Spineweb_MANet_smp.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/wangjie/anaconda3/envs/dataEngineering/lib/python3.11/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import nibabel as nib\n",
    "import matplotlib.pyplot as plt\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import Dataset\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision.transforms import ToTensor\n",
    "import torchmetrics.functional as evafunc\n",
    "import torchvision.transforms as transforms\n",
    "import numpy as np\n",
    "import torchvision.models as models\n",
    "from segmentation_models_pytorch.losses import DiceLoss, FocalLoss\n",
    "import segmentation_models_pytorch as smp\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"] = \"3\"\n",
    "from PIL import Image\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 自定义数据集类\n",
    "class SpineWeb15(Dataset):\n",
    "    def __init__(self, data_path,transform1=None,transform2=None):\n",
    "        self.data_path = data_path\n",
    "        self.transform1 = transform1\n",
    "        self.transform2 = transform2\n",
    "        \n",
    "    def __len__(self):\n",
    "        file_list = os.listdir(self.data_path)  # 列出文件夹中的所有文件和文件夹\n",
    "        file_count = len(file_list)  # 获取文件数量\n",
    "        return file_count\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        metadata = np.load(self.data_path+str(index)+\".npy\")\n",
    "        image = metadata[0]\n",
    "        label = metadata[1]\n",
    "\n",
    "        # 找到数组中的最小值和最大值\n",
    "        min_value = np.min(image)\n",
    "        max_value = np.max(image)\n",
    "\n",
    "        # # 将数组元素限制在0到255之间（类似于归一化）\n",
    "        image = (image - min_value) * (255 / (max_value - min_value))\n",
    "\n",
    "        image_data = Image.fromarray(image)\n",
    "        label_data = Image.fromarray(label)\n",
    "        # image_data = torch.from_numpy(image).float()\n",
    "        # label_data = torch.from_numpy(label).float()\n",
    "        \n",
    "        if self.transform1:\n",
    "            image_data = self.transform1(image_data)\n",
    "        if self.transform2:\n",
    "            label_data = self.transform2(label_data)\n",
    "        return image_data, label_data\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_pixel_accuracy(prediction, target):\n",
    "    # 将预测结果和目标标签进行逐像素比较，并计算正确预测像素的比例\n",
    "    correct_pixels = (prediction == target).sum().item()\n",
    "    total_pixels = target.numel()\n",
    "    return correct_pixels, total_pixels\n",
    "\n",
    "def calculate_iou(prediction, target):\n",
    "    intersection = torch.logical_and(prediction, target).sum()\n",
    "    union = torch.logical_or(prediction, target).sum()\n",
    "    iou = intersection / union\n",
    "    return iou.item()\n",
    "\n",
    "def calculate_dice(prediction, target):\n",
    "    intersection = torch.logical_and(prediction, target).sum()\n",
    "    dice = (2 * intersection) / (prediction.sum() + target.sum())\n",
    "    return dice.item()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Downloading: \"https://download.pytorch.org/models/resnet34-333f7ec4.pth\" to /home/wangjie/.cache/torch/hub/checkpoints/resnet34-333f7ec4.pth\n",
      "100%|██████████| 83.3M/83.3M [00:07<00:00, 11.3MB/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/200], Avg Loss: 0.01278, Avg IoU: 0.1392, Avg Dice: 0.1945,Test IoU: 0.4925, Test Dice: 0.6305, Train Acc: 0.9523, Test Acc: 0.9761\n",
      "Epoch [2/200], Avg Loss: 0.01117, Avg IoU: 0.6645, Avg Dice: 0.7700,Test IoU: 0.6343, Test Dice: 0.7532, Train Acc: 0.9864, Test Acc: 0.9833\n",
      "Epoch [3/200], Avg Loss: 0.01095, Avg IoU: 0.7378, Avg Dice: 0.8242,Test IoU: 0.7946, Test Dice: 0.8659, Train Acc: 0.9902, Test Acc: 0.9915\n",
      "Epoch [4/200], Avg Loss: 0.01089, Avg IoU: 0.7802, Avg Dice: 0.8554,Test IoU: 0.7245, Test Dice: 0.8210, Train Acc: 0.9923, Test Acc: 0.9885\n",
      "Epoch [5/200], Avg Loss: 0.01086, Avg IoU: 0.7927, Avg Dice: 0.8616,Test IoU: 0.7845, Test Dice: 0.8589, Train Acc: 0.9929, Test Acc: 0.9913\n",
      "Epoch [6/200], Avg Loss: 0.01084, Avg IoU: 0.8108, Avg Dice: 0.8727,Test IoU: 0.7745, Test Dice: 0.8558, Train Acc: 0.9939, Test Acc: 0.9907\n",
      "Epoch [7/200], Avg Loss: 0.01083, Avg IoU: 0.8192, Avg Dice: 0.8808,Test IoU: 0.8375, Test Dice: 0.8958, Train Acc: 0.9941, Test Acc: 0.9934\n",
      "Epoch [8/200], Avg Loss: 0.01082, Avg IoU: 0.8312, Avg Dice: 0.8912,Test IoU: 0.8091, Test Dice: 0.8803, Train Acc: 0.9946, Test Acc: 0.9922\n",
      "Epoch [9/200], Avg Loss: 0.01081, Avg IoU: 0.8440, Avg Dice: 0.8995,Test IoU: 0.8410, Test Dice: 0.8993, Train Acc: 0.9951, Test Acc: 0.9939\n",
      "Epoch [10/200], Avg Loss: 0.01081, Avg IoU: 0.8510, Avg Dice: 0.9043,Test IoU: 0.8380, Test Dice: 0.8976, Train Acc: 0.9953, Test Acc: 0.9938\n",
      "Epoch [11/200], Avg Loss: 0.01081, Avg IoU: 0.8526, Avg Dice: 0.9054,Test IoU: 0.8033, Test Dice: 0.8751, Train Acc: 0.9954, Test Acc: 0.9921\n",
      "Epoch [12/200], Avg Loss: 0.01081, Avg IoU: 0.8556, Avg Dice: 0.9071,Test IoU: 0.8484, Test Dice: 0.9045, Train Acc: 0.9955, Test Acc: 0.9942\n",
      "Epoch [13/200], Avg Loss: 0.01081, Avg IoU: 0.8549, Avg Dice: 0.9066,Test IoU: 0.8305, Test Dice: 0.8938, Train Acc: 0.9955, Test Acc: 0.9933\n",
      "Epoch [14/200], Avg Loss: 0.01080, Avg IoU: 0.8678, Avg Dice: 0.9155,Test IoU: 0.8502, Test Dice: 0.9055, Train Acc: 0.9960, Test Acc: 0.9943\n",
      "Epoch [15/200], Avg Loss: 0.01080, Avg IoU: 0.8687, Avg Dice: 0.9163,Test IoU: 0.8659, Test Dice: 0.9145, Train Acc: 0.9960, Test Acc: 0.9949\n",
      "Epoch [16/200], Avg Loss: 0.01080, Avg IoU: 0.8740, Avg Dice: 0.9193,Test IoU: 0.8658, Test Dice: 0.9140, Train Acc: 0.9962, Test Acc: 0.9950\n",
      "Epoch [17/200], Avg Loss: 0.01080, Avg IoU: 0.8690, Avg Dice: 0.9164,Test IoU: 0.8429, Test Dice: 0.9008, Train Acc: 0.9960, Test Acc: 0.9939\n",
      "Epoch [18/200], Avg Loss: 0.01080, Avg IoU: 0.8756, Avg Dice: 0.9206,Test IoU: 0.7925, Test Dice: 0.8707, Train Acc: 0.9963, Test Acc: 0.9914\n",
      "Epoch [19/200], Avg Loss: 0.01081, Avg IoU: 0.8524, Avg Dice: 0.9055,Test IoU: 0.8355, Test Dice: 0.8972, Train Acc: 0.9954, Test Acc: 0.9937\n",
      "Epoch [20/200], Avg Loss: 0.01080, Avg IoU: 0.8709, Avg Dice: 0.9170,Test IoU: 0.8539, Test Dice: 0.9079, Train Acc: 0.9961, Test Acc: 0.9944\n",
      "Epoch [21/200], Avg Loss: 0.01080, Avg IoU: 0.8817, Avg Dice: 0.9241,Test IoU: 0.8310, Test Dice: 0.8944, Train Acc: 0.9965, Test Acc: 0.9934\n",
      "Epoch [22/200], Avg Loss: 0.01080, Avg IoU: 0.8785, Avg Dice: 0.9222,Test IoU: 0.6646, Test Dice: 0.7802, Train Acc: 0.9964, Test Acc: 0.9842\n",
      "Epoch [23/200], Avg Loss: 0.01083, Avg IoU: 0.7883, Avg Dice: 0.8591,Test IoU: 0.8250, Test Dice: 0.8903, Train Acc: 0.9925, Test Acc: 0.9931\n",
      "Epoch [24/200], Avg Loss: 0.01081, Avg IoU: 0.8429, Avg Dice: 0.8986,Test IoU: 0.8426, Test Dice: 0.9004, Train Acc: 0.9950, Test Acc: 0.9939\n",
      "Epoch [25/200], Avg Loss: 0.01080, Avg IoU: 0.8565, Avg Dice: 0.9081,Test IoU: 0.8436, Test Dice: 0.9021, Train Acc: 0.9955, Test Acc: 0.9939\n",
      "Epoch [26/200], Avg Loss: 0.01080, Avg IoU: 0.8702, Avg Dice: 0.9167,Test IoU: 0.8559, Test Dice: 0.9094, Train Acc: 0.9961, Test Acc: 0.9946\n",
      "Epoch [27/200], Avg Loss: 0.01080, Avg IoU: 0.8771, Avg Dice: 0.9207,Test IoU: 0.8522, Test Dice: 0.9069, Train Acc: 0.9964, Test Acc: 0.9944\n",
      "Epoch [28/200], Avg Loss: 0.01079, Avg IoU: 0.8824, Avg Dice: 0.9241,Test IoU: 0.8677, Test Dice: 0.9158, Train Acc: 0.9966, Test Acc: 0.9951\n",
      "Epoch [29/200], Avg Loss: 0.01080, Avg IoU: 0.8824, Avg Dice: 0.9242,Test IoU: 0.8583, Test Dice: 0.9102, Train Acc: 0.9965, Test Acc: 0.9947\n",
      "Epoch [30/200], Avg Loss: 0.01080, Avg IoU: 0.8859, Avg Dice: 0.9262,Test IoU: 0.8520, Test Dice: 0.9065, Train Acc: 0.9967, Test Acc: 0.9944\n",
      "Epoch [31/200], Avg Loss: 0.01080, Avg IoU: 0.8788, Avg Dice: 0.9219,Test IoU: 0.8489, Test Dice: 0.9051, Train Acc: 0.9964, Test Acc: 0.9943\n",
      "Epoch [32/200], Avg Loss: 0.01079, Avg IoU: 0.8887, Avg Dice: 0.9278,Test IoU: 0.8714, Test Dice: 0.9176, Train Acc: 0.9968, Test Acc: 0.9953\n",
      "Epoch [33/200], Avg Loss: 0.01079, Avg IoU: 0.8903, Avg Dice: 0.9292,Test IoU: 0.8680, Test Dice: 0.9158, Train Acc: 0.9968, Test Acc: 0.9951\n",
      "Epoch [34/200], Avg Loss: 0.01079, Avg IoU: 0.8938, Avg Dice: 0.9313,Test IoU: 0.8614, Test Dice: 0.9116, Train Acc: 0.9970, Test Acc: 0.9947\n",
      "Epoch [35/200], Avg Loss: 0.01079, Avg IoU: 0.8960, Avg Dice: 0.9324,Test IoU: 0.8710, Test Dice: 0.9169, Train Acc: 0.9971, Test Acc: 0.9952\n",
      "Epoch [36/200], Avg Loss: 0.01079, Avg IoU: 0.8949, Avg Dice: 0.9320,Test IoU: 0.8560, Test Dice: 0.9087, Train Acc: 0.9970, Test Acc: 0.9945\n",
      "Epoch [37/200], Avg Loss: 0.01079, Avg IoU: 0.8963, Avg Dice: 0.9330,Test IoU: 0.8624, Test Dice: 0.9123, Train Acc: 0.9971, Test Acc: 0.9949\n",
      "Epoch [38/200], Avg Loss: 0.01079, Avg IoU: 0.8989, Avg Dice: 0.9346,Test IoU: 0.8554, Test Dice: 0.9084, Train Acc: 0.9971, Test Acc: 0.9945\n",
      "Epoch [39/200], Avg Loss: 0.01079, Avg IoU: 0.8996, Avg Dice: 0.9349,Test IoU: 0.8381, Test Dice: 0.8980, Train Acc: 0.9972, Test Acc: 0.9937\n",
      "Epoch [40/200], Avg Loss: 0.01079, Avg IoU: 0.9012, Avg Dice: 0.9363,Test IoU: 0.8586, Test Dice: 0.9100, Train Acc: 0.9972, Test Acc: 0.9946\n",
      "Epoch [41/200], Avg Loss: 0.01079, Avg IoU: 0.9044, Avg Dice: 0.9379,Test IoU: 0.6469, Test Dice: 0.7593, Train Acc: 0.9973, Test Acc: 0.9841\n",
      "Epoch [42/200], Avg Loss: 0.01081, Avg IoU: 0.8213, Avg Dice: 0.8815,Test IoU: 0.8363, Test Dice: 0.8936, Train Acc: 0.9940, Test Acc: 0.9937\n",
      "Epoch [43/200], Avg Loss: 0.01081, Avg IoU: 0.8435, Avg Dice: 0.8977,Test IoU: 0.8469, Test Dice: 0.9024, Train Acc: 0.9950, Test Acc: 0.9943\n",
      "Epoch [44/200], Avg Loss: 0.01080, Avg IoU: 0.8677, Avg Dice: 0.9138,Test IoU: 0.8546, Test Dice: 0.9081, Train Acc: 0.9960, Test Acc: 0.9945\n",
      "Epoch [45/200], Avg Loss: 0.01080, Avg IoU: 0.8797, Avg Dice: 0.9219,Test IoU: 0.8473, Test Dice: 0.9041, Train Acc: 0.9965, Test Acc: 0.9942\n",
      "Epoch [46/200], Avg Loss: 0.01080, Avg IoU: 0.8768, Avg Dice: 0.9201,Test IoU: 0.8671, Test Dice: 0.9148, Train Acc: 0.9964, Test Acc: 0.9949\n",
      "Epoch [47/200], Avg Loss: 0.01080, Avg IoU: 0.8867, Avg Dice: 0.9264,Test IoU: 0.8328, Test Dice: 0.8955, Train Acc: 0.9968, Test Acc: 0.9934\n",
      "Epoch [48/200], Avg Loss: 0.01080, Avg IoU: 0.8809, Avg Dice: 0.9230,Test IoU: 0.8565, Test Dice: 0.9092, Train Acc: 0.9965, Test Acc: 0.9946\n",
      "Epoch [49/200], Avg Loss: 0.01079, Avg IoU: 0.8900, Avg Dice: 0.9287,Test IoU: 0.8617, Test Dice: 0.9114, Train Acc: 0.9969, Test Acc: 0.9947\n",
      "Epoch [50/200], Avg Loss: 0.01079, Avg IoU: 0.8897, Avg Dice: 0.9289,Test IoU: 0.8651, Test Dice: 0.9136, Train Acc: 0.9968, Test Acc: 0.9949\n",
      "Epoch [51/200], Avg Loss: 0.01079, Avg IoU: 0.8960, Avg Dice: 0.9326,Test IoU: 0.8685, Test Dice: 0.9156, Train Acc: 0.9971, Test Acc: 0.9951\n",
      "Epoch [52/200], Avg Loss: 0.01079, Avg IoU: 0.8993, Avg Dice: 0.9349,Test IoU: 0.8553, Test Dice: 0.9081, Train Acc: 0.9972, Test Acc: 0.9944\n",
      "Epoch [53/200], Avg Loss: 0.01079, Avg IoU: 0.8977, Avg Dice: 0.9342,Test IoU: 0.8799, Test Dice: 0.9223, Train Acc: 0.9971, Test Acc: 0.9957\n",
      "Epoch [54/200], Avg Loss: 0.01079, Avg IoU: 0.9013, Avg Dice: 0.9360,Test IoU: 0.8809, Test Dice: 0.9226, Train Acc: 0.9972, Test Acc: 0.9957\n",
      "Epoch [55/200], Avg Loss: 0.01079, Avg IoU: 0.9018, Avg Dice: 0.9366,Test IoU: 0.8673, Test Dice: 0.9147, Train Acc: 0.9972, Test Acc: 0.9950\n",
      "Epoch [56/200], Avg Loss: 0.01079, Avg IoU: 0.9056, Avg Dice: 0.9388,Test IoU: 0.8715, Test Dice: 0.9175, Train Acc: 0.9974, Test Acc: 0.9953\n",
      "Epoch [57/200], Avg Loss: 0.01079, Avg IoU: 0.9084, Avg Dice: 0.9409,Test IoU: 0.8675, Test Dice: 0.9149, Train Acc: 0.9975, Test Acc: 0.9950\n",
      "Epoch [58/200], Avg Loss: 0.01079, Avg IoU: 0.9077, Avg Dice: 0.9402,Test IoU: 0.8706, Test Dice: 0.9163, Train Acc: 0.9975, Test Acc: 0.9952\n",
      "Epoch [59/200], Avg Loss: 0.01079, Avg IoU: 0.9063, Avg Dice: 0.9393,Test IoU: 0.8506, Test Dice: 0.9055, Train Acc: 0.9974, Test Acc: 0.9943\n",
      "Epoch [60/200], Avg Loss: 0.01079, Avg IoU: 0.9045, Avg Dice: 0.9385,Test IoU: 0.8718, Test Dice: 0.9171, Train Acc: 0.9973, Test Acc: 0.9953\n",
      "Epoch [61/200], Avg Loss: 0.01079, Avg IoU: 0.9116, Avg Dice: 0.9427,Test IoU: 0.8698, Test Dice: 0.9162, Train Acc: 0.9976, Test Acc: 0.9950\n",
      "Epoch [62/200], Avg Loss: 0.01079, Avg IoU: 0.9124, Avg Dice: 0.9430,Test IoU: 0.8512, Test Dice: 0.9059, Train Acc: 0.9976, Test Acc: 0.9942\n",
      "Epoch [63/200], Avg Loss: 0.01079, Avg IoU: 0.9114, Avg Dice: 0.9428,Test IoU: 0.8717, Test Dice: 0.9176, Train Acc: 0.9976, Test Acc: 0.9953\n",
      "Epoch [64/200], Avg Loss: 0.01079, Avg IoU: 0.9157, Avg Dice: 0.9453,Test IoU: 0.8680, Test Dice: 0.9155, Train Acc: 0.9977, Test Acc: 0.9950\n",
      "Epoch [65/200], Avg Loss: 0.01079, Avg IoU: 0.9139, Avg Dice: 0.9440,Test IoU: 0.8691, Test Dice: 0.9161, Train Acc: 0.9977, Test Acc: 0.9951\n",
      "Epoch [66/200], Avg Loss: 0.01079, Avg IoU: 0.9162, Avg Dice: 0.9454,Test IoU: 0.8728, Test Dice: 0.9179, Train Acc: 0.9977, Test Acc: 0.9953\n",
      "Epoch [67/200], Avg Loss: 0.01078, Avg IoU: 0.9174, Avg Dice: 0.9462,Test IoU: 0.8702, Test Dice: 0.9164, Train Acc: 0.9978, Test Acc: 0.9951\n",
      "Epoch [68/200], Avg Loss: 0.01079, Avg IoU: 0.9146, Avg Dice: 0.9442,Test IoU: 0.8746, Test Dice: 0.9188, Train Acc: 0.9977, Test Acc: 0.9954\n",
      "Epoch [69/200], Avg Loss: 0.01078, Avg IoU: 0.9175, Avg Dice: 0.9461,Test IoU: 0.8743, Test Dice: 0.9188, Train Acc: 0.9978, Test Acc: 0.9954\n",
      "Epoch [70/200], Avg Loss: 0.01079, Avg IoU: 0.9179, Avg Dice: 0.9463,Test IoU: 0.8753, Test Dice: 0.9193, Train Acc: 0.9978, Test Acc: 0.9955\n",
      "Epoch [71/200], Avg Loss: 0.01079, Avg IoU: 0.9185, Avg Dice: 0.9465,Test IoU: 0.8562, Test Dice: 0.9087, Train Acc: 0.9978, Test Acc: 0.9945\n",
      "Epoch [72/200], Avg Loss: 0.01079, Avg IoU: 0.9184, Avg Dice: 0.9465,Test IoU: 0.8516, Test Dice: 0.9060, Train Acc: 0.9978, Test Acc: 0.9943\n",
      "Epoch [73/200], Avg Loss: 0.01079, Avg IoU: 0.9169, Avg Dice: 0.9455,Test IoU: 0.8300, Test Dice: 0.8935, Train Acc: 0.9978, Test Acc: 0.9932\n",
      "Epoch [74/200], Avg Loss: 0.01079, Avg IoU: 0.9014, Avg Dice: 0.9364,Test IoU: 0.8774, Test Dice: 0.9204, Train Acc: 0.9972, Test Acc: 0.9955\n",
      "Epoch [75/200], Avg Loss: 0.01078, Avg IoU: 0.9173, Avg Dice: 0.9457,Test IoU: 0.8716, Test Dice: 0.9169, Train Acc: 0.9978, Test Acc: 0.9952\n",
      "Epoch [76/200], Avg Loss: 0.01079, Avg IoU: 0.9195, Avg Dice: 0.9472,Test IoU: 0.8509, Test Dice: 0.9044, Train Acc: 0.9979, Test Acc: 0.9943\n",
      "Epoch [77/200], Avg Loss: 0.01080, Avg IoU: 0.8613, Avg Dice: 0.9092,Test IoU: 0.8455, Test Dice: 0.9009, Train Acc: 0.9957, Test Acc: 0.9941\n",
      "Epoch [78/200], Avg Loss: 0.01079, Avg IoU: 0.8872, Avg Dice: 0.9268,Test IoU: 0.8638, Test Dice: 0.9129, Train Acc: 0.9968, Test Acc: 0.9948\n",
      "Epoch [79/200], Avg Loss: 0.01079, Avg IoU: 0.9015, Avg Dice: 0.9359,Test IoU: 0.8371, Test Dice: 0.8971, Train Acc: 0.9973, Test Acc: 0.9936\n",
      "Epoch [80/200], Avg Loss: 0.01079, Avg IoU: 0.9040, Avg Dice: 0.9377,Test IoU: 0.8718, Test Dice: 0.9170, Train Acc: 0.9973, Test Acc: 0.9952\n",
      "Epoch [81/200], Avg Loss: 0.01079, Avg IoU: 0.9097, Avg Dice: 0.9414,Test IoU: 0.8413, Test Dice: 0.8997, Train Acc: 0.9975, Test Acc: 0.9938\n",
      "Epoch [82/200], Avg Loss: 0.01079, Avg IoU: 0.8866, Avg Dice: 0.9273,Test IoU: 0.8567, Test Dice: 0.9085, Train Acc: 0.9967, Test Acc: 0.9947\n",
      "Epoch [83/200], Avg Loss: 0.01079, Avg IoU: 0.9073, Avg Dice: 0.9396,Test IoU: 0.8711, Test Dice: 0.9172, Train Acc: 0.9975, Test Acc: 0.9953\n",
      "Epoch [84/200], Avg Loss: 0.01079, Avg IoU: 0.9119, Avg Dice: 0.9424,Test IoU: 0.8527, Test Dice: 0.9061, Train Acc: 0.9976, Test Acc: 0.9944\n",
      "Epoch [85/200], Avg Loss: 0.01078, Avg IoU: 0.9144, Avg Dice: 0.9438,Test IoU: 0.8713, Test Dice: 0.9167, Train Acc: 0.9977, Test Acc: 0.9952\n",
      "Epoch [86/200], Avg Loss: 0.01079, Avg IoU: 0.9187, Avg Dice: 0.9465,Test IoU: 0.8558, Test Dice: 0.9078, Train Acc: 0.9979, Test Acc: 0.9945\n",
      "Epoch [87/200], Avg Loss: 0.01079, Avg IoU: 0.9167, Avg Dice: 0.9453,Test IoU: 0.8593, Test Dice: 0.9098, Train Acc: 0.9978, Test Acc: 0.9946\n",
      "Epoch [88/200], Avg Loss: 0.01078, Avg IoU: 0.9198, Avg Dice: 0.9472,Test IoU: 0.8577, Test Dice: 0.9090, Train Acc: 0.9979, Test Acc: 0.9946\n",
      "Epoch [89/200], Avg Loss: 0.01079, Avg IoU: 0.9074, Avg Dice: 0.9397,Test IoU: 0.8587, Test Dice: 0.9096, Train Acc: 0.9975, Test Acc: 0.9947\n",
      "Epoch [90/200], Avg Loss: 0.01079, Avg IoU: 0.9149, Avg Dice: 0.9445,Test IoU: 0.8683, Test Dice: 0.9151, Train Acc: 0.9977, Test Acc: 0.9951\n",
      "Epoch [91/200], Avg Loss: 0.01079, Avg IoU: 0.9165, Avg Dice: 0.9454,Test IoU: 0.8475, Test Dice: 0.9032, Train Acc: 0.9978, Test Acc: 0.9943\n",
      "Epoch [92/200], Avg Loss: 0.01079, Avg IoU: 0.9152, Avg Dice: 0.9444,Test IoU: 0.8497, Test Dice: 0.9044, Train Acc: 0.9977, Test Acc: 0.9943\n",
      "Epoch [93/200], Avg Loss: 0.01078, Avg IoU: 0.9183, Avg Dice: 0.9467,Test IoU: 0.8700, Test Dice: 0.9164, Train Acc: 0.9978, Test Acc: 0.9952\n",
      "Epoch [94/200], Avg Loss: 0.01079, Avg IoU: 0.8882, Avg Dice: 0.9275,Test IoU: 0.8610, Test Dice: 0.9111, Train Acc: 0.9968, Test Acc: 0.9948\n",
      "Epoch [95/200], Avg Loss: 0.01079, Avg IoU: 0.9036, Avg Dice: 0.9373,Test IoU: 0.8723, Test Dice: 0.9179, Train Acc: 0.9973, Test Acc: 0.9954\n",
      "Epoch [96/200], Avg Loss: 0.01079, Avg IoU: 0.9067, Avg Dice: 0.9394,Test IoU: 0.8509, Test Dice: 0.9049, Train Acc: 0.9974, Test Acc: 0.9945\n",
      "Epoch [97/200], Avg Loss: 0.01079, Avg IoU: 0.8878, Avg Dice: 0.9278,Test IoU: 0.8624, Test Dice: 0.9121, Train Acc: 0.9967, Test Acc: 0.9948\n",
      "Epoch [98/200], Avg Loss: 0.01079, Avg IoU: 0.9068, Avg Dice: 0.9395,Test IoU: 0.8621, Test Dice: 0.9120, Train Acc: 0.9974, Test Acc: 0.9948\n",
      "Epoch [99/200], Avg Loss: 0.01079, Avg IoU: 0.9144, Avg Dice: 0.9439,Test IoU: 0.8676, Test Dice: 0.9150, Train Acc: 0.9977, Test Acc: 0.9951\n",
      "Epoch [100/200], Avg Loss: 0.01079, Avg IoU: 0.9171, Avg Dice: 0.9457,Test IoU: 0.8658, Test Dice: 0.9137, Train Acc: 0.9978, Test Acc: 0.9950\n",
      "Epoch [101/200], Avg Loss: 0.01078, Avg IoU: 0.9201, Avg Dice: 0.9473,Test IoU: 0.8678, Test Dice: 0.9152, Train Acc: 0.9979, Test Acc: 0.9951\n",
      "Epoch [102/200], Avg Loss: 0.01079, Avg IoU: 0.9157, Avg Dice: 0.9451,Test IoU: 0.8472, Test Dice: 0.9032, Train Acc: 0.9977, Test Acc: 0.9941\n",
      "Epoch [103/200], Avg Loss: 0.01079, Avg IoU: 0.9104, Avg Dice: 0.9417,Test IoU: 0.8546, Test Dice: 0.9078, Train Acc: 0.9975, Test Acc: 0.9944\n",
      "Epoch [104/200], Avg Loss: 0.01078, Avg IoU: 0.9163, Avg Dice: 0.9454,Test IoU: 0.8646, Test Dice: 0.9132, Train Acc: 0.9978, Test Acc: 0.9949\n",
      "Epoch [105/200], Avg Loss: 0.01078, Avg IoU: 0.9158, Avg Dice: 0.9451,Test IoU: 0.8700, Test Dice: 0.9165, Train Acc: 0.9977, Test Acc: 0.9953\n",
      "Epoch [106/200], Avg Loss: 0.01078, Avg IoU: 0.9179, Avg Dice: 0.9462,Test IoU: 0.8676, Test Dice: 0.9149, Train Acc: 0.9978, Test Acc: 0.9951\n",
      "Epoch [107/200], Avg Loss: 0.01078, Avg IoU: 0.9197, Avg Dice: 0.9471,Test IoU: 0.8674, Test Dice: 0.9146, Train Acc: 0.9979, Test Acc: 0.9951\n",
      "Epoch [108/200], Avg Loss: 0.01078, Avg IoU: 0.9217, Avg Dice: 0.9485,Test IoU: 0.8739, Test Dice: 0.9184, Train Acc: 0.9979, Test Acc: 0.9955\n",
      "Epoch [109/200], Avg Loss: 0.01078, Avg IoU: 0.9234, Avg Dice: 0.9493,Test IoU: 0.8454, Test Dice: 0.9021, Train Acc: 0.9980, Test Acc: 0.9940\n",
      "Epoch [110/200], Avg Loss: 0.01078, Avg IoU: 0.9182, Avg Dice: 0.9463,Test IoU: 0.8727, Test Dice: 0.9178, Train Acc: 0.9978, Test Acc: 0.9953\n",
      "Epoch [111/200], Avg Loss: 0.01078, Avg IoU: 0.9216, Avg Dice: 0.9484,Test IoU: 0.8654, Test Dice: 0.9136, Train Acc: 0.9979, Test Acc: 0.9950\n",
      "Epoch [112/200], Avg Loss: 0.01079, Avg IoU: 0.9160, Avg Dice: 0.9451,Test IoU: 0.8476, Test Dice: 0.9034, Train Acc: 0.9977, Test Acc: 0.9942\n",
      "Epoch [113/200], Avg Loss: 0.01078, Avg IoU: 0.9135, Avg Dice: 0.9434,Test IoU: 0.8724, Test Dice: 0.9175, Train Acc: 0.9977, Test Acc: 0.9953\n",
      "Epoch [114/200], Avg Loss: 0.01079, Avg IoU: 0.9163, Avg Dice: 0.9449,Test IoU: 0.8427, Test Dice: 0.9010, Train Acc: 0.9978, Test Acc: 0.9939\n",
      "Epoch [115/200], Avg Loss: 0.01079, Avg IoU: 0.9157, Avg Dice: 0.9448,Test IoU: 0.8490, Test Dice: 0.9043, Train Acc: 0.9977, Test Acc: 0.9943\n",
      "Epoch [116/200], Avg Loss: 0.01078, Avg IoU: 0.9214, Avg Dice: 0.9480,Test IoU: 0.8619, Test Dice: 0.9120, Train Acc: 0.9980, Test Acc: 0.9948\n",
      "Epoch [117/200], Avg Loss: 0.01078, Avg IoU: 0.9237, Avg Dice: 0.9495,Test IoU: 0.8663, Test Dice: 0.9143, Train Acc: 0.9980, Test Acc: 0.9950\n",
      "Epoch [118/200], Avg Loss: 0.01078, Avg IoU: 0.9252, Avg Dice: 0.9503,Test IoU: 0.8571, Test Dice: 0.9088, Train Acc: 0.9981, Test Acc: 0.9946\n",
      "Epoch [119/200], Avg Loss: 0.01079, Avg IoU: 0.9210, Avg Dice: 0.9481,Test IoU: 0.8602, Test Dice: 0.9106, Train Acc: 0.9979, Test Acc: 0.9948\n",
      "Epoch [120/200], Avg Loss: 0.01079, Avg IoU: 0.9169, Avg Dice: 0.9437,Test IoU: 0.8603, Test Dice: 0.9111, Train Acc: 0.9979, Test Acc: 0.9948\n",
      "Epoch [121/200], Avg Loss: 0.01078, Avg IoU: 0.9224, Avg Dice: 0.9482,Test IoU: 0.8692, Test Dice: 0.9160, Train Acc: 0.9980, Test Acc: 0.9952\n",
      "Epoch [122/200], Avg Loss: 0.01078, Avg IoU: 0.9221, Avg Dice: 0.9484,Test IoU: 0.8545, Test Dice: 0.9075, Train Acc: 0.9980, Test Acc: 0.9945\n",
      "Epoch [123/200], Avg Loss: 0.01078, Avg IoU: 0.9269, Avg Dice: 0.9511,Test IoU: 0.8642, Test Dice: 0.9131, Train Acc: 0.9982, Test Acc: 0.9950\n",
      "Epoch [124/200], Avg Loss: 0.01078, Avg IoU: 0.9268, Avg Dice: 0.9514,Test IoU: 0.8782, Test Dice: 0.9210, Train Acc: 0.9981, Test Acc: 0.9956\n",
      "Epoch [125/200], Avg Loss: 0.01078, Avg IoU: 0.9267, Avg Dice: 0.9514,Test IoU: 0.8634, Test Dice: 0.9126, Train Acc: 0.9981, Test Acc: 0.9949\n",
      "Epoch [126/200], Avg Loss: 0.01079, Avg IoU: 0.9122, Avg Dice: 0.9424,Test IoU: 0.8527, Test Dice: 0.9066, Train Acc: 0.9976, Test Acc: 0.9944\n",
      "Epoch [127/200], Avg Loss: 0.01078, Avg IoU: 0.9216, Avg Dice: 0.9483,Test IoU: 0.8553, Test Dice: 0.9083, Train Acc: 0.9980, Test Acc: 0.9945\n",
      "Epoch [128/200], Avg Loss: 0.01078, Avg IoU: 0.9201, Avg Dice: 0.9473,Test IoU: 0.8625, Test Dice: 0.9118, Train Acc: 0.9979, Test Acc: 0.9948\n",
      "Epoch [129/200], Avg Loss: 0.01078, Avg IoU: 0.9252, Avg Dice: 0.9502,Test IoU: 0.8737, Test Dice: 0.9183, Train Acc: 0.9981, Test Acc: 0.9954\n",
      "Epoch [130/200], Avg Loss: 0.01078, Avg IoU: 0.9291, Avg Dice: 0.9524,Test IoU: 0.8701, Test Dice: 0.9162, Train Acc: 0.9982, Test Acc: 0.9952\n",
      "Epoch [131/200], Avg Loss: 0.01078, Avg IoU: 0.9310, Avg Dice: 0.9533,Test IoU: 0.8683, Test Dice: 0.9154, Train Acc: 0.9983, Test Acc: 0.9951\n",
      "Epoch [132/200], Avg Loss: 0.01078, Avg IoU: 0.9326, Avg Dice: 0.9545,Test IoU: 0.8707, Test Dice: 0.9167, Train Acc: 0.9983, Test Acc: 0.9952\n",
      "Epoch [133/200], Avg Loss: 0.01078, Avg IoU: 0.9314, Avg Dice: 0.9538,Test IoU: 0.8745, Test Dice: 0.9186, Train Acc: 0.9983, Test Acc: 0.9954\n",
      "Epoch [134/200], Avg Loss: 0.01078, Avg IoU: 0.9332, Avg Dice: 0.9550,Test IoU: 0.8654, Test Dice: 0.9136, Train Acc: 0.9984, Test Acc: 0.9950\n",
      "Epoch [135/200], Avg Loss: 0.01078, Avg IoU: 0.9321, Avg Dice: 0.9542,Test IoU: 0.8664, Test Dice: 0.9142, Train Acc: 0.9983, Test Acc: 0.9950\n",
      "Epoch [136/200], Avg Loss: 0.01078, Avg IoU: 0.9317, Avg Dice: 0.9538,Test IoU: 0.8640, Test Dice: 0.9128, Train Acc: 0.9983, Test Acc: 0.9949\n",
      "Epoch [137/200], Avg Loss: 0.01078, Avg IoU: 0.9271, Avg Dice: 0.9514,Test IoU: 0.8637, Test Dice: 0.9126, Train Acc: 0.9982, Test Acc: 0.9949\n",
      "Epoch [138/200], Avg Loss: 0.01078, Avg IoU: 0.9326, Avg Dice: 0.9544,Test IoU: 0.8703, Test Dice: 0.9165, Train Acc: 0.9983, Test Acc: 0.9952\n",
      "Epoch [139/200], Avg Loss: 0.01078, Avg IoU: 0.9306, Avg Dice: 0.9532,Test IoU: 0.8767, Test Dice: 0.9201, Train Acc: 0.9983, Test Acc: 0.9955\n",
      "Epoch [140/200], Avg Loss: 0.01078, Avg IoU: 0.9323, Avg Dice: 0.9542,Test IoU: 0.8756, Test Dice: 0.9195, Train Acc: 0.9983, Test Acc: 0.9954\n",
      "Epoch [141/200], Avg Loss: 0.01078, Avg IoU: 0.9347, Avg Dice: 0.9557,Test IoU: 0.8637, Test Dice: 0.9128, Train Acc: 0.9984, Test Acc: 0.9949\n",
      "Epoch [142/200], Avg Loss: 0.01078, Avg IoU: 0.9323, Avg Dice: 0.9544,Test IoU: 0.8542, Test Dice: 0.9073, Train Acc: 0.9983, Test Acc: 0.9945\n",
      "Epoch [143/200], Avg Loss: 0.01078, Avg IoU: 0.9296, Avg Dice: 0.9528,Test IoU: 0.8773, Test Dice: 0.9203, Train Acc: 0.9982, Test Acc: 0.9955\n",
      "Epoch [144/200], Avg Loss: 0.01078, Avg IoU: 0.9311, Avg Dice: 0.9535,Test IoU: 0.8595, Test Dice: 0.9104, Train Acc: 0.9983, Test Acc: 0.9947\n",
      "Epoch [145/200], Avg Loss: 0.01078, Avg IoU: 0.9291, Avg Dice: 0.9523,Test IoU: 0.8775, Test Dice: 0.9203, Train Acc: 0.9982, Test Acc: 0.9955\n",
      "Epoch [146/200], Avg Loss: 0.01078, Avg IoU: 0.9344, Avg Dice: 0.9556,Test IoU: 0.8643, Test Dice: 0.9132, Train Acc: 0.9984, Test Acc: 0.9949\n",
      "Epoch [147/200], Avg Loss: 0.01079, Avg IoU: 0.9084, Avg Dice: 0.9402,Test IoU: 0.8554, Test Dice: 0.9077, Train Acc: 0.9975, Test Acc: 0.9944\n",
      "Epoch [148/200], Avg Loss: 0.01078, Avg IoU: 0.9188, Avg Dice: 0.9463,Test IoU: 0.8553, Test Dice: 0.9078, Train Acc: 0.9979, Test Acc: 0.9944\n",
      "Epoch [149/200], Avg Loss: 0.01078, Avg IoU: 0.9255, Avg Dice: 0.9505,Test IoU: 0.8691, Test Dice: 0.9157, Train Acc: 0.9981, Test Acc: 0.9952\n",
      "Epoch [150/200], Avg Loss: 0.01078, Avg IoU: 0.9307, Avg Dice: 0.9534,Test IoU: 0.8695, Test Dice: 0.9159, Train Acc: 0.9983, Test Acc: 0.9952\n",
      "Epoch [151/200], Avg Loss: 0.01078, Avg IoU: 0.9304, Avg Dice: 0.9528,Test IoU: 0.8758, Test Dice: 0.9194, Train Acc: 0.9983, Test Acc: 0.9955\n",
      "Epoch [152/200], Avg Loss: 0.01078, Avg IoU: 0.9339, Avg Dice: 0.9550,Test IoU: 0.8662, Test Dice: 0.9139, Train Acc: 0.9984, Test Acc: 0.9951\n",
      "Epoch [153/200], Avg Loss: 0.01078, Avg IoU: 0.9300, Avg Dice: 0.9527,Test IoU: 0.8629, Test Dice: 0.9123, Train Acc: 0.9983, Test Acc: 0.9949\n",
      "Epoch [154/200], Avg Loss: 0.01078, Avg IoU: 0.9199, Avg Dice: 0.9471,Test IoU: 0.8640, Test Dice: 0.9128, Train Acc: 0.9979, Test Acc: 0.9950\n",
      "Epoch [155/200], Avg Loss: 0.01078, Avg IoU: 0.9203, Avg Dice: 0.9474,Test IoU: 0.8783, Test Dice: 0.9209, Train Acc: 0.9979, Test Acc: 0.9956\n",
      "Epoch [156/200], Avg Loss: 0.01078, Avg IoU: 0.9287, Avg Dice: 0.9518,Test IoU: 0.8700, Test Dice: 0.9162, Train Acc: 0.9982, Test Acc: 0.9951\n",
      "Epoch [157/200], Avg Loss: 0.01079, Avg IoU: 0.9312, Avg Dice: 0.9535,Test IoU: 0.8607, Test Dice: 0.9110, Train Acc: 0.9983, Test Acc: 0.9948\n",
      "Epoch [158/200], Avg Loss: 0.01078, Avg IoU: 0.9241, Avg Dice: 0.9488,Test IoU: 0.8566, Test Dice: 0.9083, Train Acc: 0.9981, Test Acc: 0.9946\n",
      "Epoch [159/200], Avg Loss: 0.01079, Avg IoU: 0.9148, Avg Dice: 0.9441,Test IoU: 0.8488, Test Dice: 0.9036, Train Acc: 0.9977, Test Acc: 0.9941\n",
      "Epoch [160/200], Avg Loss: 0.01078, Avg IoU: 0.9247, Avg Dice: 0.9498,Test IoU: 0.8621, Test Dice: 0.9116, Train Acc: 0.9981, Test Acc: 0.9948\n",
      "Epoch [161/200], Avg Loss: 0.01079, Avg IoU: 0.8975, Avg Dice: 0.9333,Test IoU: 0.8742, Test Dice: 0.9183, Train Acc: 0.9971, Test Acc: 0.9953\n",
      "Epoch [162/200], Avg Loss: 0.01079, Avg IoU: 0.9129, Avg Dice: 0.9429,Test IoU: 0.8592, Test Dice: 0.9096, Train Acc: 0.9977, Test Acc: 0.9946\n",
      "Epoch [163/200], Avg Loss: 0.01079, Avg IoU: 0.9049, Avg Dice: 0.9377,Test IoU: 0.8329, Test Dice: 0.8904, Train Acc: 0.9973, Test Acc: 0.9935\n",
      "Epoch [164/200], Avg Loss: 0.01080, Avg IoU: 0.8757, Avg Dice: 0.9186,Test IoU: 0.8522, Test Dice: 0.9065, Train Acc: 0.9963, Test Acc: 0.9945\n",
      "Epoch [165/200], Avg Loss: 0.01079, Avg IoU: 0.8936, Avg Dice: 0.9308,Test IoU: 0.8485, Test Dice: 0.9040, Train Acc: 0.9970, Test Acc: 0.9942\n",
      "Epoch [166/200], Avg Loss: 0.01079, Avg IoU: 0.8858, Avg Dice: 0.9267,Test IoU: 0.8508, Test Dice: 0.9058, Train Acc: 0.9966, Test Acc: 0.9944\n",
      "Epoch [167/200], Avg Loss: 0.01078, Avg IoU: 0.9048, Avg Dice: 0.9385,Test IoU: 0.8679, Test Dice: 0.9155, Train Acc: 0.9973, Test Acc: 0.9951\n",
      "Epoch [168/200], Avg Loss: 0.01079, Avg IoU: 0.9123, Avg Dice: 0.9427,Test IoU: 0.8679, Test Dice: 0.9154, Train Acc: 0.9976, Test Acc: 0.9951\n",
      "Epoch [169/200], Avg Loss: 0.01078, Avg IoU: 0.9180, Avg Dice: 0.9461,Test IoU: 0.8636, Test Dice: 0.9131, Train Acc: 0.9978, Test Acc: 0.9949\n",
      "Epoch [170/200], Avg Loss: 0.01078, Avg IoU: 0.9211, Avg Dice: 0.9477,Test IoU: 0.8672, Test Dice: 0.9146, Train Acc: 0.9979, Test Acc: 0.9950\n",
      "Epoch [171/200], Avg Loss: 0.01079, Avg IoU: 0.9231, Avg Dice: 0.9489,Test IoU: 0.8604, Test Dice: 0.9108, Train Acc: 0.9980, Test Acc: 0.9948\n",
      "Epoch [172/200], Avg Loss: 0.01078, Avg IoU: 0.9145, Avg Dice: 0.9438,Test IoU: 0.8615, Test Dice: 0.9116, Train Acc: 0.9977, Test Acc: 0.9948\n",
      "Epoch [173/200], Avg Loss: 0.01078, Avg IoU: 0.9242, Avg Dice: 0.9495,Test IoU: 0.8710, Test Dice: 0.9166, Train Acc: 0.9980, Test Acc: 0.9953\n",
      "Epoch [174/200], Avg Loss: 0.01078, Avg IoU: 0.9254, Avg Dice: 0.9502,Test IoU: 0.8615, Test Dice: 0.9114, Train Acc: 0.9981, Test Acc: 0.9948\n",
      "Epoch [175/200], Avg Loss: 0.01078, Avg IoU: 0.9260, Avg Dice: 0.9504,Test IoU: 0.8636, Test Dice: 0.9128, Train Acc: 0.9981, Test Acc: 0.9949\n",
      "Epoch [176/200], Avg Loss: 0.01078, Avg IoU: 0.9292, Avg Dice: 0.9522,Test IoU: 0.8715, Test Dice: 0.9172, Train Acc: 0.9982, Test Acc: 0.9953\n",
      "Epoch [177/200], Avg Loss: 0.01078, Avg IoU: 0.9298, Avg Dice: 0.9525,Test IoU: 0.8652, Test Dice: 0.9135, Train Acc: 0.9983, Test Acc: 0.9950\n",
      "Epoch [178/200], Avg Loss: 0.01078, Avg IoU: 0.9244, Avg Dice: 0.9494,Test IoU: 0.8661, Test Dice: 0.9141, Train Acc: 0.9981, Test Acc: 0.9951\n",
      "Epoch [179/200], Avg Loss: 0.01078, Avg IoU: 0.9293, Avg Dice: 0.9522,Test IoU: 0.8654, Test Dice: 0.9136, Train Acc: 0.9982, Test Acc: 0.9950\n",
      "Epoch [180/200], Avg Loss: 0.01078, Avg IoU: 0.9321, Avg Dice: 0.9539,Test IoU: 0.8789, Test Dice: 0.9208, Train Acc: 0.9983, Test Acc: 0.9956\n",
      "Epoch [181/200], Avg Loss: 0.01078, Avg IoU: 0.9325, Avg Dice: 0.9539,Test IoU: 0.8659, Test Dice: 0.9139, Train Acc: 0.9984, Test Acc: 0.9950\n",
      "Epoch [182/200], Avg Loss: 0.01079, Avg IoU: 0.9185, Avg Dice: 0.9461,Test IoU: 0.8710, Test Dice: 0.9168, Train Acc: 0.9979, Test Acc: 0.9953\n",
      "Epoch [183/200], Avg Loss: 0.01078, Avg IoU: 0.9295, Avg Dice: 0.9523,Test IoU: 0.8703, Test Dice: 0.9165, Train Acc: 0.9983, Test Acc: 0.9953\n",
      "Epoch [184/200], Avg Loss: 0.01078, Avg IoU: 0.9278, Avg Dice: 0.9516,Test IoU: 0.8761, Test Dice: 0.9193, Train Acc: 0.9982, Test Acc: 0.9955\n",
      "Epoch [185/200], Avg Loss: 0.01078, Avg IoU: 0.9325, Avg Dice: 0.9541,Test IoU: 0.8675, Test Dice: 0.9148, Train Acc: 0.9984, Test Acc: 0.9951\n",
      "Epoch [186/200], Avg Loss: 0.01078, Avg IoU: 0.9296, Avg Dice: 0.9525,Test IoU: 0.8607, Test Dice: 0.9108, Train Acc: 0.9983, Test Acc: 0.9948\n",
      "Epoch [187/200], Avg Loss: 0.01078, Avg IoU: 0.9280, Avg Dice: 0.9515,Test IoU: 0.8730, Test Dice: 0.9174, Train Acc: 0.9982, Test Acc: 0.9954\n",
      "Epoch [188/200], Avg Loss: 0.01078, Avg IoU: 0.9342, Avg Dice: 0.9548,Test IoU: 0.8728, Test Dice: 0.9174, Train Acc: 0.9984, Test Acc: 0.9953\n",
      "Epoch [189/200], Avg Loss: 0.01078, Avg IoU: 0.9364, Avg Dice: 0.9561,Test IoU: 0.8739, Test Dice: 0.9181, Train Acc: 0.9985, Test Acc: 0.9954\n",
      "Epoch [190/200], Avg Loss: 0.01078, Avg IoU: 0.9369, Avg Dice: 0.9564,Test IoU: 0.8539, Test Dice: 0.9071, Train Acc: 0.9985, Test Acc: 0.9944\n",
      "Epoch [191/200], Avg Loss: 0.01080, Avg IoU: 0.8666, Avg Dice: 0.9138,Test IoU: 0.8532, Test Dice: 0.9072, Train Acc: 0.9959, Test Acc: 0.9943\n",
      "Epoch [192/200], Avg Loss: 0.01079, Avg IoU: 0.8979, Avg Dice: 0.9335,Test IoU: 0.8402, Test Dice: 0.8997, Train Acc: 0.9972, Test Acc: 0.9937\n",
      "Epoch [193/200], Avg Loss: 0.01079, Avg IoU: 0.9053, Avg Dice: 0.9384,Test IoU: 0.8655, Test Dice: 0.9139, Train Acc: 0.9974, Test Acc: 0.9949\n",
      "Epoch [194/200], Avg Loss: 0.01079, Avg IoU: 0.9163, Avg Dice: 0.9445,Test IoU: 0.8431, Test Dice: 0.9007, Train Acc: 0.9978, Test Acc: 0.9940\n",
      "Epoch [195/200], Avg Loss: 0.01079, Avg IoU: 0.8975, Avg Dice: 0.9335,Test IoU: 0.8447, Test Dice: 0.9003, Train Acc: 0.9971, Test Acc: 0.9940\n",
      "Epoch [196/200], Avg Loss: 0.01079, Avg IoU: 0.9152, Avg Dice: 0.9441,Test IoU: 0.8603, Test Dice: 0.9102, Train Acc: 0.9978, Test Acc: 0.9947\n",
      "Epoch [197/200], Avg Loss: 0.01078, Avg IoU: 0.9208, Avg Dice: 0.9474,Test IoU: 0.8579, Test Dice: 0.9092, Train Acc: 0.9980, Test Acc: 0.9945\n",
      "Epoch [198/200], Avg Loss: 0.01078, Avg IoU: 0.9235, Avg Dice: 0.9488,Test IoU: 0.8500, Test Dice: 0.9040, Train Acc: 0.9981, Test Acc: 0.9941\n",
      "Epoch [199/200], Avg Loss: 0.01078, Avg IoU: 0.9268, Avg Dice: 0.9508,Test IoU: 0.8546, Test Dice: 0.9071, Train Acc: 0.9982, Test Acc: 0.9944\n",
      "Epoch [200/200], Avg Loss: 0.01078, Avg IoU: 0.9290, Avg Dice: 0.9520,Test IoU: 0.8476, Test Dice: 0.9030, Train Acc: 0.9982, Test Acc: 0.9941\n",
      "-----------------------------------------\n",
      "Best Loss: 0.0108, Best Train IoU: 0.9369, Best Train Dice: 0.9564, Best Test IoU: 0.8809, Best Test Dice: 0.9226, Best Train Acc: 0.9985, Best Test Acc: 0.9957\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# -----------------------------\n",
    "# UNet++ Train and Test Set\n",
    "# -----------------------------\n",
    "\n",
    "# empty the cache for model and training process\n",
    "torch.cuda.empty_cache()\n",
    "\n",
    "# set the hyper parameters and the paths 设置超参数和路径\n",
    "data_path = \"./dataset/Spineweb_dataset15/train_9/\"\n",
    "test_path = \"./dataset/Spineweb_dataset15/test_1/\"\n",
    "\n",
    "#hyper parameters set\n",
    "batch_size = 64\n",
    "num_epochs = 200\n",
    "learning_rate = 1e-03\n",
    "confidence = 0.7\n",
    "poster_threshold = 200\n",
    "in_channels = 1  # according to demand to adjust the number of channels 根据实际情况修改通道数\n",
    "out_channels = 1  # according to demand to adjust the number of channels 根据实际情况修改通道数\n",
    "\n",
    "# 创建数据集和数据加载器\n",
    "transform = transforms.Compose([\n",
    "    # transforms.ToPILImage(),  # 转换为 PIL 图像对象\n",
    "    transforms.Resize((128, 128)),  # 调整大小为 128x128\n",
    "    transforms.ToTensor()  # 转换为张量\n",
    "])\n",
    "\n",
    "dataset = SpineWeb15(data_path=data_path, transform1=transform, transform2=transform)\n",
    "dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)\n",
    "testset = SpineWeb15(data_path=test_path, transform1=transform, transform2=transform)\n",
    "test_loader = DataLoader(testset, batch_size=batch_size, shuffle=True)\n",
    "\n",
    "# 创建模型和优化器\n",
    "model = smp.MAnet(encoder_name='resnet34', encoder_depth=5, encoder_weights='imagenet', decoder_use_batchnorm=True, decoder_channels=(256, 128, 64, 32, 16), decoder_pab_channels=64, in_channels=1, classes=1, activation=None, aux_params=None)\n",
    "optimizer = optim.Adam(model.parameters(), lr=learning_rate)\n",
    "criterion = nn.BCEWithLogitsLoss()  # 二分类任务可以使用BCEWithLogitsLoss\n",
    "# criterion = DiceLoss(mode=\"binary\")   #DiceLoss\n",
    "# criterion = FocalLoss(mode=\"binary\")\n",
    "# criterion = smp.losses.TverskyLoss(mode=\"binary\")\n",
    "# criterion = smp.losses.LovaszLoss(mode=\"binary\")\n",
    "\n",
    "# 设置设备\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = model.to(device)\n",
    "\n",
    "# check the parameters of the model could be trained 检查模型的参数是否设置为可训练\n",
    "# for name, param in model.named_parameters():\n",
    "#     print(f\"Parameter: {name}, Requires Grad: {param.requires_grad}\")\n",
    "\n",
    "best_loss_train = 100.0\n",
    "best_IoU_train = 0.0\n",
    "best_Dice_train = 0.0\n",
    "\n",
    "best_IoU_test = 0.0\n",
    "best_Dice_test = 0.0\n",
    "\n",
    "best_acc_train = 0.0\n",
    "best_acc_test = 0.0\n",
    "\n",
    "loss_train_list = []\n",
    "IoU_train_list = []\n",
    "Dice_train_list = []\n",
    "IoU_test_list = []\n",
    "Dice_test_list = []\n",
    "acc_train_list = []\n",
    "acc_test_list = []\n",
    "\n",
    "# 训练模型\n",
    "for epoch in range(num_epochs):\n",
    "    # torch.cuda.empty_cache()\n",
    "    model.train()\n",
    "    train_loss = 0.0\n",
    "    total_iou = 0.0\n",
    "    total_dice = 0.0\n",
    "    total_iou_test = 0.0\n",
    "    total_dice_test = 0.0\n",
    "    total_acc_train = 0.0\n",
    "    total_acc_test = 0.0\n",
    "    total_correct_pixels = 0\n",
    "    total_pixels = 0\n",
    "    total_correct_pixels_test = 0\n",
    "    total_pixels_test = 0\n",
    "\n",
    "    if epoch > 100:\n",
    "        confidence = 0.9\n",
    "\n",
    "    for images, labels in dataloader:\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "    \n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        outputs = model(images)\n",
    "        # squeeze the dimension that pytorch add for channels\n",
    "        outputs = outputs.squeeze(1)     # tensor size ([16,128,128])\n",
    "        labels = labels.squeeze(1)\n",
    "        labels = labels * 255\n",
    "        labels[labels <= poster_threshold] = 0\n",
    "        labels[labels > poster_threshold] = 1\n",
    "        # apply sigmoid on the outputs\n",
    "        outputs = torch.sigmoid(outputs)\n",
    "        # process the result, bigger than confidence is reliable\n",
    "        predicted = (outputs > confidence).float()     # tensor size ([16,128,128]), True->1.0 and False->0.0\n",
    "        intersection = torch.logical_and(predicted, labels).sum((1, 2))\n",
    "        union = torch.logical_or(predicted, labels).sum((1, 2))\n",
    "        iou = (intersection / (union + 1e-7))  # 每个样本的IoU\n",
    "        dice = (2 * intersection / (predicted.sum((1, 2)) + labels.sum((1, 2))+1e-7))  # 每个样本的Dice Coefficient\n",
    "        # 计算训练集的像素精度\n",
    "        batch_correct_pixels,bacth_total_pixels = calculate_pixel_accuracy(predicted, labels)\n",
    "\n",
    "        loss = criterion(outputs, labels)\n",
    "\n",
    "        train_loss += loss.item()\n",
    "        total_iou += torch.sum(iou).item()\n",
    "        total_dice += torch.sum(dice).item()\n",
    "        total_correct_pixels += batch_correct_pixels\n",
    "        total_pixels += bacth_total_pixels\n",
    "\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "    # train set results\n",
    "    avg_loss = train_loss / len(dataset)\n",
    "    avg_iou = total_iou / len(dataset)\n",
    "    avg_dice = total_dice / len(dataset)\n",
    "    avg_acc_train = total_correct_pixels / total_pixels\n",
    "\n",
    "    # test set results\n",
    "    with torch.no_grad():\n",
    "        for inputs, targets in test_loader:     \n",
    "            inputs = inputs.to(device)\n",
    "            targets = targets.to(device)\n",
    "            targets = targets * 255\n",
    "            targets[targets <= poster_threshold] = 0\n",
    "            targets[targets > poster_threshold] = 1\n",
    "            outputs = model(inputs)\n",
    "            # squeeze the dimension that pytorch add for channels\n",
    "            outputs = outputs.squeeze(1)     # tensor size ([16,128,128])\n",
    "            targets = targets.squeeze(1)\n",
    "            outputs = torch.sigmoid(outputs)\n",
    "            # process the result, bigger than 0.5 is reliable\n",
    "            predicted = (outputs > confidence).float()     # tensor size ([16,128,128])\n",
    "            intersection = torch.logical_and(predicted, targets).sum((1, 2))\n",
    "            union = torch.logical_or(predicted, targets).sum((1, 2))\n",
    "            iou = (intersection / (union + 1e-7))  # 每个样本的IoU\n",
    "            dice = (2 * intersection / (predicted.sum((1, 2)) + targets.sum((1, 2))+1e-7))  # 每个样本的Dice Coefficient\n",
    "\n",
    "\n",
    "            batch_correct_pixels,batch_total_pixels = calculate_pixel_accuracy(predicted, targets)\n",
    "            total_iou_test += torch.sum(iou).item()\n",
    "            total_dice_test += torch.sum(dice).item()\n",
    "            total_correct_pixels_test += batch_correct_pixels\n",
    "            total_pixels_test += batch_total_pixels\n",
    "\n",
    "    test_iou = total_iou_test / len(testset)\n",
    "    test_dice = total_dice_test / len(testset)\n",
    "    test_acc = total_correct_pixels_test / total_pixels_test\n",
    "\n",
    "    # get the best IoU and Dice\n",
    "    best_loss_train = best_loss_train if best_loss_train < avg_loss else avg_loss\n",
    "    best_IoU_train = best_IoU_train if best_IoU_train > avg_iou else avg_iou\n",
    "    best_Dice_train = best_Dice_train if best_Dice_train > avg_dice else avg_dice\n",
    "    best_IoU_test = best_IoU_test if best_IoU_test > test_iou else test_iou\n",
    "    # best_Dice_test = best_Dice_test if best_Dice_test > test_dice else test_dice\n",
    "\n",
    "    if(best_Dice_test <= test_dice):\n",
    "        best_Dice_test = test_dice\n",
    "        torch.save(model.state_dict(), f\"./model/MANet_checkpoint.pth\")\n",
    "\n",
    "    # get the best accuracy of the trainloader and testloader\n",
    "    best_acc_train = best_acc_train if best_acc_train > avg_acc_train else avg_acc_train\n",
    "    best_acc_test = best_acc_test if best_acc_test > test_acc else test_acc\n",
    "\n",
    "    # store the data in each epoch\n",
    "    loss_train_list.append(avg_loss)\n",
    "    IoU_train_list.append(avg_iou)\n",
    "    Dice_train_list.append(avg_dice)\n",
    "    IoU_test_list.append(test_iou)\n",
    "    Dice_test_list.append(test_dice)\n",
    "    acc_train_list.append(avg_acc_train)\n",
    "    acc_test_list.append(test_acc)\n",
    "\n",
    "    print(f\"Epoch [{epoch+1}/{num_epochs}], Avg Loss: {avg_loss:.5f}, Avg IoU: {avg_iou:.4f}, Avg Dice: {avg_dice:.4f},Test IoU: {test_iou:.4f}, Test Dice: {test_dice:.4f}, Train Acc: {avg_acc_train:.4f}, Test Acc: {test_acc:.4f}\")\n",
    "    \n",
    "print(\"-----------------------------------------\")\n",
    "print(f\"Best Loss: {best_loss_train:.4f}, Best Train IoU: {best_IoU_train:.4f}, Best Train Dice: {best_Dice_train:.4f}, Best Test IoU: {best_IoU_test:.4f}, Best Test Dice: {best_Dice_test:.4f}, Best Train Acc: {best_acc_train:.4f}, Best Test Acc: {best_acc_test:.4f}\")\n",
    "# draw the plot for the loss and dice\n",
    "# 定义epochs的数量，用于x轴坐标\n",
    "epochs = len(loss_train_list)\n",
    "fit,(ax1,ax2) = plt.subplots(2,1,figsize=(8,8))\n",
    "# 绘制训练损失曲线\n",
    "ax1.plot(range(1, epochs + 1), loss_train_list, label='Train Loss')\n",
    "# 设置图例1\n",
    "ax1.legend()\n",
    "ax1.set_xlabel('Epochs')\n",
    "ax1.set_ylabel('Loss')\n",
    "ax1.set_title('Training Loss')\n",
    "ax1.grid(True)\n",
    "# 绘制IoU和Dice\n",
    "ax2.plot(range(1,epochs +1),IoU_train_list,label='Train IoU')\n",
    "ax2.plot(range(1,epochs +1),Dice_train_list,label='Train Dice')\n",
    "# 设置图例2\n",
    "ax2.legend()\n",
    "ax2.set_xlabel('Epochs')\n",
    "ax2.set_ylabel('IoU and Dice')\n",
    "ax2.set_title('Training IoU and Dice')\n",
    "ax2.grid(True)# show the plots\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "# empty the cache for model and training process\n",
    "torch.cuda.empty_cache()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Test IoU: 0.8830, Best Test Dice: 0.9236 Best Test Acc: 0.9958\n"
     ]
    }
   ],
   "source": [
    "# Using the best model to evaluate on the test set\n",
    "\n",
    "model = smp.MAnet(encoder_name='resnet34', encoder_depth=5, encoder_weights='imagenet', decoder_use_batchnorm=True, decoder_channels=(256, 128, 64, 32, 16), decoder_pab_channels=64, in_channels=1, classes=1, activation=None, aux_params=None)\n",
    "\n",
    "# 设置设备\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = model.to(device)\n",
    "\n",
    "test_path = \"./dataset/Spineweb_dataset15/test_1/\"\n",
    "\n",
    "transform = transforms.Compose([\n",
    "    # transforms.ToPILImage(),  # 转换为 PIL 图像对象\n",
    "    transforms.Resize((128, 128)),  # 调整大小为 128x128\n",
    "    transforms.ToTensor()  # 转换为张量\n",
    "])\n",
    "\n",
    "# 加载已保存的模型参数\n",
    "checkpoint_path = \"./model/MANet_checkpoint.pth\"  # 模型文件路径\n",
    "checkpoint = torch.load(checkpoint_path)\n",
    "model.load_state_dict(checkpoint)\n",
    "\n",
    "# 设置模型为推理模式\n",
    "model.eval()\n",
    "\n",
    "batch_size = 128\n",
    "\n",
    "confidence = 0.9\n",
    "\n",
    "testset = SpineWeb15(data_path=test_path, transform1=transform,transform2=transform)\n",
    "test_loader = DataLoader(testset, batch_size=batch_size, shuffle=True)\n",
    "\n",
    "# data_iter = iter(test_loader)\n",
    "# images, labels = next(data_iter)\n",
    "# print(\"加载的样本数量:\", images.shape[0])\n",
    "\n",
    "total_iou_test = 0.0\n",
    "total_dice_test = 0.0\n",
    "total_acc_test = 0.0\n",
    "total_correct_pixels_test = 0\n",
    "total_pixels_test = 0\n",
    "\n",
    "# test set results\n",
    "with torch.no_grad():\n",
    "    for inputs, targets in test_loader:\n",
    "        inputs = inputs.to(device)\n",
    "        targets = targets.to(device)\n",
    "        targets = targets * 255\n",
    "        targets[targets <= poster_threshold] = 0\n",
    "        targets[targets > poster_threshold] = 1\n",
    "        outputs = model(inputs)\n",
    "        # squeeze the dimension that pytorch add for channels\n",
    "        outputs = outputs.squeeze(1)     # tensor size ([16,128,128])\n",
    "        targets = targets.squeeze(1)\n",
    "        outputs = torch.sigmoid(outputs)\n",
    "        # process the result, bigger than 0.5 is reliable\n",
    "        predicted = (outputs > confidence).float()     # tensor size ([16,128,128])\n",
    "        intersection = torch.logical_and(predicted, targets).sum((1, 2))\n",
    "        union = torch.logical_or(predicted, targets).sum((1, 2))\n",
    "        iou = (intersection / (union + 1e-7))  # 平均每个样本的IoU\n",
    "        dice = (2 * intersection / (predicted.sum((1, 2)) + targets.sum((1, 2))+1e-7))  # 平均每个样本的Dice Coefficient\n",
    "\n",
    "\n",
    "        batch_correct_pixels,batch_total_pixels = calculate_pixel_accuracy(predicted, targets)\n",
    "        total_iou_test += torch.sum(iou).item()\n",
    "        total_dice_test += torch.sum(dice).item()\n",
    "        total_correct_pixels_test += batch_correct_pixels\n",
    "        total_pixels_test += batch_total_pixels\n",
    "        \n",
    "inputs_npimage = inputs.squeeze(1)[0].detach().cpu().numpy()\n",
    "predicted_npimage = predicted[0].detach().cpu().numpy()\n",
    "targets_npimage = targets[0].detach().cpu().numpy()\n",
    "fig, axes = plt.subplots(3, 1)\n",
    "axes[0].imshow(inputs_npimage, cmap='gray')\n",
    "axes[0].set_title('Original')\n",
    "axes[1].imshow(predicted_npimage, cmap='gray')\n",
    "axes[1].set_title('Predicted')\n",
    "axes[2].imshow(targets_npimage, cmap='gray')\n",
    "axes[2].set_title('Target')\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "test_iou = total_iou_test / len(testset)\n",
    "test_dice = total_dice_test / len(testset)\n",
    "test_acc = total_correct_pixels_test / total_pixels_test\n",
    "\n",
    "# output the result, the result may vary a little due to the resize function in transform\n",
    "print(f\"Best Test IoU: {test_iou:.4f}, Best Test Dice: {test_dice:.4f} Best Test Acc: {test_acc:.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 3 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best Train IoU: 0.8993, Best Train Dice: 0.9347 Best Train Acc: 0.9972\n"
     ]
    }
   ],
   "source": [
    "# Using the best model to evaluate on the train set\n",
    "\n",
    "model = smp.MAnet(encoder_name='resnet34', encoder_depth=5, encoder_weights='imagenet', decoder_use_batchnorm=True, decoder_channels=(256, 128, 64, 32, 16), decoder_pab_channels=64, in_channels=1, classes=1, activation=None, aux_params=None)\n",
    "\n",
    "# 设置设备\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "model = model.to(device)\n",
    "\n",
    "test_path = \"./dataset/Spineweb_dataset15/train_9/\"\n",
    "\n",
    "transform = transforms.Compose([\n",
    "    # transforms.ToPILImage(),  # 转换为 PIL 图像对象\n",
    "    transforms.Resize((128, 128)),  # 调整大小为 128x128\n",
    "    transforms.ToTensor()  # 转换为张量\n",
    "])\n",
    "\n",
    "# 加载已保存的模型参数\n",
    "checkpoint_path = \"./model/MANet_checkpoint.pth\"  # 模型文件路径\n",
    "checkpoint = torch.load(checkpoint_path)\n",
    "model.load_state_dict(checkpoint)\n",
    "\n",
    "# 设置模型为推理模式\n",
    "model.eval()\n",
    "\n",
    "batch_size = 128\n",
    "\n",
    "confidence = 0.9\n",
    "\n",
    "testset = SpineWeb15(data_path=test_path, transform1=transform, transform2=transform)\n",
    "test_loader = DataLoader(testset, batch_size=batch_size, shuffle=True)\n",
    "\n",
    "# data_iter = iter(test_loader)\n",
    "# images, labels = next(data_iter)\n",
    "# print(\"加载的样本数量:\", images.shape[0])\n",
    "\n",
    "total_iou_test = 0.0\n",
    "total_dice_test = 0.0\n",
    "total_acc_test = 0.0\n",
    "total_correct_pixels_test = 0\n",
    "total_pixels_test = 0\n",
    "\n",
    "# test set results\n",
    "with torch.no_grad():\n",
    "    for inputs, targets in test_loader:\n",
    "        inputs = inputs.to(device)\n",
    "        targets = targets.to(device)\n",
    "        targets = targets * 255\n",
    "        targets[targets <= poster_threshold] = 0\n",
    "        targets[targets > poster_threshold] = 1\n",
    "        outputs = model(inputs)\n",
    "        # squeeze the dimension that pytorch add for channels\n",
    "        outputs = outputs.squeeze(1)     # tensor size ([16,128,128])\n",
    "        targets = targets.squeeze(1)\n",
    "        outputs = torch.sigmoid(outputs)\n",
    "        # process the result, bigger than 0.5 is reliable\n",
    "        predicted = (outputs > confidence).float()     # tensor size ([16,128,128])\n",
    "        intersection = torch.logical_and(predicted, targets).sum((1, 2))\n",
    "        union = torch.logical_or(predicted, targets).sum((1, 2))\n",
    "        iou = (intersection / (union + 1e-7))  # 平均每个样本的IoU\n",
    "        dice = (2 * intersection / (predicted.sum((1, 2)) + targets.sum((1, 2))+1e-7))  # 平均每个样本的Dice Coefficient\n",
    "\n",
    "\n",
    "        batch_correct_pixels,batch_total_pixels = calculate_pixel_accuracy(predicted, targets)\n",
    "        total_iou_test += torch.sum(iou).item()\n",
    "        total_dice_test += torch.sum(dice).item()\n",
    "        total_correct_pixels_test += batch_correct_pixels\n",
    "        total_pixels_test += batch_total_pixels\n",
    "        \n",
    "\n",
    "inputs_npimage = inputs.squeeze(1)[0].detach().cpu().numpy()\n",
    "predicted_npimage = predicted[0].detach().cpu().numpy()\n",
    "targets_npimage = targets[0].detach().cpu().numpy()\n",
    "fig, axes = plt.subplots(3, 1)\n",
    "axes[0].imshow(inputs_npimage, cmap='gray')\n",
    "axes[0].set_title('Original')\n",
    "axes[1].imshow(predicted_npimage, cmap='gray')\n",
    "axes[1].set_title('Predicted')\n",
    "axes[2].imshow(targets_npimage, cmap='gray')\n",
    "axes[2].set_title('Target')\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "\n",
    "\n",
    "test_iou = total_iou_test / len(testset)\n",
    "test_dice = total_dice_test / len(testset)\n",
    "test_acc = total_correct_pixels_test / total_pixels_test\n",
    "\n",
    "# output the result, the result may vary a little due to the resize function in transform\n",
    "print(f\"Best Train IoU: {test_iou:.4f}, Best Train Dice: {test_dice:.4f} Best Train Acc: {test_acc:.4f}\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dataEngineering",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.0"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}