
      "Processing image: 3698/5000\n",
      "Processing image: 3699/5000\n",
      "Processing image: 3700/5000\n",
      "Processing image: 3701/5000\n",
      "Processing image: 3702/5000\n",
      "Processing image: 3703/5000\n",
      "Processing image: 3704/5000\n",
      "Processing image: 3705/5000\n",
      "Processing image: 3706/5000\n",
      "Processing image: 3707/5000\n",
      "Processing image: 3708/5000\n",
      "Processing image: 3709/5000\n",
      "Processing image: 3710/5000\n",
      "Processing image: 3711/5000\n",
      "Processing image: 3712/5000\n",
      "Processing image: 3713/5000\n",
      "Processing image: 3714/5000\n",
      "Processing image: 3715/5000\n",
      "Processing image: 3716/5000\n",
      "Processing image: 3717/5000\n",
      "Processing image: 3718/5000\n",
      "Processing image: 3719/5000\n",
      "Processing image: 3720/5000\n",
      "Processing image: 3721/5000\n",
      "Processing image: 3722/5000\n",
      "Processing image: 3723/5000\n",
      "Processing image: 3724/5000\n",
      "Processing image: 3725/5000\n",
      "Processing image: 3726/5000\n",
      "Processing image: 3727/5000\n",
      "Processing image: 3728/5000\n",
      "Processing image: 3729/5000\n",
      "Processing image: 3730/5000\n",
      "Processing image: 3731/5000\n",
      "Processing image: 3732/5000\n",
      "Processing image: 3733/5000\n",
      "Processing image: 3734/5000\n",
      "Processing image: 3735/5000\n",
      "Processing image: 3736/5000\n",
      "Processing image: 3737/5000\n",
      "Processing image: 3738/5000\n",
      "Processing image: 3739/5000\n",
      "Processing image: 3740/5000\n",
      "Processing image: 3741/5000\n",
      "Processing image: 3742/5000\n",
      "Processing image: 3743/5000\n",
      "Processing image: 3744/5000\n",
      "Processing image: 3745/5000\n",
      "Processing image: 3746/5000\n",
      "Processing image: 3747/5000\n",
      "Processing image: 3748/5000\n",
      "Processing image: 3749/5000\n",
      "Processing image: 3750/5000\n",
      "Processing image: 3751/5000\n",
      "Processing image: 3752/5000\n",
      "Processing image: 3753/5000\n",
      "Processing image: 3754/5000\n",
      "Processing image: 3755/5000\n",
      "Processing image: 3756/5000\n",
      "Processing image: 3757/5000\n",
      "Processing image: 3758/5000\n",
      "Processing image: 3759/5000\n",
      "Processing image: 3760/5000\n",
      "Processing image: 3761/5000\n",
      "Processing image: 3762/5000\n",
      "Processing image: 3763/5000\n",
      "Processing image: 3764/5000\n",
      "Processing image: 3765/5000\n",
      "Processing image: 3766/5000\n",
      "Processing image: 3767/5000\n",
      "Processing image: 3768/5000\n",
      "Processing image: 3769/5000\n",
      "Processing image: 3770/5000\n",
      "Processing image: 3771/5000\n",
      "Processing image: 3772/5000\n",
      "Processing image: 3773/5000\n",
      "Processing image: 3774/5000\n",
      "Processing image: 3775/5000\n",
      "Processing image: 3776/5000\n",
      "Processing image: 3777/5000\n",
      "Processing image: 3778/5000\n",
      "Processing image: 3779/5000\n",
      "Processing image: 3780/5000\n",
      "Processing image: 3781/5000\n",
      "Processing image: 3782/5000\n",
      "Processing image: 3783/5000\n",
      "Processing image: 3784/5000\n",
      "Processing image: 3785/5000\n",
      "Processing image: 3786/5000\n",
      "Processing image: 3787/5000\n",
      "Processing image: 3788/5000\n",
      "Processing image: 3789/5000\n",
      "Processing image: 3790/5000\n",
      "Processing image: 3791/5000\n",
      "Processing image: 3792/5000\n",
      "Processing image: 3793/5000\n",
      "Processing image: 3794/5000\n",
      "Processing image: 3795/5000\n",
      "Processing image: 3796/5000\n",
      "Processing image: 3797/5000\n",
      "Processing image: 3798/5000\n",
      "Processing image: 3799/5000\n",
      "Processing image: 3800/5000\n",
      "Processing image: 3801/5000\n",
      "Processing image: 3802/5000\n",
      "Processing image: 3803/5000\n",
      "Processing image: 3804/5000\n",
      "Processing image: 3805/5000\n",
      "Processing image: 3806/5000\n",
      "Processing image: 3807/5000\n",
      "Processing image: 3808/5000\n",
      "Processing image: 3809/5000\n",
      "Processing image: 3810/5000\n",
      "Processing image: 3811/5000\n",
      "Processing image: 3812/5000\n",
      "Processing image: 3813/5000\n",
      "Processing image: 3814/5000\n",
      "Processing image: 3815/5000\n",
      "Processing image: 3816/5000\n",
      "Processing image: 3817/5000\n",
      "Processing image: 3818/5000\n",
      "Processing image: 3819/5000\n",
      "Processing image: 3820/5000\n",
      "Processing image: 3821/5000\n",
      "Processing image: 3822/5000\n",
      "Processing image: 3823/5000\n",
      "Processing image: 3824/5000\n",
      "Processing image: 3825/5000\n",
      "Processing image: 3826/5000\n",
      "Processing image: 3827/5000\n",
      "Processing image: 3828/5000\n",
      "Processing image: 3829/5000\n",
      "Processing image: 3830/5000\n",
      "Processing image: 3831/5000\n",
      "Processing image: 3832/5000\n",
      "Processing image: 3833/5000\n",
      "Processing image: 3834/5000\n",
      "Processing image: 3835/5000\n",
      "Processing image: 3836/5000\n",
      "Processing image: 3837/5000\n",
      "Processing image: 3838/5000\n",
      "Processing image: 3839/5000\n",
      "Processing image: 3840/5000\n",
      "Processing image: 3841/5000\n",
      "Processing image: 3842/5000\n",
      "Processing image: 3843/5000\n",
      "Processing image: 3844/5000\n",
      "Processing image: 3845/5000\n",
      "Processing image: 3846/5000\n",
      "Processing image: 3847/5000\n",
      "Processing image: 3848/5000\n",
      "Processing image: 3849/5000\n",
      "Processing image: 3850/5000\n",
      "Processing image: 3851/5000\n",
      "Processing image: 3852/5000\n",
      "Processing image: 3853/5000\n",
      "Processing image: 3854/5000\n",
      "Processing image: 3855/5000\n",
      "Processing image: 3856/5000\n",
      "Processing image: 3857/5000\n",
      "Processing image: 3858/5000\n",
      "Processing image: 3859/5000\n",
      "Processing image: 3860/5000\n",
      "Processing image: 3861/5000\n",
      "Processing image: 3862/5000\n",
      "Processing image: 3863/5000\n",
      "Processing image: 3864/5000\n",
      "Processing image: 3865/5000\n",
      "Processing image: 3866/5000\n",
      "Processing image: 3867/5000\n",
      "Processing image: 3868/5000\n",
      "Processing image: 3869/5000\n",
      "Processing image: 3870/5000\n",
      "Processing image: 3871/5000\n",
      "Processing image: 3872/5000\n",
      "Processing image: 3873/5000\n",
      "Processing image: 3874/5000\n",
      "Processing image: 3875/5000\n",
      "Processing image: 3876/5000\n",
      "Processing image: 3877/5000\n",
      "Processing image: 3878/5000\n",
      "Processing image: 3879/5000\n",
      "Processing image: 3880/5000\n",
      "Processing image: 3881/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 3882/5000\n",
      "Processing image: 3883/5000\n",
      "Processing image: 3884/5000\n",
      "Processing image: 3885/5000\n",
      "Processing image: 3886/5000\n",
      "Processing image: 3887/5000\n",
      "Processing image: 3888/5000\n",
      "Processing image: 3889/5000\n",
      "Processing image: 3890/5000\n",
      "Processing image: 3891/5000\n",
      "Processing image: 3892/5000\n",
      "Processing image: 3893/5000\n",
      "Processing image: 3894/5000\n",
      "Processing image: 3895/5000\n",
      "Processing image: 3896/5000\n",
      "Processing image: 3897/5000\n",
      "Processing image: 3898/5000\n",
      "Processing image: 3899/5000\n",
      "Processing image: 3900/5000\n",
      "Processing image: 3901/5000\n",
      "Processing image: 3902/5000\n",
      "Processing image: 3903/5000\n",
      "Processing image: 3904/5000\n",
      "Processing image: 3905/5000\n",
      "Processing image: 3906/5000\n",
      "Processing image: 3907/5000\n",
      "Processing image: 3908/5000\n",
      "Processing image: 3909/5000\n",
      "Processing image: 3910/5000\n",
      "Processing image: 3911/5000\n",
      "Processing image: 3912/5000\n",
      "Processing image: 3913/5000\n",
      "Processing image: 3914/5000\n",
      "Processing image: 3915/5000\n",
      "Processing image: 3916/5000\n",
      "Processing image: 3917/5000\n",
      "Processing image: 3918/5000\n",
      "Processing image: 3919/5000\n",
      "Processing image: 3920/5000\n",
      "Processing image: 3921/5000\n",
      "Processing image: 3922/5000\n",
      "Processing image: 3923/5000\n",
      "Processing image: 3924/5000\n",
      "Processing image: 3925/5000\n",
      "Processing image: 3926/5000\n",
      "Processing image: 3927/5000\n",
      "Processing image: 3928/5000\n",
      "Processing image: 3929/5000\n",
      "Processing image: 3930/5000\n",
      "Processing image: 3931/5000\n",
      "Processing image: 3932/5000\n",
      "Processing image: 3933/5000\n",
      "Processing image: 3934/5000\n",
      "Processing image: 3935/5000\n",
      "Processing image: 3936/5000\n",
      "Processing image: 3937/5000\n",
      "Processing image: 3938/5000\n",
      "Processing image: 3939/5000\n",
      "Processing image: 3940/5000\n",
      "Processing image: 3941/5000\n",
      "Processing image: 3942/5000\n",
      "Processing image: 3943/5000\n",
      "Processing image: 3944/5000\n",
      "Processing image: 3945/5000\n",
      "Processing image: 3946/5000\n",
      "Processing image: 3947/5000\n",
      "Processing image: 3948/5000\n",
      "Processing image: 3949/5000\n",
      "Processing image: 3950/5000\n",
      "Processing image: 3951/5000\n",
      "Processing image: 3952/5000\n",
      "Processing image: 3953/5000\n",
      "Processing image: 3954/5000\n",
      "Processing image: 3955/5000\n",
      "Processing image: 3956/5000\n",
      "Processing image: 3957/5000\n",
      "Processing image: 3958/5000\n",
      "Processing image: 3959/5000\n",
      "Processing image: 3960/5000\n",
      "Processing image: 3961/5000\n",
      "Processing image: 3962/5000\n",
      "Processing image: 3963/5000\n",
      "Processing image: 3964/5000\n",
      "Processing image: 3965/5000\n",
      "Processing image: 3966/5000\n",
      "Processing image: 3967/5000\n",
      "Processing image: 3968/5000\n",
      "Processing image: 3969/5000\n",
      "Processing image: 3970/5000\n",
      "Processing image: 3971/5000\n",
      "Processing image: 3972/5000\n",
      "Processing image: 3973/5000\n",
      "Processing image: 3974/5000\n",
      "Processing image: 3975/5000\n",
      "Processing image: 3976/5000\n",
      "Processing image: 3977/5000\n",
      "Processing image: 3978/5000\n",
      "Processing image: 3979/5000\n",
      "Processing image: 3980/5000\n",
      "Processing image: 3981/5000\n",
      "Processing image: 3982/5000\n",
      "Processing image: 3983/5000\n",
      "Processing image: 3984/5000\n",
      "Processing image: 3985/5000\n",
      "Processing image: 3986/5000\n",
      "Processing image: 3987/5000\n",
      "Processing image: 3988/5000\n",
      "Processing image: 3989/5000\n",
      "Processing image: 3990/5000\n",
      "Processing image: 3991/5000\n",
      "Processing image: 3992/5000\n",
      "Processing image: 3993/5000\n",
      "Processing image: 3994/5000\n",
      "Processing image: 3995/5000\n",
      "Processing image: 3996/5000\n",
      "Processing image: 3997/5000\n",
      "Processing image: 3998/5000\n",
      "Processing image: 3999/5000\n",
      "Processing image: 4000/5000\n",
      "Processing image: 4001/5000\n",
      "Processing image: 4002/5000\n",
      "Processing image: 4003/5000\n",
      "Processing image: 4004/5000\n",
      "Processing image: 4005/5000\n",
      "Processing image: 4006/5000\n",
      "Processing image: 4007/5000\n",
      "Processing image: 4008/5000\n",
      "Processing image: 4009/5000\n",
      "Processing image: 4010/5000\n",
      "Processing image: 4011/5000\n",
      "Processing image: 4012/5000\n",
      "Processing image: 4013/5000\n",
      "Processing image: 4014/5000\n",
      "Processing image: 4015/5000\n",
      "Processing image: 4016/5000\n",
      "Processing image: 4017/5000\n",
      "Processing image: 4018/5000\n",
      "Processing image: 4019/5000\n",
      "Processing image: 4020/5000\n",
      "Processing image: 4021/5000\n",
      "Processing image: 4022/5000\n",
      "Processing image: 4023/5000\n",
      "Processing image: 4024/5000\n",
      "Processing image: 4025/5000\n",
      "Processing image: 4026/5000\n",
      "Processing image: 4027/5000\n",
      "Processing image: 4028/5000\n",
      "Processing image: 4029/5000\n",
      "Processing image: 4030/5000\n",
      "Processing image: 4031/5000\n",
      "Processing image: 4032/5000\n",
      "Processing image: 4033/5000\n",
      "Processing image: 4034/5000\n",
      "Processing image: 4035/5000\n",
      "Processing image: 4036/5000\n",
      "Processing image: 4037/5000\n",
      "Processing image: 4038/5000\n",
      "Processing image: 4039/5000\n",
      "Processing image: 4040/5000\n",
      "Processing image: 4041/5000\n",
      "Processing image: 4042/5000\n",
      "Processing image: 4043/5000\n",
      "Processing image: 4044/5000\n",
      "Processing image: 4045/5000\n",
      "Processing image: 4046/5000\n",
      "Processing image: 4047/5000\n",
      "Processing image: 4048/5000\n",
      "Processing image: 4049/5000\n",
      "Processing image: 4050/5000\n",
      "Processing image: 4051/5000\n",
      "Processing image: 4052/5000\n",
      "Processing image: 4053/5000\n",
      "Processing image: 4054/5000\n",
      "Processing image: 4055/5000\n",
      "Processing image: 4056/5000\n",
      "Processing image: 4057/5000\n",
      "Processing image: 4058/5000\n",
      "Processing image: 4059/5000\n",
      "Processing image: 4060/5000\n",
      "Processing image: 4061/5000\n",
      "Processing image: 4062/5000\n",
      "Processing image: 4063/5000\n",
      "Processing image: 4064/5000\n",
      "Processing image: 4065/5000\n",
      "Processing image: 4066/5000\n",
      "Processing image: 4067/5000\n",
      "Processing image: 4068/5000\n",
      "Processing image: 4069/5000\n",
      "Processing image: 4070/5000\n",
      "Processing image: 4071/5000\n",
      "Processing image: 4072/5000\n",
      "Processing image: 4073/5000\n",
      "Processing image: 4074/5000\n",
      "Processing image: 4075/5000\n",
      "Processing image: 4076/5000\n",
      "Processing image: 4077/5000\n",
      "Processing image: 4078/5000\n",
      "Processing image: 4079/5000\n",
      "Processing image: 4080/5000\n",
      "Processing image: 4081/5000\n",
      "Processing image: 4082/5000\n",
      "Processing image: 4083/5000\n",
      "Processing image: 4084/5000\n",
      "Processing image: 4085/5000\n",
      "Processing image: 4086/5000\n",
      "Processing image: 4087/5000\n",
      "Processing image: 4088/5000\n",
      "Processing image: 4089/5000\n",
      "Processing image: 4090/5000\n",
      "Processing image: 4091/5000\n",
      "Processing image: 4092/5000\n",
      "Processing image: 4093/5000\n",
      "Processing image: 4094/5000\n",
      "Processing image: 4095/5000\n",
      "Processing image: 4096/5000\n",
      "Processing image: 4097/5000\n",
      "Processing image: 4098/5000\n",
      "Processing image: 4099/5000\n",
      "Processing image: 4100/5000\n",
      "Processing image: 4101/5000\n",
      "Processing image: 4102/5000\n",
      "Processing image: 4103/5000\n",
      "Processing image: 4104/5000\n",
      "Processing image: 4105/5000\n",
      "Processing image: 4106/5000\n",
      "Processing image: 4107/5000\n",
      "Processing image: 4108/5000\n",
      "Processing image: 4109/5000\n",
      "Processing image: 4110/5000\n",
      "Processing image: 4111/5000\n",
      "Processing image: 4112/5000\n",
      "Processing image: 4113/5000\n",
      "Processing image: 4114/5000\n",
      "Processing image: 4115/5000\n",
      "Processing image: 4116/5000\n",
      "Processing image: 4117/5000\n",
      "Processing image: 4118/5000\n",
      "Processing image: 4119/5000\n",
      "Processing image: 4120/5000\n",
      "Processing image: 4121/5000\n",
      "Processing image: 4122/5000\n",
      "Processing image: 4123/5000\n",
      "Processing image: 4124/5000\n",
      "Processing image: 4125/5000\n",
      "Processing image: 4126/5000\n",
      "Processing image: 4127/5000\n",
      "Processing image: 4128/5000\n",
      "Processing image: 4129/5000\n",
      "Processing image: 4130/5000\n",
      "Processing image: 4131/5000\n",
      "Processing image: 4132/5000\n",
      "Processing image: 4133/5000\n",
      "Processing image: 4134/5000\n",
      "Processing image: 4135/5000\n",
      "Processing image: 4136/5000\n",
      "Processing image: 4137/5000\n",
      "Processing image: 4138/5000\n",
      "Processing image: 4139/5000\n",
      "Processing image: 4140/5000\n",
      "Processing image: 4141/5000\n",
      "Processing image: 4142/5000\n",
      "Processing image: 4143/5000\n",
      "Processing image: 4144/5000\n",
      "Processing image: 4145/5000\n",
      "Processing image: 4146/5000\n",
      "Processing image: 4147/5000\n",
      "Processing image: 4148/5000\n",
      "Processing image: 4149/5000\n",
      "Processing image: 4150/5000\n",
      "Processing image: 4151/5000\n",
      "Processing image: 4152/5000\n",
      "Processing image: 4153/5000\n",
      "Processing image: 4154/5000\n",
      "Processing image: 4155/5000\n",
      "Processing image: 4156/5000\n",
      "Processing image: 4157/5000\n",
      "Processing image: 4158/5000\n",
      "Processing image: 4159/5000\n",
      "Processing image: 4160/5000\n",
      "Processing image: 4161/5000\n",
      "Processing image: 4162/5000\n",
      "Processing image: 4163/5000\n",
      "Processing image: 4164/5000\n",
      "Processing image: 4165/5000\n",
      "Processing image: 4166/5000\n",
      "Processing image: 4167/5000\n",
      "Processing image: 4168/5000\n",
      "Processing image: 4169/5000\n",
      "Processing image: 4170/5000\n",
      "Processing image: 4171/5000\n",
      "Processing image: 4172/5000\n",
      "Processing image: 4173/5000\n",
      "Processing image: 4174/5000\n",
      "Processing image: 4175/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 4176/5000\n",
      "Processing image: 4177/5000\n",
      "Processing image: 4178/5000\n",
      "Processing image: 4179/5000\n",
      "Processing image: 4180/5000\n",
      "Processing image: 4181/5000\n",
      "Processing image: 4182/5000\n",
      "Processing image: 4183/5000\n",
      "Processing image: 4184/5000\n",
      "Processing image: 4185/5000\n",
      "Processing image: 4186/5000\n",
      "Processing image: 4187/5000\n",
      "Processing image: 4188/5000\n",
      "Processing image: 4189/5000\n",
      "Processing image: 4190/5000\n",
      "Processing image: 4191/5000\n",
      "Processing image: 4192/5000\n",
      "Processing image: 4193/5000\n",
      "Processing image: 4194/5000\n",
      "Processing image: 4195/5000\n",
      "Processing image: 4196/5000\n",
      "Processing image: 4197/5000\n",
      "Processing image: 4198/5000\n",
      "Processing image: 4199/5000\n",
      "Processing image: 4200/5000\n",
      "Processing image: 4201/5000\n",
      "Processing image: 4202/5000\n",
      "Processing image: 4203/5000\n",
      "Processing image: 4204/5000\n",
      "Processing image: 4205/5000\n",
      "Processing image: 4206/5000\n",
      "Processing image: 4207/5000\n",
      "Processing image: 4208/5000\n",
      "Processing image: 4209/5000\n",
      "Processing image: 4210/5000\n",
      "Processing image: 4211/5000\n",
      "Processing image: 4212/5000\n",
      "Processing image: 4213/5000\n",
      "Processing image: 4214/5000\n",
      "Processing image: 4215/5000\n",
      "Processing image: 4216/5000\n",
      "Processing image: 4217/5000\n",
      "Processing image: 4218/5000\n",
      "Processing image: 4219/5000\n",
      "Processing image: 4220/5000\n",
      "Processing image: 4221/5000\n",
      "Processing image: 4222/5000\n",
      "Processing image: 4223/5000\n",
      "Processing image: 4224/5000\n",
      "Processing image: 4225/5000\n",
      "Processing image: 4226/5000\n",
      "Processing image: 4227/5000\n",
      "Processing image: 4228/5000\n",
      "Processing image: 4229/5000\n",
      "Processing image: 4230/5000\n",
      "Processing image: 4231/5000\n",
      "Processing image: 4232/5000\n",
      "Processing image: 4233/5000\n",
      "Processing image: 4234/5000\n",
      "Processing image: 4235/5000\n",
      "Processing image: 4236/5000\n",
      "Processing image: 4237/5000\n",
      "Processing image: 4238/5000\n",
      "Processing image: 4239/5000\n",
      "Processing image: 4240/5000\n",
      "Processing image: 4241/5000\n",
      "Processing image: 4242/5000\n",
      "Processing image: 4243/5000\n",
      "Processing image: 4244/5000\n",
      "Processing image: 4245/5000\n",
      "Processing image: 4246/5000\n",
      "Processing image: 4247/5000\n",
      "Processing image: 4248/5000\n",
      "Processing image: 4249/5000\n",
      "Processing image: 4250/5000\n",
      "Processing image: 4251/5000\n",
      "Processing image: 4252/5000\n",
      "Processing image: 4253/5000\n",
      "Processing image: 4254/5000\n",
      "Processing image: 4255/5000\n",
      "Processing image: 4256/5000\n",
      "Processing image: 4257/5000\n",
      "Processing image: 4258/5000\n",
      "Processing image: 4259/5000\n",
      "Processing image: 4260/5000\n",
      "Processing image: 4261/5000\n",
      "Processing image: 4262/5000\n",
      "Processing image: 4263/5000\n",
      "Processing image: 4264/5000\n",
      "Processing image: 4265/5000\n",
      "Processing image: 4266/5000\n",
      "Processing image: 4267/5000\n",
      "Processing image: 4268/5000\n",
      "Processing image: 4269/5000\n",
      "Processing image: 4270/5000\n",
      "Processing image: 4271/5000\n",
      "Processing image: 4272/5000\n",
      "Processing image: 4273/5000\n",
      "Processing image: 4274/5000\n",
      "Processing image: 4275/5000\n",
      "Processing image: 4276/5000\n",
      "Processing image: 4277/5000\n",
      "Processing image: 4278/5000\n",
      "Processing image: 4279/5000\n",
      "Processing image: 4280/5000\n",
      "Processing image: 4281/5000\n",
      "Processing image: 4282/5000\n",
      "Processing image: 4283/5000\n",
      "Processing image: 4284/5000\n",
      "Processing image: 4285/5000\n",
      "Processing image: 4286/5000\n",
      "Processing image: 4287/5000\n",
      "Processing image: 4288/5000\n",
      "Processing image: 4289/5000\n",
      "Processing image: 4290/5000\n",
      "Processing image: 4291/5000\n",
      "Processing image: 4292/5000\n",
      "Processing image: 4293/5000\n",
      "Processing image: 4294/5000\n",
      "Processing image: 4295/5000\n",
      "Processing image: 4296/5000\n",
      "Processing image: 4297/5000\n",
      "Processing image: 4298/5000\n",
      "Processing image: 4299/5000\n",
      "Processing image: 4300/5000\n",
      "Processing image: 4301/5000\n",
      "Processing image: 4302/5000\n",
      "Processing image: 4303/5000\n",
      "Processing image: 4304/5000\n",
      "Processing image: 4305/5000\n",
      "Processing image: 4306/5000\n",
      "Processing image: 4307/5000\n",
      "Processing image: 4308/5000\n",
      "Processing image: 4309/5000\n",
      "Processing image: 4310/5000\n",
      "Processing image: 4311/5000\n",
      "Processing image: 4312/5000\n",
      "Processing image: 4313/5000\n",
      "Processing image: 4314/5000\n",
      "Processing image: 4315/5000\n",
      "Processing image: 4316/5000\n",
      "Processing image: 4317/5000\n",
      "Processing image: 4318/5000\n",
      "Processing image: 4319/5000\n",
      "Processing image: 4320/5000\n",
      "Processing image: 4321/5000\n",
      "Processing image: 4322/5000\n",
      "Processing image: 4323/5000\n",
      "Processing image: 4324/5000\n",
      "Processing image: 4325/5000\n",
      "Processing image: 4326/5000\n",
      "Processing image: 4327/5000\n",
      "Processing image: 4328/5000\n",
      "Processing image: 4329/5000\n",
      "Processing image: 4330/5000\n",
      "Processing image: 4331/5000\n",
      "Processing image: 4332/5000\n",
      "Processing image: 4333/5000\n",
      "Processing image: 4334/5000\n",
      "Processing image: 4335/5000\n",
      "Processing image: 4336/5000\n",
      "Processing image: 4337/5000\n",
      "Processing image: 4338/5000\n",
      "Processing image: 4339/5000\n",
      "Processing image: 4340/5000\n",
      "Processing image: 4341/5000\n",
      "Processing image: 4342/5000\n",
      "Processing image: 4343/5000\n",
      "Processing image: 4344/5000\n",
      "Processing image: 4345/5000\n",
      "Processing image: 4346/5000\n",
      "Processing image: 4347/5000\n",
      "Processing image: 4348/5000\n",
      "Processing image: 4349/5000\n",
      "Processing image: 4350/5000\n",
      "Processing image: 4351/5000\n",
      "Processing image: 4352/5000\n",
      "Processing image: 4353/5000\n",
      "Processing image: 4354/5000\n",
      "Processing image: 4355/5000\n",
      "Processing image: 4356/5000\n",
      "Processing image: 4357/5000\n",
      "Processing image: 4358/5000\n",
      "Processing image: 4359/5000\n",
      "Processing image: 4360/5000\n",
      "Processing image: 4361/5000\n",
      "Processing image: 4362/5000\n",
      "Processing image: 4363/5000\n",
      "Processing image: 4364/5000\n",
      "Processing image: 4365/5000\n",
      "Processing image: 4366/5000\n",
      "Processing image: 4367/5000\n",
      "Processing image: 4368/5000\n",
      "Processing image: 4369/5000\n",
      "Processing image: 4370/5000\n",
      "Processing image: 4371/5000\n",
      "Processing image: 4372/5000\n",
      "Processing image: 4373/5000\n",
      "Processing image: 4374/5000\n",
      "Processing image: 4375/5000\n",
      "Processing image: 4376/5000\n",
      "Processing image: 4377/5000\n",
      "Processing image: 4378/5000\n",
      "Processing image: 4379/5000\n",
      "Processing image: 4380/5000\n",
      "Processing image: 4381/5000\n",
      "Processing image: 4382/5000\n",
      "Processing image: 4383/5000\n",
      "Processing image: 4384/5000\n",
      "Processing image: 4385/5000\n",
      "Processing image: 4386/5000\n",
      "Processing image: 4387/5000\n",
      "Processing image: 4388/5000\n",
      "Processing image: 4389/5000\n",
      "Processing image: 4390/5000\n",
      "Processing image: 4391/5000\n",
      "Processing image: 4392/5000\n",
      "Processing image: 4393/5000\n",
      "Processing image: 4394/5000\n",
      "Processing image: 4395/5000\n",
      "Processing image: 4396/5000\n",
      "Processing image: 4397/5000\n",
      "Processing image: 4398/5000\n",
      "Processing image: 4399/5000\n",
      "Processing image: 4400/5000\n",
      "Processing image: 4401/5000\n",
      "Processing image: 4402/5000\n",
      "Processing image: 4403/5000\n",
      "Processing image: 4404/5000\n",
      "Processing image: 4405/5000\n",
      "Processing image: 4406/5000\n",
      "Processing image: 4407/5000\n",
      "Processing image: 4408/5000\n",
      "Processing image: 4409/5000\n",
      "Processing image: 4410/5000\n",
      "Processing image: 4411/5000\n",
      "Processing image: 4412/5000\n",
      "Processing image: 4413/5000\n",
      "Processing image: 4414/5000\n",
      "Processing image: 4415/5000\n",
      "Processing image: 4416/5000\n",
      "Processing image: 4417/5000\n",
      "Processing image: 4418/5000\n",
      "Processing image: 4419/5000\n",
      "Processing image: 4420/5000\n",
      "Processing image: 4421/5000\n",
      "Processing image: 4422/5000\n",
      "Processing image: 4423/5000\n",
      "Processing image: 4424/5000\n",
      "Processing image: 4425/5000\n",
      "Processing image: 4426/5000\n",
      "Processing image: 4427/5000\n",
      "Processing image: 4428/5000\n",
      "Processing image: 4429/5000\n",
      "Processing image: 4430/5000\n",
      "Processing image: 4431/5000\n",
      "Processing image: 4432/5000\n",
      "Processing image: 4433/5000\n",
      "Processing image: 4434/5000\n",
      "Processing image: 4435/5000\n",
      "Processing image: 4436/5000\n",
      "Processing image: 4437/5000\n",
      "Processing image: 4438/5000\n",
      "Processing image: 4439/5000\n",
      "Processing image: 4440/5000\n",
      "Processing image: 4441/5000\n",
      "Processing image: 4442/5000\n",
      "Processing image: 4443/5000\n",
      "Processing image: 4444/5000\n",
      "Processing image: 4445/5000\n",
      "Processing image: 4446/5000\n",
      "Processing image: 4447/5000\n",
      "Processing image: 4448/5000\n",
      "Processing image: 4449/5000\n",
      "Processing image: 4450/5000\n",
      "Processing image: 4451/5000\n",
      "Processing image: 4452/5000\n",
      "Processing image: 4453/5000\n",
      "Processing image: 4454/5000\n",
      "Processing image: 4455/5000\n",
      "Processing image: 4456/5000\n",
      "Processing image: 4457/5000\n",
      "Processing image: 4458/5000\n",
      "Processing image: 4459/5000\n",
      "Processing image: 4460/5000\n",
      "Processing image: 4461/5000\n",
      "Processing image: 4462/5000\n",
      "Processing image: 4463/5000\n",
      "Processing image: 4464/5000\n",
      "Processing image: 4465/5000\n",
      "Processing image: 4466/5000\n",
      "Processing image: 4467/5000\n",
      "Processing image: 4468/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 4469/5000\n",
      "Processing image: 4470/5000\n",
      "Processing image: 4471/5000\n",
      "Processing image: 4472/5000\n",
      "Processing image: 4473/5000\n",
      "Processing image: 4474/5000\n",
      "Processing image: 4475/5000\n",
      "Processing image: 4476/5000\n",
      "Processing image: 4477/5000\n",
      "Processing image: 4478/5000\n",
      "Processing image: 4479/5000\n",
      "Processing image: 4480/5000\n",
      "Processing image: 4481/5000\n",
      "Processing image: 4482/5000\n",
      "Processing image: 4483/5000\n",
      "Processing image: 4484/5000\n",
      "Processing image: 4485/5000\n",
      "Processing image: 4486/5000\n",
      "Processing image: 4487/5000\n",
      "Processing image: 4488/5000\n",
      "Processing image: 4489/5000\n",
      "Processing image: 4490/5000\n",
      "Processing image: 4491/5000\n",
      "Processing image: 4492/5000\n",
      "Processing image: 4493/5000\n",
      "Processing image: 4494/5000\n",
      "Processing image: 4495/5000\n",
      "Processing image: 4496/5000\n",
      "Processing image: 4497/5000\n",
      "Processing image: 4498/5000\n",
      "Processing image: 4499/5000\n",
      "Processing image: 4500/5000\n",
      "Processing image: 4501/5000\n",
      "Processing image: 4502/5000\n",
      "Processing image: 4503/5000\n",
      "Processing image: 4504/5000\n",
      "Processing image: 4505/5000\n",
      "Processing image: 4506/5000\n",
      "Processing image: 4507/5000\n",
      "Processing image: 4508/5000\n",
      "Processing image: 4509/5000\n",
      "Processing image: 4510/5000\n",
      "Processing image: 4511/5000\n",
      "Processing image: 4512/5000\n",
      "Processing image: 4513/5000\n",
      "Processing image: 4514/5000\n",
      "Processing image: 4515/5000\n",
      "Processing image: 4516/5000\n",
      "Processing image: 4517/5000\n",
      "Processing image: 4518/5000\n",
      "Processing image: 4519/5000\n",
      "Processing image: 4520/5000\n",
      "Processing image: 4521/5000\n",
      "Processing image: 4522/5000\n",
      "Processing image: 4523/5000\n",
      "Processing image: 4524/5000\n",
      "Processing image: 4525/5000\n",
      "Processing image: 4526/5000\n",
      "Processing image: 4527/5000\n",
      "Processing image: 4528/5000\n",
      "Processing image: 4529/5000\n",
      "Processing image: 4530/5000\n",
      "Processing image: 4531/5000\n",
      "Processing image: 4532/5000\n",
      "Processing image: 4533/5000\n",
      "Processing image: 4534/5000\n",
      "Processing image: 4535/5000\n",
      "Processing image: 4536/5000\n",
      "Processing image: 4537/5000\n",
      "Processing image: 4538/5000\n",
      "Processing image: 4539/5000\n",
      "Processing image: 4540/5000\n",
      "Processing image: 4541/5000\n",
      "Processing image: 4542/5000\n",
      "Processing image: 4543/5000\n",
      "Processing image: 4544/5000\n",
      "Processing image: 4545/5000\n",
      "Processing image: 4546/5000\n",
      "Processing image: 4547/5000\n",
      "Processing image: 4548/5000\n",
      "Processing image: 4549/5000\n",
      "Processing image: 4550/5000\n",
      "Processing image: 4551/5000\n",
      "Processing image: 4552/5000\n",
      "Processing image: 4553/5000\n",
      "Processing image: 4554/5000\n",
      "Processing image: 4555/5000\n",
      "Processing image: 4556/5000\n",
      "Processing image: 4557/5000\n",
      "Processing image: 4558/5000\n",
      "Processing image: 4559/5000\n",
      "Processing image: 4560/5000\n",
      "Processing image: 4561/5000\n",
      "Processing image: 4562/5000\n",
      "Processing image: 4563/5000\n",
      "Processing image: 4564/5000\n",
      "Processing image: 4565/5000\n",
      "Processing image: 4566/5000\n",
      "Processing image: 4567/5000\n",
      "Processing image: 4568/5000\n",
      "Processing image: 4569/5000\n",
      "Processing image: 4570/5000\n",
      "Processing image: 4571/5000\n",
      "Processing image: 4572/5000\n",
      "Processing image: 4573/5000\n",
      "Processing image: 4574/5000\n",
      "Processing image: 4575/5000\n",
      "Processing image: 4576/5000\n",
      "Processing image: 4577/5000\n",
      "Processing image: 4578/5000\n",
      "Processing image: 4579/5000\n",
      "Processing image: 4580/5000\n",
      "Processing image: 4581/5000\n",
      "Processing image: 4582/5000\n",
      "Processing image: 4583/5000\n",
      "Processing image: 4584/5000\n",
      "Processing image: 4585/5000\n",
      "Processing image: 4586/5000\n",
      "Processing image: 4587/5000\n",
      "Processing image: 4588/5000\n",
      "Processing image: 4589/5000\n",
      "Processing image: 4590/5000\n",
      "Processing image: 4591/5000\n",
      "Processing image: 4592/5000\n",
      "Processing image: 4593/5000\n",
      "Processing image: 4594/5000\n",
      "Processing image: 4595/5000\n",
      "Processing image: 4596/5000\n",
      "Processing image: 4597/5000\n",
      "Processing image: 4598/5000\n",
      "Processing image: 4599/5000\n",
      "Processing image: 4600/5000\n",
      "Processing image: 4601/5000\n",
      "Processing image: 4602/5000\n",
      "Processing image: 4603/5000\n",
      "Processing image: 4604/5000\n",
      "Processing image: 4605/5000\n",
      "Processing image: 4606/5000\n",
      "Processing image: 4607/5000\n",
      "Processing image: 4608/5000\n",
      "Processing image: 4609/5000\n",
      "Processing image: 4610/5000\n",
      "Processing image: 4611/5000\n",
      "Processing image: 4612/5000\n",
      "Processing image: 4613/5000\n",
      "Processing image: 4614/5000\n",
      "Processing image: 4615/5000\n",
      "Processing image: 4616/5000\n",
      "Processing image: 4617/5000\n",
      "Processing image: 4618/5000\n",
      "Processing image: 4619/5000\n",
      "Processing image: 4620/5000\n",
      "Processing image: 4621/5000\n",
      "Processing image: 4622/5000\n",
      "Processing image: 4623/5000\n",
      "Processing image: 4624/5000\n",
      "Processing image: 4625/5000\n",
      "Processing image: 4626/5000\n",
      "Processing image: 4627/5000\n",
      "Processing image: 4628/5000\n",
      "Processing image: 4629/5000\n",
      "Processing image: 4630/5000\n",
      "Processing image: 4631/5000\n",
      "Processing image: 4632/5000\n",
      "Processing image: 4633/5000\n",
      "Processing image: 4634/5000\n",
      "Processing image: 4635/5000\n",
      "Processing image: 4636/5000\n",
      "Processing image: 4637/5000\n",
      "Processing image: 4638/5000\n",
      "Processing image: 4639/5000\n",
      "Processing image: 4640/5000\n",
      "Processing image: 4641/5000\n",
      "Processing image: 4642/5000\n",
      "Processing image: 4643/5000\n",
      "Processing image: 4644/5000\n",
      "Processing image: 4645/5000\n",
      "Processing image: 4646/5000\n",
      "Processing image: 4647/5000\n",
      "Processing image: 4648/5000\n",
      "Processing image: 4649/5000\n",
      "Processing image: 4650/5000\n",
      "Processing image: 4651/5000\n",
      "Processing image: 4652/5000\n",
      "Processing image: 4653/5000\n",
      "Processing image: 4654/5000\n",
      "Processing image: 4655/5000\n",
      "Processing image: 4656/5000\n",
      "Processing image: 4657/5000\n",
      "Processing image: 4658/5000\n",
      "Processing image: 4659/5000\n",
      "Processing image: 4660/5000\n",
      "Processing image: 4661/5000\n",
      "Processing image: 4662/5000\n",
      "Processing image: 4663/5000\n",
      "Processing image: 4664/5000\n",
      "Processing image: 4665/5000\n",
      "Processing image: 4666/5000\n",
      "Processing image: 4667/5000\n",
      "Processing image: 4668/5000\n",
      "Processing image: 4669/5000\n",
      "Processing image: 4670/5000\n",
      "Processing image: 4671/5000\n",
      "Processing image: 4672/5000\n",
      "Processing image: 4673/5000\n",
      "Processing image: 4674/5000\n",
      "Processing image: 4675/5000\n",
      "Processing image: 4676/5000\n",
      "Processing image: 4677/5000\n",
      "Processing image: 4678/5000\n",
      "Processing image: 4679/5000\n",
      "Processing image: 4680/5000\n",
      "Processing image: 4681/5000\n",
      "Processing image: 4682/5000\n",
      "Processing image: 4683/5000\n",
      "Processing image: 4684/5000\n",
      "Processing image: 4685/5000\n",
      "Processing image: 4686/5000\n",
      "Processing image: 4687/5000\n",
      "Processing image: 4688/5000\n",
      "Processing image: 4689/5000\n",
      "Processing image: 4690/5000\n",
      "Processing image: 4691/5000\n",
      "Processing image: 4692/5000\n",
      "Processing image: 4693/5000\n",
      "Processing image: 4694/5000\n",
      "Processing image: 4695/5000\n",
      "Processing image: 4696/5000\n",
      "Processing image: 4697/5000\n",
      "Processing image: 4698/5000\n",
      "Processing image: 4699/5000\n",
      "Processing image: 4700/5000\n",
      "Processing image: 4701/5000\n",
      "Processing image: 4702/5000\n",
      "Processing image: 4703/5000\n",
      "Processing image: 4704/5000\n",
      "Processing image: 4705/5000\n",
      "Processing image: 4706/5000\n",
      "Processing image: 4707/5000\n",
      "Processing image: 4708/5000\n",
      "Processing image: 4709/5000\n",
      "Processing image: 4710/5000\n",
      "Processing image: 4711/5000\n",
      "Processing image: 4712/5000\n",
      "Processing image: 4713/5000\n",
      "Processing image: 4714/5000\n",
      "Processing image: 4715/5000\n",
      "Processing image: 4716/5000\n",
      "Processing image: 4717/5000\n",
      "Processing image: 4718/5000\n",
      "Processing image: 4719/5000\n",
      "Processing image: 4720/5000\n",
      "Processing image: 4721/5000\n",
      "Processing image: 4722/5000\n",
      "Processing image: 4723/5000\n",
      "Processing image: 4724/5000\n",
      "Processing image: 4725/5000\n",
      "Processing image: 4726/5000\n",
      "Processing image: 4727/5000\n",
      "Processing image: 4728/5000\n",
      "Processing image: 4729/5000\n",
      "Processing image: 4730/5000\n",
      "Processing image: 4731/5000\n",
      "Processing image: 4732/5000\n",
      "Processing image: 4733/5000\n",
      "Processing image: 4734/5000\n",
      "Processing image: 4735/5000\n",
      "Processing image: 4736/5000\n",
      "Processing image: 4737/5000\n",
      "Processing image: 4738/5000\n",
      "Processing image: 4739/5000\n",
      "Processing image: 4740/5000\n",
      "Processing image: 4741/5000\n",
      "Processing image: 4742/5000\n",
      "Processing image: 4743/5000\n",
      "Processing image: 4744/5000\n",
      "Processing image: 4745/5000\n",
      "Processing image: 4746/5000\n",
      "Processing image: 4747/5000\n",
      "Processing image: 4748/5000\n",
      "Processing image: 4749/5000\n",
      "Processing image: 4750/5000\n",
      "Processing image: 4751/5000\n",
      "Processing image: 4752/5000\n",
      "Processing image: 4753/5000\n",
      "Processing image: 4754/5000\n",
      "Processing image: 4755/5000\n",
      "Processing image: 4756/5000\n",
      "Processing image: 4757/5000\n",
      "Processing image: 4758/5000\n",
      "Processing image: 4759/5000\n",
      "Processing image: 4760/5000\n",
      "Processing image: 4761/5000\n",
      "Processing image: 4762/5000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Processing image: 4763/5000\n",
      "Processing image: 4764/5000\n",
      "Processing image: 4765/5000\n",
      "Processing image: 4766/5000\n",
      "Processing image: 4767/5000\n",
      "Processing image: 4768/5000\n",
      "Processing image: 4769/5000\n",
      "Processing image: 4770/5000\n",
      "Processing image: 4771/5000\n",
      "Processing image: 4772/5000\n",
      "Processing image: 4773/5000\n",
      "Processing image: 4774/5000\n",
      "Processing image: 4775/5000\n",
      "Processing image: 4776/5000\n",
      "Processing image: 4777/5000\n",
      "Processing image: 4778/5000\n",
      "Processing image: 4779/5000\n",
      "Processing image: 4780/5000\n",
      "Processing image: 4781/5000\n",
      "Processing image: 4782/5000\n",
      "Processing image: 4783/5000\n",
      "Processing image: 4784/5000\n",
      "Processing image: 4785/5000\n",
      "Processing image: 4786/5000\n",
      "Processing image: 4787/5000\n",
      "Processing image: 4788/5000\n",
      "Processing image: 4789/5000\n",
      "Processing image: 4790/5000\n",
      "Processing image: 4791/5000\n",
      "Processing image: 4792/5000\n",
      "Processing image: 4793/5000\n",
      "Processing image: 4794/5000\n",
      "Processing image: 4795/5000\n",
      "Processing image: 4796/5000\n",
      "Processing image: 4797/5000\n",
      "Processing image: 4798/5000\n",
      "Processing image: 4799/5000\n",
      "Processing image: 4800/5000\n",
      "Processing image: 4801/5000\n",
      "Processing image: 4802/5000\n",
      "Processing image: 4803/5000\n",
      "Processing image: 4804/5000\n",
      "Processing image: 4805/5000\n",
      "Processing image: 4806/5000\n",
      "Processing image: 4807/5000\n",
      "Processing image: 4808/5000\n",
      "Processing image: 4809/5000\n",
      "Processing image: 4810/5000\n",
      "Processing image: 4811/5000\n",
      "Processing image: 4812/5000\n",
      "Processing image: 4813/5000\n",
      "Processing image: 4814/5000\n",
      "Processing image: 4815/5000\n",
      "Processing image: 4816/5000\n",
      "Processing image: 4817/5000\n",
      "Processing image: 4818/5000\n",
      "Processing image: 4819/5000\n",
      "Processing image: 4820/5000\n",
      "Processing image: 4821/5000\n",
      "Processing image: 4822/5000\n",
      "Processing image: 4823/5000\n",
      "Processing image: 4824/5000\n",
      "Processing image: 4825/5000\n",
      "Processing image: 4826/5000\n",
      "Processing image: 4827/5000\n",
      "Processing image: 4828/5000\n",
      "Processing image: 4829/5000\n",
      "Processing image: 4830/5000\n",
      "Processing image: 4831/5000\n",
      "Processing image: 4832/5000\n",
      "Processing image: 4833/5000\n",
      "Processing image: 4834/5000\n",
      "Processing image: 4835/5000\n",
      "Processing image: 4836/5000\n",
      "Processing image: 4837/5000\n",
      "Processing image: 4838/5000\n",
      "Processing image: 4839/5000\n",
      "Processing image: 4840/5000\n",
      "Processing image: 4841/5000\n",
      "Processing image: 4842/5000\n",
      "Processing image: 4843/5000\n",
      "Processing image: 4844/5000\n",
      "Processing image: 4845/5000\n",
      "Processing image: 4846/5000\n",
      "Processing image: 4847/5000\n",
      "Processing image: 4848/5000\n",
      "Processing image: 4849/5000\n",
      "Processing image: 4850/5000\n",
      "Processing image: 4851/5000\n",
      "Processing image: 4852/5000\n",
      "Processing image: 4853/5000\n",
      "Processing image: 4854/5000\n",
      "Processing image: 4855/5000\n",
      "Processing image: 4856/5000\n",
      "Processing image: 4857/5000\n",
      "Processing image: 4858/5000\n",
      "Processing image: 4859/5000\n",
      "Processing image: 4860/5000\n",
      "Processing image: 4861/5000\n",
      "Processing image: 4862/5000\n",
      "Processing image: 4863/5000\n",
      "Processing image: 4864/5000\n",
      "Processing image: 4865/5000\n",
      "Processing image: 4866/5000\n",
      "Processing image: 4867/5000\n",
      "Processing image: 4868/5000\n",
      "Processing image: 4869/5000\n",
      "Processing image: 4870/5000\n",
      "Processing image: 4871/5000\n",
      "Processing image: 4872/5000\n",
      "Processing image: 4873/5000\n",
      "Processing image: 4874/5000\n",
      "Processing image: 4875/5000\n",
      "Processing image: 4876/5000\n",
      "Processing image: 4877/5000\n",
      "Processing image: 4878/5000\n",
      "Processing image: 4879/5000\n",
      "Processing image: 4880/5000\n",
      "Processing image: 4881/5000\n",
      "Processing image: 4882/5000\n",
      "Processing image: 4883/5000\n",
      "Processing image: 4884/5000\n",
      "Processing image: 4885/5000\n",
      "Processing image: 4886/5000\n",
      "Processing image: 4887/5000\n",
      "Processing image: 4888/5000\n",
      "Processing image: 4889/5000\n",
      "Processing image: 4890/5000\n",
      "Processing image: 4891/5000\n",
      "Processing image: 4892/5000\n",
      "Processing image: 4893/5000\n",
      "Processing image: 4894/5000\n",
      "Processing image: 4895/5000\n",
      "Processing image: 4896/5000\n",
      "Processing image: 4897/5000\n",
      "Processing image: 4898/5000\n",
      "Processing image: 4899/5000\n",
      "Processing image: 4900/5000\n",
      "Processing image: 4901/5000\n",
      "Processing image: 4902/5000\n",
      "Processing image: 4903/5000\n",
      "Processing image: 4904/5000\n",
      "Processing image: 4905/5000\n",
      "Processing image: 4906/5000\n",
      "Processing image: 4907/5000\n",
      "Processing image: 4908/5000\n",
      "Processing image: 4909/5000\n",
      "Processing image: 4910/5000\n",
      "Processing image: 4911/5000\n",
      "Processing image: 4912/5000\n",
      "Processing image: 4913/5000\n",
      "Processing image: 4914/5000\n",
      "Processing image: 4915/5000\n",
      "Processing image: 4916/5000\n",
      "Processing image: 4917/5000\n",
      "Processing image: 4918/5000\n",
      "Processing image: 4919/5000\n",
      "Processing image: 4920/5000\n",
      "Processing image: 4921/5000\n",
      "Processing image: 4922/5000\n",
      "Processing image: 4923/5000\n",
      "Processing image: 4924/5000\n",
      "Processing image: 4925/5000\n",
      "Processing image: 4926/5000\n",
      "Processing image: 4927/5000\n",
      "Processing image: 4928/5000\n",
      "Processing image: 4929/5000\n",
      "Processing image: 4930/5000\n",
      "Processing image: 4931/5000\n",
      "Processing image: 4932/5000\n",
      "Processing image: 4933/5000\n",
      "Processing image: 4934/5000\n",
      "Processing image: 4935/5000\n",
      "Processing image: 4936/5000\n",
      "Processing image: 4937/5000\n",
      "Processing image: 4938/5000\n",
      "Processing image: 4939/5000\n",
      "Processing image: 4940/5000\n",
      "Processing image: 4941/5000\n",
      "Processing image: 4942/5000\n",
      "Processing image: 4943/5000\n",
      "Processing image: 4944/5000\n",
      "Processing image: 4945/5000\n",
      "Processing image: 4946/5000\n",
      "Processing image: 4947/5000\n",
      "Processing image: 4948/5000\n",
      "Processing image: 4949/5000\n",
      "Processing image: 4950/5000\n",
      "Processing image: 4951/5000\n",
      "Processing image: 4952/5000\n",
      "Processing image: 4953/5000\n",
      "Processing image: 4954/5000\n",
      "Processing image: 4955/5000\n",
      "Processing image: 4956/5000\n",
      "Processing image: 4957/5000\n",
      "Processing image: 4958/5000\n",
      "Processing image: 4959/5000\n",
      "Processing image: 4960/5000\n",
      "Processing image: 4961/5000\n",
      "Processing image: 4962/5000\n",
      "Processing image: 4963/5000\n",
      "Processing image: 4964/5000\n",
      "Processing image: 4965/5000\n",
      "Processing image: 4966/5000\n",
      "Processing image: 4967/5000\n",
      "Processing image: 4968/5000\n",
      "Processing image: 4969/5000\n",
      "Processing image: 4970/5000\n",
      "Processing image: 4971/5000\n",
      "Processing image: 4972/5000\n",
      "Processing image: 4973/5000\n",
      "Processing image: 4974/5000\n",
      "Processing image: 4975/5000\n",
      "Processing image: 4976/5000\n",
      "Processing image: 4977/5000\n",
      "Processing image: 4978/5000\n",
      "Processing image: 4979/5000\n",
      "Processing image: 4980/5000\n",
      "Processing image: 4981/5000\n",
      "Processing image: 4982/5000\n",
      "Processing image: 4983/5000\n",
      "Processing image: 4984/5000\n",
      "Processing image: 4985/5000\n",
      "Processing image: 4986/5000\n",
      "Processing image: 4987/5000\n",
      "Processing image: 4988/5000\n",
      "Processing image: 4989/5000\n",
      "Processing image: 4990/5000\n",
      "Processing image: 4991/5000\n",
      "Processing image: 4992/5000\n",
      "Processing image: 4993/5000\n",
      "Processing image: 4994/5000\n",
      "Processing image: 4995/5000\n",
      "Processing image: 4996/5000\n",
      "Processing image: 4997/5000\n",
      "Processing image: 4998/5000\n",
      "Processing image: 4999/5000\n",
      "Processing image: 5000/5000\n"
     ]
    }
   ],
   "source": [
    "evaluate(ic_paths, ic_res_path, model, device, img_size, con_thr, iou_thr)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}