diff --git a/cv/unet/train_unet_segmentation.ipynb b/cv/unet/train_unet_segmentation.ipynb
index 04a9f1c..98d968a 100644
--- a/cv/unet/train_unet_segmentation.ipynb
+++ b/cv/unet/train_unet_segmentation.ipynb
@@ -1,1466 +1,1082 @@
-{
- "cells": [
-  {
-   "cell_type": "markdown",
-   "id": "f060eb4c",
-   "metadata": {},
-   "source": [
-    "## 基于MindSpore框架的UNet-2D案例实现\n",
-    "\n",
-    "### 1 模型简介\n",
-    "\n",
-    "Unet模型于2015年在论文《U-Net: Convolutional Networks for Biomedical Image Segmentation》中被提出，最初的提出是为了解决医学图像分割问题，用于细胞层面的图像分割任务。\n",
-    "\n",
-    "Unet模型是在FCN网络的基础上构建的，但由于FCN无法获取上下文信息以及位置信息，导致准确性较低，Unet模型由此引入了U型结构获取上述两种信息，并且模型结构简单高效、容易构建，在较小的数据集上也能实现较高的准确率。\n",
-    "\n",
-    "#### 1.1 模型结构\n",
-    "Unet模型的整体结构由两部分组成，即特征提取网络和特征融合网络，其结构也被称为“编码器-解码器结构”，并且由于网络整体结构类似于大写的英文字母“U”，故得名Unet，在其原始论文中定义的网络结构如图1所示。\n",
-    "\n",
-    "<center>\n",
-    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bw5io8tqj20n10fa3zo.jpg\" alt=\"image-20220819101847606\" style=\"zoom:67%;\" />\n",
-    "    <br>\n",
-    "    <div style=\"color:orange;\n",
-    "    display: inline-block;\n",
-    "    color: #999;\n",
-    "    padding: 2px;\">图1 Unet网络结构图</div>\n",
-    "</center>\n",
-    "\n",
-    "整个模型结构就是在原始图像输入后，首先进行特征提取，再进行特征融合：\n",
-    "\n",
-    "a) 左半部分负责特征提取的网络结构（即编码器结构）需要利用两个3x3的卷积核与2x2的池化层组成一个“下采样模块”，每一个下采样模块首先会对特征图进行两次valid卷积，再进行一次池化操作。由此经过4个下采样模块后，原始尺寸为572x572大小、通道数为1的原始图像，转换为了大小为28x28、通道数为1024的特征图。\n",
-    "\n",
-    "b) 右半部分负责进行上采样的网络结构（即解码器结构）需要利用1次反卷积操作、特征拼接操作以及两个3x3的卷积核作为一个“上采样模块”，每一个上采样模块首先会对特征图通过反卷积操作使图像尺寸增加1倍，再通过拼接编码器结构中的特征图使得通道数增加，最后经过两次valid卷积。由此经过4个上采样模块后，经过下采样模块的、大小为28x28、通道数为1024的特征图，转换为了大小为388x388、通道数为64的特征图。\n",
-    "\n",
-    "c) 网络结构的最后一部分是通过两个1x1的卷积核将经过上采样得到的通道数为64的特征图，转换为了通道数为2的图像作为预测结果输出。\n",
-    "#### 1.2 模型特点\n",
-    "\n",
-    "a) 利用拼接操作将低级特征图与高级特征图进行特征融合\n",
-    "\n",
-    "b) 完全对称的U型结构使得高分辨率信息和低分辨率信息在目标图片中增加，前后特征融合更为彻底。\n",
-    "\n",
-    "c) 结合了下采样时的低分辨率信息（提供物体类别识别依据）和上采样时的高分辨率信息（提供精准分割定位依据），此外还通过融合操作填补底层信息以提高分割精度。\n"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "ce58c71c",
-   "metadata": {},
-   "source": [
-    "### 2 案例实现\n",
-    "\n",
-    "#### 2.1 **环境准备与数据读取**\n",
-    "\n",
-    "本案例基于MindSpore-CPU版本实现，在CPU上完成模型训练。\n",
-    "\n",
-    "案例实现所使用的数据即ISBI果蝇电镜图数据集，可以从http://brainiac2.mit.edu/isbi_challenge/ 中下载，下载好的数据集包括3个tif文件，分别对应测试集样本、训练集标签、训练集样本，文件路径结构如下：\n",
-    "\n",
-    "```\n",
-    ".datasets/\n",
-    "└── ISBI\n",
-    "    ├── test-volume.tif\n",
-    "    ├── train-labels.tif\n",
-    "    └── train-volume.tif\n",
-    "```\n",
-    "\n",
-    "其中每个tif文件都由30副图片压缩而成，所以接下来需要获取每个tif文件中所存储的所有图片，将其转换为png格式存储，得到训练集样本对应的30张png图片、训练集标签对应的30张png图片以及测试集样本对应的30张png图片。"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 1,
-   "id": "6ae6e6ec",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "/home/ma-user/anaconda3/envs/MindSpore/bin/python\n",
-      "MindSpore version:  1.8.1\n",
-      "The result of multiplication calculation is correct, MindSpore has been installed successfully!\n",
-      "None\n"
-     ]
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 12 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 12 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "import sys\n",
-    "import mindspore\n",
-    "print(sys.executable)\n",
-    "print(mindspore.run_check())\n",
-    "from PIL import Image, ImageSequence\n",
-    "import math\n",
-    "import numpy as np\n",
-    "import matplotlib.pyplot as plt\n",
-    "#显示下载好的数据\n",
-    "train_image_path = \"data/train-volume.tif\"\n",
-    "train_masks_path = \"data/train-labels.tif\"\n",
-    "image = np.array([np.array(p) for p in ImageSequence.Iterator(Image.open(train_image_path))])\n",
-    "masks = np.array([np.array(p) for p in ImageSequence.Iterator(Image.open(train_masks_path))])\n",
-    "\n",
-    "def show_image(image_list,num = 6):\n",
-    "    img_titles = []\n",
-    "    img_draws = []\n",
-    "    for ind,img in enumerate(image_list):\n",
-    "        if ind == num:\n",
-    "            break\n",
-    "        img_titles.append(ind)\n",
-    "        img_draws.append(img)\n",
-    "\n",
-    "    for i in range(len(img_titles)):\n",
-    "        if len(img_titles) > 6:\n",
-    "            row = 3\n",
-    "        elif 3<len(img_titles)<=6:\n",
-    "            row = 2\n",
-    "        else:\n",
-    "            row = 1\n",
-    "        col = math.ceil(len(img_titles)/row)\n",
-    "        plt.subplot(row,col,i+1),plt.imshow(img_draws[i],'gray')\n",
-    "        plt.title(img_titles[i])\n",
-    "        plt.xticks([]),plt.yticks([])\n",
-    "    plt.show()\n",
-    "    \n",
-    "show_image(image,num = 12)\n",
-    "show_image(masks,num = 12)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "294ec00c",
-   "metadata": {},
-   "source": [
-    "具体的实现方式首先是将tif文件转换为数组形式，之后通过io操作将每张图片对应的数组存储为png图像，处理过后的训练集样本及其对应的标签图像如图2所示。将3个tif文件转换为png格式后，针对训练集的样本与标签，将其以2:1的比例，重新划分为了训练集与验证集，划分完成后的文件路径结构如下：\n",
-    "\n",
-    "```\n",
-    ".datasets/\n",
-    "└── ISBI\n",
-    "    ├── test_imgs\n",
-    "    │   ├── 00000.png\n",
-    "    │   ├── 00001.png\n",
-    "    │   └── . . . . .\n",
-    "    ├── train\n",
-    "    │   ├── image\n",
-    "    │   │   ├── 00001.png\n",
-    "    │   │   ├── 00002.png\n",
-    "    │   │   └── . . . . .\n",
-    "    │   └── mask\n",
-    "    │       ├── 00001.png\n",
-    "    │       ├── 00002.png\n",
-    "    │       └── . . . . .\n",
-    "    └── val\n",
-    "        ├── image\n",
-    "        │   ├── 00000.png\n",
-    "        │   ├── 00003.png\n",
-    "        │   └── . . . . .\n",
-    "        └── mask\n",
-    "            ├── 00000.png\n",
-    "            ├── 00003.png\n",
-    "            └── . . . . .\n",
-    "```\n",
-    "\n",
-    "<center>\n",
-    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwhda4g2j20fz0chabb.jpg\" alt=\"image-20220819101847606\" style=\"zoom:50%;\" />\n",
-    "    <br>\n",
-    "    <div style=\"color:orange;\n",
-    "    display: inline-block;\n",
-    "    color: #999;\n",
-    "    padding: 2px;\">图2 训练集样本及其对应标签</div>\n",
-    "</center>\n",
-    "\n",
-    "#### 2.2 数据集创建\n",
-    "\n",
-    "在进行上述tif文件格式转换，以及测试集和验证集的进一步划分后，就完成了数据读取所需的所有工作，接下来就需要利用处理好的图像数据，通过一定的图像变换来进行数据增强，并完成数据集的创建。\n",
-    "\n",
-    "数据增强部分是引入了mindspore.dataset.vision，针对训练集样本和标签，首先通过A.resize()方法将图像尺寸重新调整为统一大小，之后再进行转置以及水平翻转、垂直翻转，完成针对训练集样本和标签的数据增强。针对验证集的样本和标签，仅通过resize()方法将图像尺寸重新调整为统一大小。\n",
-    "\n",
-    "其次数据集的创建部分，首先是定义了Data_Loader类，在该类的__init__函数中，根据传入的data_path参数，确定在数据读取阶段设置好的、训练集和验证集的存储路径，再设置对应的样本和标签路径，并针对训练集和验证集的不同数据增强方法。在该类的__getitem__函数中，通过传入索引值读取训练集或验证集存储路径下的样本和标签图像，并对图像进行对应的数据增强操作，之后再对样本和标签的形状进行转置，就完成了__getitem__函数对样本和标签图像的读取。最后通过定义create_dataset函数，传入data_dir、batch_size等参数，在函数中实例化Data_Loader类获取data_dir，也就是训练集或验证集对应路径下的样本和标签元组对，再通过mindspore.dataset中的GeneratorDataset将元组转换为Tensor，最后通过设定好的batch_size将样本和标签按照batch_size大小分组，由此完成数据集的创建，上述流程对应代码如下："
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 2,
-   "id": "dc6df9cd",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import os\n",
-    "import cv2\n",
-    "import mindspore.dataset as ds\n",
-    "import glob\n",
-    "import mindspore.dataset.vision as vision_C  #.c_transforms\n",
-    "import mindspore.dataset.transforms as C_transforms #.c_transform\n",
-    "import random\n",
-    "import mindspore\n",
-    "from mindspore.dataset.vision import Inter\n",
-    "\n",
-    "def train_transforms(img_size):\n",
-    "    return [\n",
-    "    vision_C.Resize(img_size, interpolation=Inter.NEAREST),\n",
-    "    vision_C.Rescale(1./255., 0.0),\n",
-    "    vision_C.RandomHorizontalFlip(prob=0.5),\n",
-    "    vision_C.RandomVerticalFlip(prob=0.5),\n",
-    "    vision_C.HWC2CHW()\n",
-    "    ]\n",
-    "\n",
-    "\n",
-    "def val_transforms(img_size):\n",
-    "    return [\n",
-    "    vision_C.Resize(img_size, interpolation=Inter.NEAREST),\n",
-    "    vision_C.Rescale(1/255., 0),\n",
-    "    vision_C.HWC2CHW()\n",
-    "    ]\n",
-    "\n",
-    "\n",
-    "\n",
-    "class Data_Loader:\n",
-    "    def __init__(self, data_path):\n",
-    "        # 初始化函数，读取所有data_path下的图片\n",
-    "        self.data_path = data_path\n",
-    "        self.imgs_path = glob.glob(os.path.join(data_path, 'image/*.png'))\n",
-    "        self.label_path = glob.glob(os.path.join(data_path, 'mask/*.png'))\n",
-    "\n",
-    "    def __getitem__(self, index):\n",
-    "        # 根据index读取图片\n",
-    "        image = cv2.imread(self.imgs_path[index])\n",
-    "        label = cv2.imread(self.label_path[index], cv2.IMREAD_GRAYSCALE)\n",
-    "        label = label.reshape((label.shape[0], label.shape[1], 1))\n",
-    "    \n",
-    "        return image, label\n",
-    "\n",
-    "    @property\n",
-    "    def column_names(self):\n",
-    "        column_names = ['image', 'label']\n",
-    "        return column_names\n",
-    "\n",
-    "    def __len__(self):\n",
-    "        # 返回训练集大小\n",
-    "        return len(self.imgs_path)\n",
-    "\n",
-    "\n",
-    "def create_dataset(data_dir, img_size, batch_size, augment, shuffle):\n",
-    "    mc_dataset = Data_Loader(data_path=data_dir)\n",
-    "    dataset = ds.GeneratorDataset(mc_dataset, mc_dataset.column_names, shuffle=shuffle)\n",
-    "\n",
-    "    if augment:\n",
-    "        transform_img = train_transforms(img_size)\n",
-    "    else:\n",
-    "        transform_img = val_transforms(img_size)\n",
-    "\n",
-    "    seed = random.randint(1,1000)\n",
-    "    mindspore.set_seed(seed)\n",
-    "    dataset = dataset.map(input_columns='image', num_parallel_workers=1, operations=transform_img)\n",
-    "    mindspore.set_seed(seed)\n",
-    "    dataset = dataset.map(input_columns=\"label\", num_parallel_workers=1, operations=transform_img)\n",
-    "\n",
-    "    if shuffle:\n",
-    "        dataset = dataset.shuffle(buffer_size=10000)\n",
-    "    dataset = dataset.batch(batch_size, num_parallel_workers=1)\n",
-    "    if augment == True and shuffle == True:\n",
-    "        print(\"训练集数据量：\", len(mc_dataset))\n",
-    "    elif augment == False and shuffle == False:\n",
-    "        print(\"验证集数据量：\", len(mc_dataset))\n",
-    "    else:\n",
-    "        pass\n",
-    "    return dataset"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 3,
-   "id": "0170fc4a",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "验证集数据量： 10\n",
-      "Shape of image [N, C, H, W]: (3, 3, 224, 224) Float32 --- Shape of label [N, C, H, W]: (3, 1, 224, 224) Float32\n",
-      "Shape of image [N, C, H, W]: (3, 3, 224, 224) Float32 --- Shape of label [N, C, H, W]: (3, 1, 224, 224) Float32\n",
-      "Shape of image [N, C, H, W]: (3, 3, 224, 224) Float32 --- Shape of label [N, C, H, W]: (3, 1, 224, 224) Float32\n",
-      "Shape of image [N, C, H, W]: (1, 3, 224, 224) Float32 --- Shape of label [N, C, H, W]: (1, 1, 224, 224) Float32\n"
-     ]
-    }
-   ],
-   "source": [
-    "if __name__ == '__main__':\n",
-    "    train_dataset = create_dataset('src/datasets/ISBI/val', img_size=224, batch_size=3, augment=False, shuffle=False)\n",
-    "    for item, (image, label) in enumerate(train_dataset):\n",
-    "        if item < 5:\n",
-    "            print(f\"Shape of image [N, C, H, W]: {image.shape} {image.dtype}\",'---',f\"Shape of label [N, C, H, W]: {label.shape} {label.dtype}\")"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "7eb80c9d",
-   "metadata": {},
-   "source": [
-    "#### 2.3 模型构建\n",
-    "\n",
-    "本案例实现中所构建的Unet模型结构与2015年论文中提出的Unet结构大致相同，但本案例中Unet网络模型的“下采样模块”与“上采样模块”使用的卷积类型都为Same卷积，而原论文中使用的是Valid卷积。此外，原论文的网络模型最终使用两个1x1的卷积核，输出了通道数2的预测图像，而本案例的网络模型最终使用的是1个1x1的卷积核，输出通道数为1的灰度图，和标签图像格式保持一致。实际构建的Unet模型结构如图3所示。\n",
-    "\n",
-    "<center>\n",
-    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwhpesqpj20no0g7jso.jpg\" alt=\"image-20220819101847606\" style=\"zoom:67%;\" />\n",
-    "    <br>\n",
-    "    <div style=\"color:orange;\n",
-    "    display: inline-block;\n",
-    "    color: #999;\n",
-    "    padding: 2px;\">图3 实际构建的Unet模型结构</div>\n",
-    "</center>\n",
-    "\n",
-    "MindSpore框架构建网络的流程与PyTorch类似，在定义模型类时需要继承Cell类，并重写__init__和construct方法。具体的实现方式首先是定义了一个double_conv模型类，在类中重写__init__方法，通过使用nn.Conv2d层定义“下采样模块”与“上采样模块”中都使用到的两个卷积函数，并且在每个卷积层后加入nn.BatchNorm2d层来对每次卷积后的特征图进行标准化，防止过拟合，以及使用nn.ReLU层加入非线性的激活函数。之后在construct方法中使用定义好的运算构建前向网络。\n",
-    "\n",
-    "在doubel_conv模型类定义好之后，接下来就是通过定义UNet模型类来完成整个UNet网络的构建。在UNet模型类的__init__方法中实例化double_conv类来表示两个连续的卷积层，接着使用nn.MaxPool2d来进行最大池化，由此完成了1个“下采样模块”的构建，重复4次即可完成网络中的编码器部分。针对解码器部分，使用了nn.ResizeBilinear层来表示反卷积层，接着实例化了double_conv类来表示两个卷积层，由此完成了1个“上采样模块”的构建，重复4次即完成网络中解码器部分的搭建。之后通过1个nn.Conv2d层来完成预测图像的输出。最后在construct方法中使用定义好的运算构建前向网络，由此完成整个Unet网络模型的构建。上述构建流程的对应代码如下所示：\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 4,
-   "id": "791b240a",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "from mindspore import nn\n",
-    "import mindspore.numpy as np\n",
-    "import mindspore.ops as ops\n",
-    "import mindspore.ops.operations as F\n",
-    "\n",
-    "def double_conv(in_ch, out_ch):\n",
-    "    return nn.SequentialCell(nn.Conv2d(in_ch, out_ch, 3),\n",
-    "                              nn.BatchNorm2d(out_ch), nn.ReLU(),\n",
-    "                              nn.Conv2d(out_ch, out_ch, 3),\n",
-    "                              nn.BatchNorm2d(out_ch), nn.ReLU())\n",
-    "class UNet(nn.Cell):\n",
-    "    def __init__(self, in_ch = 3, n_classes = 1):\n",
-    "        super(UNet, self).__init__()\n",
-    "        self.concat1 = F.Concat(axis=1)\n",
-    "        self.concat2 = F.Concat(axis=1)\n",
-    "        self.concat3 = F.Concat(axis=1)\n",
-    "        self.concat4 = F.Concat(axis=1)\n",
-    "        self.double_conv1 = double_conv(in_ch, 64)\n",
-    "        self.maxpool1 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
-    "        self.double_conv2 = double_conv(64, 128)\n",
-    "        self.maxpool2 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
-    "        self.double_conv3 = double_conv(128, 256)\n",
-    "        self.maxpool3 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
-    "        self.double_conv4 = double_conv(256, 512)\n",
-    "        self.maxpool4 = nn.MaxPool2d(kernel_size=2, stride=2)\n",
-    "        self.double_conv5 = double_conv(512, 1024)\n",
-    "\n",
-    "        self.upsample1 = nn.ResizeBilinear()\n",
-    "        self.double_conv6 = double_conv(1024 + 512, 512)\n",
-    "        self.upsample2 = nn.ResizeBilinear()\n",
-    "        self.double_conv7 = double_conv(512 + 256, 256)\n",
-    "        self.upsample3 = nn.ResizeBilinear()\n",
-    "        self.double_conv8 = double_conv(256 + 128, 128)\n",
-    "        self.upsample4 = nn.ResizeBilinear()\n",
-    "        self.double_conv9 = double_conv(128 + 64, 64)\n",
-    "\n",
-    "        self.final = nn.Conv2d(64, n_classes, 1)\n",
-    "        self.sigmoid = ops.Sigmoid()\n",
-    "\n",
-    "    def construct(self, x):\n",
-    "\n",
-    "        feature1 = self.double_conv1(x)\n",
-    "        tmp = self.maxpool1(feature1)\n",
-    "        feature2 = self.double_conv2(tmp)\n",
-    "        tmp = self.maxpool2(feature2)\n",
-    "        feature3 = self.double_conv3(tmp)\n",
-    "        tmp = self.maxpool3(feature3)\n",
-    "        feature4 = self.double_conv4(tmp)\n",
-    "        tmp = self.maxpool4(feature4)\n",
-    "        feature5 = self.double_conv5(tmp)\n",
-    "\n",
-    "        up_feature1 = self.upsample1(feature5, scale_factor=2)\n",
-    "        tmp = self.concat1((feature4, up_feature1))\n",
-    "        tmp = self.double_conv6(tmp)\n",
-    "        up_feature2 = self.upsample2(tmp, scale_factor=2)\n",
-    "        tmp = self.concat2((feature3, up_feature2))\n",
-    "        tmp = self.double_conv7(tmp)\n",
-    "        up_feature3 = self.upsample3(tmp, scale_factor=2)\n",
-    "        tmp = self.concat3((feature2, up_feature3))\n",
-    "        tmp = self.double_conv8(tmp)\n",
-    "        up_feature4 = self.upsample4(tmp, scale_factor=2)\n",
-    "        tmp = self.concat4((feature1, up_feature4))\n",
-    "        tmp = self.double_conv9(tmp)\n",
-    "        output = self.sigmoid(self.final(tmp))\n",
-    "\n",
-    "        return output"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0afb1c99",
-   "metadata": {},
-   "source": [
-    "#### 2.4 自定义评估指标\n",
-    "\n",
-    "为了能够更加全面和直观的观察网络模型训练效果，本案例实现中还使用了MindSpore框架来自定义Metrics，在自定义的metrics类中使用了多种评价函数来评估模型的好坏，分别为准确率Acc、交并比IoU、Dice系数、灵敏度Sens、特异性Spec。\n",
-    "\n",
-    "a) 其中准确率Acc是图像中正确分类的像素百分比。即分类正确的像素占总像素的比例，用公式可表示为：\n",
-    "$$\n",
-    "A c c=\\frac{T P+T N}{T P+T N+F P+F N}\n",
-    "$$\n",
-    "其中：\n",
-    "\n",
-    "- TP：真阳性数，在label中为阳性，在预测值中也为阳性的个数。\n",
-    "- TN：真阴性数，在label中为阴性，在预测值中也为阴性的个数。\n",
-    "- FP：假阳性数，在label中为阴性，在预测值中为阳性的个数。\n",
-    "- FN：假阴性数，在label中为阳性，在预测值中为阴性的个数。\n",
-    "\n",
-    "b) 交并比IoU是预测分割和标签之间的重叠区域除以预测分割和标签之间的联合区域（两者的交集/两者的并集），是语义分割中最常用的指标之一，其计算公式为：\n",
-    "$$\n",
-    "I o U=\\frac{|A \\cap B|}{|A \\cup B|}=\\frac{T P}{T P+F P+F N}\n",
-    "$$\n",
-    "c) Dice系数定义为两倍的交集除以像素和，也叫F1 score，与IoU呈正相关关系，其计算公式为：\n",
-    "$$\n",
-    "\\text { Dice }=\\frac{2|A \\cap B|}{|A|+|B|}=\\frac{2 T P}{2 T P+F P+F N}\n",
-    "$$\n",
-    "d) 敏感度Sens和特异性Spec分别是描述识别出的阳性占所有阳性的比例，以及描述识别出的负例占所有负例的比例，计算公式分别为：\n",
-    "$$\n",
-    "\\text { Sens }=\\frac{T P}{T P+F N}\n",
-    "$$\n",
-    "\n",
-    "$$\n",
-    "\\text { Spec }=\\frac{T N}{F P+T N}\n",
-    "$$\n",
-    "\n",
-    "具体的实现方法首先是自定义metrics_类，并按照MindSpore官方文档继承nn.Metric父类，接着根据上述5个评价指标的计算公式，在类中定义5个指标的计算方法，之后通过重新实现clear方法来初始化相关参数；重新实现update方法来传入模型预测值和标签，通过上述定义的各评价指标计算方法，计算每个指标的值并存入一个列表；最后通过重新实现eval方法来讲存储各评估指标值的列表返回。上述流程对应的代码如下：\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 5,
-   "id": "92ee9cce",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "import numpy as np\n",
-    "from mindspore._checkparam import Validator as validator\n",
-    "from mindspore.nn import Metric\n",
-    "from mindspore import Tensor\n",
-    "\n",
-    "class metrics_(Metric):\n",
-    "    def __init__(self, metrics, smooth=1e-5):\n",
-    "        super(metrics_, self).__init__()\n",
-    "        self.metrics = metrics\n",
-    "        self.smooth = validator.check_positive_float(smooth, \"smooth\")\n",
-    "        self.metrics_list = [0. for i in range(len(self.metrics))]\n",
-    "        self._samples_num = 0\n",
-    "        self.clear()\n",
-    "\n",
-    "    def Acc_metrics(self,y_pred, y):\n",
-    "        tp = np.sum(y_pred.flatten() == y.flatten(), dtype=y_pred.dtype)\n",
-    "        total = len(y_pred.flatten())\n",
-    "        single_acc = float(tp) / float(total)\n",
-    "        return single_acc\n",
-    "\n",
-    "    def IoU_metrics(self,y_pred, y):\n",
-    "        intersection = np.sum(y_pred.flatten() * y.flatten())\n",
-    "        unionset = np.sum(y_pred.flatten() + y.flatten()) - intersection\n",
-    "        single_iou = float(intersection) / float(unionset + self.smooth)\n",
-    "        return single_iou\n",
-    "\n",
-    "    def Dice_metrics(self,y_pred, y):\n",
-    "        intersection = np.sum(y_pred.flatten() * y.flatten())\n",
-    "        unionset = np.sum(y_pred.flatten()) + np.sum(y.flatten())\n",
-    "        single_dice = 2*float(intersection) / float(unionset + self.smooth)\n",
-    "        return single_dice\n",
-    "\n",
-    "    def Sens_metrics(self,y_pred, y):\n",
-    "        tp = np.sum(y_pred.flatten() * y.flatten())\n",
-    "        actual_positives = np.sum(y.flatten())\n",
-    "        single_sens = float(tp) / float(actual_positives + self.smooth)\n",
-    "        return single_sens\n",
-    "\n",
-    "    def Spec_metrics(self,y_pred, y):\n",
-    "        true_neg = np.sum((1 - y.flatten()) * (1 - y_pred.flatten()))\n",
-    "        total_neg = np.sum((1 - y.flatten()))\n",
-    "        single_spec = float(true_neg) / float(total_neg + self.smooth)\n",
-    "        return single_spec\n",
-    "\n",
-    "    def clear(self):\n",
-    "        \"\"\"Clears the internal evaluation result.\"\"\"\n",
-    "        self.metrics_list = [0. for i in range(len(self.metrics))]\n",
-    "        self._samples_num = 0\n",
-    "\n",
-    "    def update(self, *inputs):\n",
-    "\n",
-    "        if len(inputs) != 2:\n",
-    "            raise ValueError(\"For 'update', it needs 2 inputs (predicted value, true value), \"\"but got {}.\".format(len(inputs)))\n",
-    "\n",
-    "        \n",
-    "        y_pred = Tensor(inputs[0]).asnumpy()  #modelarts,cpu\n",
-    "        # y_pred = np.array(Tensor(inputs[0]))  #cpu\n",
-    "        \n",
-    "        y_pred[y_pred > 0.5] = float(1)\n",
-    "        y_pred[y_pred <= 0.5] = float(0)\n",
-    "        \n",
-    "        y = Tensor(inputs[1]).asnumpy()  #modelarts,cpu\n",
-    "        # y = np.array(Tensor(inputs[1]))  #cpu\n",
-    "        \n",
-    "        self._samples_num += y.shape[0]\n",
-    "\n",
-    "        if y_pred.shape != y.shape:\n",
-    "            raise ValueError(f\"For 'update', predicted value (input[0]) and true value (input[1]) \"\n",
-    "                             f\"should have same shape, but got predicted value shape: {y_pred.shape}, \"\n",
-    "                             f\"true value shape: {y.shape}.\")\n",
-    "\n",
-    "        for i in range(y.shape[0]):\n",
-    "            if \"acc\" in self.metrics:\n",
-    "                single_acc = self.Acc_metrics(y_pred[i], y[i])\n",
-    "                self.metrics_list[0] += single_acc\n",
-    "            if \"iou\" in self.metrics:\n",
-    "                single_iou = self.IoU_metrics(y_pred[i], y[i])\n",
-    "                self.metrics_list[1] += single_iou\n",
-    "            if \"dice\" in self.metrics:\n",
-    "                single_dice = self.Dice_metrics(y_pred[i], y[i])\n",
-    "                self.metrics_list[2] += single_dice\n",
-    "            if \"sens\" in self.metrics:\n",
-    "                single_sens = self.Sens_metrics(y_pred[i], y[i])\n",
-    "                self.metrics_list[3] += single_sens\n",
-    "            if \"spec\" in self.metrics:\n",
-    "                single_spec = self.Spec_metrics(y_pred[i], y[i])\n",
-    "                self.metrics_list[4] += single_spec\n",
-    "\n",
-    "    def eval(self):\n",
-    "        if self._samples_num == 0:\n",
-    "            raise RuntimeError(\"The 'metrics' can not be calculated, because the number of samples is 0, \"\n",
-    "                               \"please check whether your inputs(predicted value, true value) are empty, or has \"\n",
-    "                               \"called update method before calling eval method.\")\n",
-    "        for i in range(len(self.metrics_list)):\n",
-    "            self.metrics_list[i] = self.metrics_list[i] / float(self._samples_num)\n",
-    "\n",
-    "        return self.metrics_list"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 6,
-   "id": "31b91f11-f00a-435c-980f-cb6be444cb98",
-   "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "丨acc: 0.6667丨丨iou: 0.5000丨丨dice: 0.6667丨丨sens: 0.6000丨丨spec: 0.7500丨\n"
-     ]
-    }
-   ],
-   "source": [
-    "x = Tensor(np.array([[[[0.2, 0.5, 0.7], [0.3, 0.1, 0.2], [0.9, 0.6, 0.8]]]]))\n",
-    "y = Tensor(np.array([[[[0, 1, 1], [1, 0, 0], [0, 1, 1]]]]))\n",
-    "metric = metrics_([\"acc\", \"iou\", \"dice\", \"sens\", \"spec\"],smooth=1e-5)\n",
-    "metric.clear()\n",
-    "metric.update(x, y)\n",
-    "res = metric.eval()\n",
-    "print( '丨acc: %.4f丨丨iou: %.4f丨丨dice: %.4f丨丨sens: %.4f丨丨spec: %.4f丨' % (res[0], res[1], res[2], res[3],res[4]), flush=True)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "497d2efd",
-   "metadata": {},
-   "source": [
-    "#### 2.5 模型训练及评估\n",
-    "\n",
-    "在模型训练时，首先是设置模型训练的epoch次数为50，再通过2.1节中自定义的create_dataset方法创建了训练集和验证集，其中训练集batch_size大小为4，验证集batch_size大小为2，图像尺寸统一调整为224x224；损失函数使用nn.BCELoss，优化器使用nn.Adam，并设置学习率为0.01。回调函数方面使用了LossMonitor和TimeMonitor来监控训练过程中每个epoch结束后，损失值Loss的变化情况以及每个epoch、每个step的运行时间，还实例化了2.5节中自定义的回调类EvalCallBack，实现计算每个epoch结束后，在2.4节中定义的5个评估指标，并保存当前最优模型。在50个epcoh结束后，模型在训练集和验证集上的评估指标如表1所示：\n",
-    "\n",
-    "模型训练部分的代码如下："
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 7,
-   "id": "c274b40c",
-   "metadata": {
-    "scrolled": true
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Looking in indexes: http://192.168.0.122:8888/repository/pypi/simple\n",
-      "Requirement already satisfied: ml_collections in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (0.1.1)\n",
-      "Requirement already satisfied: six in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from ml_collections) (1.16.0)\n",
-      "Requirement already satisfied: PyYAML in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from ml_collections) (5.3.1)\n",
-      "Requirement already satisfied: contextlib2 in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from ml_collections) (21.6.0)\n",
-      "Requirement already satisfied: absl-py in /home/ma-user/anaconda3/envs/MindSpore/lib/python3.7/site-packages (from ml_collections) (1.3.0)\n",
-      "训练集数据量： 20\n",
-      "验证集数据量： 10\n",
-      "iters_per_epoch:  5\n",
-      "total_train_steps:  500\n",
-      "Epoch [1 / 100]\n",
-      "Train loss:0.583257 丨acc: 0.559丨丨iou: 0.483丨丨dice: 0.648丨丨sens: 0.531丨丨spec: 0.648丨\n",
-      "Val loss:0.582231 丨acc: 0.219丨丨iou: 0.000丨丨dice: 0.000丨丨sens: 0.000丨丨spec: 1.000丨\n",
-      "Epoch [2 / 100]\n",
-      "Train loss:0.583249 丨acc: 0.560丨丨iou: 0.484丨丨dice: 0.649丨丨sens: 0.533丨丨spec: 0.646丨\n",
-      "Val loss:0.582301 丨acc: 0.219丨丨iou: 0.000丨丨dice: 0.000丨丨sens: 0.000丨丨spec: 1.000丨\n",
-      "Epoch [3 / 100]\n",
-      "Train loss:0.583241 丨acc: 0.561丨丨iou: 0.485丨丨dice: 0.650丨丨sens: 0.534丨丨spec: 0.645丨\n",
-      "Val loss:0.582389 丨acc: 0.219丨丨iou: 0.000丨丨dice: 0.000丨丨sens: 0.000丨丨spec: 1.000丨\n",
-      "Epoch [4 / 100]\n",
-      "Train loss:0.583233 丨acc: 0.561丨丨iou: 0.486丨丨dice: 0.652丨丨sens: 0.536丨丨spec: 0.644丨\n",
-      "Val loss:0.582496 丨acc: 0.219丨丨iou: 0.000丨丨dice: 0.000丨丨sens: 0.000丨丨spec: 1.000丨\n",
-      "IoU improved from 0.0000 to 0.0000\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [5 / 100]\n",
-      "Train loss:0.583225 丨acc: 0.562丨丨iou: 0.488丨丨dice: 0.653丨丨sens: 0.537丨丨spec: 0.642丨\n",
-      "Val loss:0.582530 丨acc: 0.219丨丨iou: 0.000丨丨dice: 0.001丨丨sens: 0.000丨丨spec: 1.000丨\n",
-      "IoU improved from 0.0000 to 0.0003\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [6 / 100]\n",
-      "Train loss:0.583218 丨acc: 0.563丨丨iou: 0.489丨丨dice: 0.654丨丨sens: 0.539丨丨spec: 0.641丨\n",
-      "Val loss:0.582752 丨acc: 0.247丨丨iou: 0.041丨丨dice: 0.079丨丨sens: 0.041丨丨spec: 0.978丨\n",
-      "IoU improved from 0.0003 to 0.0412\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [7 / 100]\n",
-      "Train loss:0.583209 丨acc: 0.564丨丨iou: 0.490丨丨dice: 0.655丨丨sens: 0.540丨丨spec: 0.639丨\n",
-      "Val loss:0.582824 丨acc: 0.305丨丨iou: 0.134丨丨dice: 0.237丨丨sens: 0.138丨丨spec: 0.901丨\n",
-      "IoU improved from 0.0412 to 0.1342\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [8 / 100]\n",
-      "Train loss:0.583201 丨acc: 0.565丨丨iou: 0.491丨丨dice: 0.656丨丨sens: 0.542丨丨spec: 0.638丨\n",
-      "Val loss:0.582034 丨acc: 0.576丨丨iou: 0.525丨丨dice: 0.689丨丨sens: 0.602丨丨spec: 0.482丨\n",
-      "IoU improved from 0.1342 to 0.5252\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [9 / 100]\n",
-      "Train loss:0.583194 丨acc: 0.566丨丨iou: 0.492丨丨dice: 0.657丨丨sens: 0.543丨丨spec: 0.637丨\n",
-      "Val loss:0.581645 丨acc: 0.635丨丨iou: 0.606丨丨dice: 0.755丨丨sens: 0.720丨丨spec: 0.335丨\n",
-      "IoU improved from 0.5252 to 0.6062\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [10 / 100]\n",
-      "Train loss:0.583186 丨acc: 0.566丨丨iou: 0.493丨丨dice: 0.658丨丨sens: 0.545丨丨spec: 0.635丨\n",
-      "Val loss:0.581545 丨acc: 0.616丨丨iou: 0.583丨丨dice: 0.736丨丨sens: 0.688丨丨spec: 0.363丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [11 / 100]\n",
-      "Train loss:0.583178 丨acc: 0.567丨丨iou: 0.495丨丨dice: 0.659丨丨sens: 0.546丨丨spec: 0.634丨\n",
-      "Val loss:0.581647 丨acc: 0.572丨丨iou: 0.525丨丨dice: 0.688丨丨sens: 0.606丨丨spec: 0.456丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [12 / 100]\n",
-      "Train loss:0.583170 丨acc: 0.568丨丨iou: 0.496丨丨dice: 0.660丨丨sens: 0.548丨丨spec: 0.633丨\n",
-      "Val loss:0.581873 丨acc: 0.527丨丨iou: 0.466丨丨dice: 0.635丨丨sens: 0.529丨丨spec: 0.525丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [13 / 100]\n",
-      "Train loss:0.583162 丨acc: 0.569丨丨iou: 0.497丨丨dice: 0.661丨丨sens: 0.549丨丨spec: 0.631丨\n",
-      "Val loss:0.581992 丨acc: 0.500丨丨iou: 0.423丨丨dice: 0.593丨丨sens: 0.470丨丨spec: 0.606丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [14 / 100]\n",
-      "Train loss:0.583154 丨acc: 0.570丨丨iou: 0.498丨丨dice: 0.662丨丨sens: 0.551丨丨spec: 0.630丨\n",
-      "Val loss:0.581849 丨acc: 0.511丨丨iou: 0.434丨丨dice: 0.604丨丨sens: 0.481丨丨spec: 0.616丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [15 / 100]\n",
-      "Train loss:0.583146 丨acc: 0.571丨丨iou: 0.499丨丨dice: 0.663丨丨sens: 0.552丨丨spec: 0.628丨\n",
-      "Val loss:0.581558 丨acc: 0.544丨丨iou: 0.475丨丨dice: 0.643丨丨sens: 0.530丨丨spec: 0.592丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [16 / 100]\n",
-      "Train loss:0.583138 丨acc: 0.571丨丨iou: 0.500丨丨dice: 0.664丨丨sens: 0.554丨丨spec: 0.627丨\n",
-      "Val loss:0.581365 丨acc: 0.563丨丨iou: 0.497丨丨dice: 0.663丨丨sens: 0.554丨丨spec: 0.595丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [17 / 100]\n",
-      "Train loss:0.583130 丨acc: 0.572丨丨iou: 0.501丨丨dice: 0.665丨丨sens: 0.555丨丨spec: 0.625丨\n",
-      "Val loss:0.581302 丨acc: 0.564丨丨iou: 0.492丨丨dice: 0.659丨丨sens: 0.543丨丨spec: 0.637丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [18 / 100]\n",
-      "Train loss:0.583122 丨acc: 0.573丨丨iou: 0.503丨丨dice: 0.666丨丨sens: 0.556丨丨spec: 0.624丨\n",
-      "Val loss:0.581352 丨acc: 0.555丨丨iou: 0.477丨丨dice: 0.645丨丨sens: 0.521丨丨spec: 0.673丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [19 / 100]\n",
-      "Train loss:0.583115 丨acc: 0.574丨丨iou: 0.504丨丨dice: 0.667丨丨sens: 0.558丨丨spec: 0.623丨\n",
-      "Val loss:0.581279 丨acc: 0.560丨丨iou: 0.482丨丨dice: 0.649丨丨sens: 0.526丨丨spec: 0.676丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [20 / 100]\n",
-      "Train loss:0.583107 丨acc: 0.575丨丨iou: 0.505丨丨dice: 0.668丨丨sens: 0.559丨丨spec: 0.621丨\n",
-      "Val loss:0.581145 丨acc: 0.570丨丨iou: 0.495丨丨dice: 0.660丨丨sens: 0.542丨丨spec: 0.659丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [21 / 100]\n",
-      "Train loss:0.583099 丨acc: 0.576丨丨iou: 0.506丨丨dice: 0.669丨丨sens: 0.561丨丨spec: 0.620丨\n",
-      "Val loss:0.581002 丨acc: 0.579丨丨iou: 0.508丨丨dice: 0.672丨丨sens: 0.560丨丨spec: 0.639丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [22 / 100]\n",
-      "Train loss:0.583091 丨acc: 0.576丨丨iou: 0.507丨丨dice: 0.670丨丨sens: 0.562丨丨spec: 0.619丨\n",
-      "Val loss:0.580899 丨acc: 0.587丨丨iou: 0.520丨丨dice: 0.683丨丨sens: 0.575丨丨spec: 0.621丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [23 / 100]\n",
-      "Train loss:0.583083 丨acc: 0.577丨丨iou: 0.508丨丨dice: 0.671丨丨sens: 0.564丨丨spec: 0.617丨\n",
-      "Val loss:0.580825 丨acc: 0.592丨丨iou: 0.528丨丨dice: 0.689丨丨sens: 0.585丨丨spec: 0.607丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [24 / 100]\n",
-      "Train loss:0.583075 丨acc: 0.578丨丨iou: 0.509丨丨dice: 0.672丨丨sens: 0.565丨丨spec: 0.616丨\n",
-      "Val loss:0.580772 丨acc: 0.596丨丨iou: 0.533丨丨dice: 0.694丨丨sens: 0.593丨丨spec: 0.595丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [25 / 100]\n",
-      "Train loss:0.583067 丨acc: 0.579丨丨iou: 0.510丨丨dice: 0.673丨丨sens: 0.567丨丨spec: 0.614丨\n",
-      "Val loss:0.580734 丨acc: 0.599丨丨iou: 0.537丨丨dice: 0.697丨丨sens: 0.599丨丨spec: 0.589丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [26 / 100]\n",
-      "Train loss:0.583059 丨acc: 0.580丨丨iou: 0.512丨丨dice: 0.674丨丨sens: 0.568丨丨spec: 0.613丨\n",
-      "Val loss:0.580708 丨acc: 0.601丨丨iou: 0.540丨丨dice: 0.700丨丨sens: 0.603丨丨spec: 0.584丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [27 / 100]\n",
-      "Train loss:0.583051 丨acc: 0.580丨丨iou: 0.513丨丨dice: 0.675丨丨sens: 0.569丨丨spec: 0.612丨\n",
-      "Val loss:0.580688 丨acc: 0.602丨丨iou: 0.542丨丨dice: 0.702丨丨sens: 0.606丨丨spec: 0.580丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [28 / 100]\n",
-      "Train loss:0.583043 丨acc: 0.581丨丨iou: 0.514丨丨dice: 0.676丨丨sens: 0.571丨丨spec: 0.610丨\n",
-      "Val loss:0.580673 丨acc: 0.604丨丨iou: 0.544丨丨dice: 0.703丨丨sens: 0.608丨丨spec: 0.578丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [29 / 100]\n",
-      "Train loss:0.583036 丨acc: 0.582丨丨iou: 0.515丨丨dice: 0.677丨丨sens: 0.572丨丨spec: 0.609丨\n",
-      "Val loss:0.580660 丨acc: 0.605丨丨iou: 0.546丨丨dice: 0.704丨丨sens: 0.610丨丨spec: 0.576丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [30 / 100]\n",
-      "Train loss:0.583028 丨acc: 0.583丨丨iou: 0.516丨丨dice: 0.678丨丨sens: 0.574丨丨spec: 0.607丨\n",
-      "Val loss:0.580648 丨acc: 0.606丨丨iou: 0.547丨丨dice: 0.706丨丨sens: 0.612丨丨spec: 0.574丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [31 / 100]\n",
-      "Train loss:0.583020 丨acc: 0.584丨丨iou: 0.517丨丨dice: 0.679丨丨sens: 0.575丨丨spec: 0.606丨\n",
-      "Val loss:0.580639 丨acc: 0.607丨丨iou: 0.548丨丨dice: 0.707丨丨sens: 0.613丨丨spec: 0.572丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [32 / 100]\n",
-      "Train loss:0.583011 丨acc: 0.585丨丨iou: 0.518丨丨dice: 0.680丨丨sens: 0.577丨丨spec: 0.605丨\n",
-      "Val loss:0.580630 丨acc: 0.608丨丨iou: 0.549丨丨dice: 0.708丨丨sens: 0.615丨丨spec: 0.571丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [33 / 100]\n",
-      "Train loss:0.583003 丨acc: 0.585丨丨iou: 0.519丨丨dice: 0.681丨丨sens: 0.578丨丨spec: 0.603丨\n",
-      "Val loss:0.580620 丨acc: 0.608丨丨iou: 0.550丨丨dice: 0.708丨丨sens: 0.616丨丨spec: 0.569丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [34 / 100]\n",
-      "Train loss:0.582995 丨acc: 0.586丨丨iou: 0.521丨丨dice: 0.682丨丨sens: 0.580丨丨spec: 0.602丨\n",
-      "Val loss:0.580611 丨acc: 0.609丨丨iou: 0.552丨丨dice: 0.709丨丨sens: 0.618丨丨spec: 0.568丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [35 / 100]\n",
-      "Train loss:0.582987 丨acc: 0.587丨丨iou: 0.522丨丨dice: 0.683丨丨sens: 0.581丨丨spec: 0.601丨\n",
-      "Val loss:0.580601 丨acc: 0.610丨丨iou: 0.553丨丨dice: 0.710丨丨sens: 0.619丨丨spec: 0.566丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [36 / 100]\n",
-      "Train loss:0.582979 丨acc: 0.588丨丨iou: 0.523丨丨dice: 0.684丨丨sens: 0.582丨丨spec: 0.599丨\n",
-      "Val loss:0.580592 丨acc: 0.611丨丨iou: 0.554丨丨dice: 0.711丨丨sens: 0.621丨丨spec: 0.565丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [37 / 100]\n",
-      "Train loss:0.582970 丨acc: 0.589丨丨iou: 0.524丨丨dice: 0.685丨丨sens: 0.584丨丨spec: 0.598丨\n",
-      "Val loss:0.580584 丨acc: 0.612丨丨iou: 0.555丨丨dice: 0.712丨丨sens: 0.622丨丨spec: 0.564丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [38 / 100]\n",
-      "Train loss:0.582961 丨acc: 0.590丨丨iou: 0.525丨丨dice: 0.686丨丨sens: 0.585丨丨spec: 0.597丨\n",
-      "Val loss:0.580575 丨acc: 0.612丨丨iou: 0.556丨丨dice: 0.713丨丨sens: 0.623丨丨spec: 0.562丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [39 / 100]\n",
-      "Train loss:0.582953 丨acc: 0.590丨丨iou: 0.526丨丨dice: 0.687丨丨sens: 0.587丨丨spec: 0.595丨\n",
-      "Val loss:0.580566 丨acc: 0.613丨丨iou: 0.557丨丨dice: 0.714丨丨sens: 0.625丨丨spec: 0.561丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [40 / 100]\n",
-      "Train loss:0.582945 丨acc: 0.591丨丨iou: 0.528丨丨dice: 0.688丨丨sens: 0.588丨丨spec: 0.594丨\n",
-      "Val loss:0.580557 丨acc: 0.614丨丨iou: 0.558丨丨dice: 0.715丨丨sens: 0.626丨丨spec: 0.560丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [41 / 100]\n",
-      "Train loss:0.582936 丨acc: 0.592丨丨iou: 0.529丨丨dice: 0.689丨丨sens: 0.590丨丨spec: 0.592丨\n",
-      "Val loss:0.580547 丨acc: 0.615丨丨iou: 0.559丨丨dice: 0.716丨丨sens: 0.628丨丨spec: 0.558丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [42 / 100]\n",
-      "Train loss:0.582928 丨acc: 0.593丨丨iou: 0.530丨丨dice: 0.690丨丨sens: 0.591丨丨spec: 0.591丨\n",
-      "Val loss:0.580538 丨acc: 0.616丨丨iou: 0.560丨丨dice: 0.717丨丨sens: 0.629丨丨spec: 0.557丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [43 / 100]\n",
-      "Train loss:0.582919 丨acc: 0.594丨丨iou: 0.531丨丨dice: 0.691丨丨sens: 0.593丨丨spec: 0.590丨\n",
-      "Val loss:0.580528 丨acc: 0.617丨丨iou: 0.561丨丨dice: 0.718丨丨sens: 0.631丨丨spec: 0.555丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [44 / 100]\n",
-      "Train loss:0.582910 丨acc: 0.595丨丨iou: 0.532丨丨dice: 0.692丨丨sens: 0.594丨丨spec: 0.588丨\n",
-      "Val loss:0.580519 丨acc: 0.617丨丨iou: 0.562丨丨dice: 0.718丨丨sens: 0.632丨丨spec: 0.554丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [45 / 100]\n",
-      "Train loss:0.582902 丨acc: 0.595丨丨iou: 0.533丨丨dice: 0.693丨丨sens: 0.596丨丨spec: 0.587丨\n",
-      "Val loss:0.580510 丨acc: 0.618丨丨iou: 0.563丨丨dice: 0.719丨丨sens: 0.634丨丨spec: 0.553丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [46 / 100]\n",
-      "Train loss:0.582893 丨acc: 0.596丨丨iou: 0.534丨丨dice: 0.694丨丨sens: 0.597丨丨spec: 0.585丨\n",
-      "Val loss:0.580500 丨acc: 0.619丨丨iou: 0.564丨丨dice: 0.720丨丨sens: 0.635丨丨spec: 0.551丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [47 / 100]\n",
-      "Train loss:0.582885 丨acc: 0.597丨丨iou: 0.535丨丨dice: 0.695丨丨sens: 0.599丨丨spec: 0.584丨\n",
-      "Val loss:0.580492 丨acc: 0.620丨丨iou: 0.565丨丨dice: 0.721丨丨sens: 0.636丨丨spec: 0.550丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [48 / 100]\n",
-      "Train loss:0.582876 丨acc: 0.598丨丨iou: 0.536丨丨dice: 0.696丨丨sens: 0.600丨丨spec: 0.583丨\n",
-      "Val loss:0.580482 丨acc: 0.621丨丨iou: 0.567丨丨dice: 0.722丨丨sens: 0.638丨丨spec: 0.549丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [49 / 100]\n",
-      "Train loss:0.582867 丨acc: 0.599丨丨iou: 0.537丨丨dice: 0.697丨丨sens: 0.601丨丨spec: 0.581丨\n",
-      "Val loss:0.580473 丨acc: 0.621丨丨iou: 0.568丨丨dice: 0.723丨丨sens: 0.639丨丨spec: 0.547丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [50 / 100]\n",
-      "Train loss:0.582859 丨acc: 0.599丨丨iou: 0.538丨丨dice: 0.698丨丨sens: 0.603丨丨spec: 0.580丨\n",
-      "Val loss:0.580463 丨acc: 0.622丨丨iou: 0.569丨丨dice: 0.724丨丨sens: 0.641丨丨spec: 0.546丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [51 / 100]\n",
-      "Train loss:0.582850 丨acc: 0.600丨丨iou: 0.540丨丨dice: 0.699丨丨sens: 0.604丨丨spec: 0.578丨\n",
-      "Val loss:0.580454 丨acc: 0.623丨丨iou: 0.570丨丨dice: 0.725丨丨sens: 0.642丨丨spec: 0.544丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [52 / 100]\n",
-      "Train loss:0.582841 丨acc: 0.601丨丨iou: 0.541丨丨dice: 0.700丨丨sens: 0.606丨丨spec: 0.577丨\n",
-      "Val loss:0.580445 丨acc: 0.624丨丨iou: 0.571丨丨dice: 0.725丨丨sens: 0.643丨丨spec: 0.543丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [53 / 100]\n",
-      "Train loss:0.582833 丨acc: 0.602丨丨iou: 0.542丨丨dice: 0.700丨丨sens: 0.607丨丨spec: 0.576丨\n",
-      "Val loss:0.580436 丨acc: 0.625丨丨iou: 0.572丨丨dice: 0.726丨丨sens: 0.645丨丨spec: 0.542丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [54 / 100]\n",
-      "Train loss:0.582824 丨acc: 0.603丨丨iou: 0.543丨丨dice: 0.701丨丨sens: 0.609丨丨spec: 0.574丨\n",
-      "Val loss:0.580428 丨acc: 0.625丨丨iou: 0.573丨丨dice: 0.727丨丨sens: 0.646丨丨spec: 0.541丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [55 / 100]\n",
-      "Train loss:0.582814 丨acc: 0.604丨丨iou: 0.544丨丨dice: 0.702丨丨sens: 0.610丨丨spec: 0.573丨\n",
-      "Val loss:0.580418 丨acc: 0.626丨丨iou: 0.574丨丨dice: 0.728丨丨sens: 0.647丨丨spec: 0.540丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [56 / 100]\n",
-      "Train loss:0.582805 丨acc: 0.605丨丨iou: 0.545丨丨dice: 0.703丨丨sens: 0.612丨丨spec: 0.572丨\n",
-      "Val loss:0.580410 丨acc: 0.627丨丨iou: 0.575丨丨dice: 0.729丨丨sens: 0.648丨丨spec: 0.539丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [57 / 100]\n",
-      "Train loss:0.582795 丨acc: 0.606丨丨iou: 0.546丨丨dice: 0.704丨丨sens: 0.613丨丨spec: 0.571丨\n",
-      "Val loss:0.580400 丨acc: 0.628丨丨iou: 0.576丨丨dice: 0.729丨丨sens: 0.650丨丨spec: 0.538丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [58 / 100]\n",
-      "Train loss:0.582785 丨acc: 0.606丨丨iou: 0.548丨丨dice: 0.705丨丨sens: 0.614丨丨spec: 0.569丨\n",
-      "Val loss:0.580390 丨acc: 0.628丨丨iou: 0.577丨丨dice: 0.730丨丨sens: 0.651丨丨spec: 0.537丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [59 / 100]\n",
-      "Train loss:0.582775 丨acc: 0.607丨丨iou: 0.549丨丨dice: 0.706丨丨sens: 0.616丨丨spec: 0.568丨\n",
-      "Val loss:0.580380 丨acc: 0.629丨丨iou: 0.578丨丨dice: 0.731丨丨sens: 0.653丨丨spec: 0.535丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [60 / 100]\n",
-      "Train loss:0.582764 丨acc: 0.608丨丨iou: 0.550丨丨dice: 0.707丨丨sens: 0.618丨丨spec: 0.567丨\n",
-      "Val loss:0.580370 丨acc: 0.630丨丨iou: 0.579丨丨dice: 0.732丨丨sens: 0.654丨丨spec: 0.535丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [61 / 100]\n",
-      "Train loss:0.582752 丨acc: 0.609丨丨iou: 0.551丨丨dice: 0.708丨丨sens: 0.619丨丨spec: 0.566丨\n",
-      "Val loss:0.580357 丨acc: 0.631丨丨iou: 0.580丨丨dice: 0.733丨丨sens: 0.655丨丨spec: 0.534丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [62 / 100]\n",
-      "Train loss:0.582739 丨acc: 0.610丨丨iou: 0.552丨丨dice: 0.709丨丨sens: 0.621丨丨spec: 0.565丨\n",
-      "Val loss:0.580346 丨acc: 0.632丨丨iou: 0.581丨丨dice: 0.734丨丨sens: 0.656丨丨spec: 0.533丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [63 / 100]\n",
-      "Train loss:0.582727 丨acc: 0.611丨丨iou: 0.554丨丨dice: 0.711丨丨sens: 0.622丨丨spec: 0.565丨\n",
-      "Val loss:0.580333 丨acc: 0.633丨丨iou: 0.582丨丨dice: 0.735丨丨sens: 0.658丨丨spec: 0.532丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [64 / 100]\n",
-      "Train loss:0.582713 丨acc: 0.612丨丨iou: 0.555丨丨dice: 0.712丨丨sens: 0.624丨丨spec: 0.564丨\n",
-      "Val loss:0.580318 丨acc: 0.634丨丨iou: 0.584丨丨dice: 0.736丨丨sens: 0.660丨丨spec: 0.531丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [65 / 100]\n",
-      "Train loss:0.582699 丨acc: 0.614丨丨iou: 0.556丨丨dice: 0.713丨丨sens: 0.625丨丨spec: 0.564丨\n",
-      "Val loss:0.580304 丨acc: 0.635丨丨iou: 0.585丨丨dice: 0.737丨丨sens: 0.661丨丨spec: 0.531丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [66 / 100]\n",
-      "Train loss:0.582685 丨acc: 0.615丨丨iou: 0.558丨丨dice: 0.714丨丨sens: 0.627丨丨spec: 0.563丨\n",
-      "Val loss:0.580290 丨acc: 0.636丨丨iou: 0.586丨丨dice: 0.738丨丨sens: 0.663丨丨spec: 0.530丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [67 / 100]\n",
-      "Train loss:0.582671 丨acc: 0.616丨丨iou: 0.559丨丨dice: 0.715丨丨sens: 0.629丨丨spec: 0.563丨\n",
-      "Val loss:0.580274 丨acc: 0.637丨丨iou: 0.588丨丨dice: 0.739丨丨sens: 0.664丨丨spec: 0.529丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [68 / 100]\n",
-      "Train loss:0.582656 丨acc: 0.617丨丨iou: 0.560丨丨dice: 0.716丨丨sens: 0.630丨丨spec: 0.562丨\n",
-      "Val loss:0.580257 丨acc: 0.639丨丨iou: 0.589丨丨dice: 0.740丨丨sens: 0.667丨丨spec: 0.529丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [69 / 100]\n",
-      "Train loss:0.582641 丨acc: 0.618丨丨iou: 0.562丨丨dice: 0.717丨丨sens: 0.632丨丨spec: 0.562丨\n",
-      "Val loss:0.580239 丨acc: 0.640丨丨iou: 0.591丨丨dice: 0.741丨丨sens: 0.668丨丨spec: 0.527丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [70 / 100]\n",
-      "Train loss:0.582624 丨acc: 0.619丨丨iou: 0.563丨丨dice: 0.719丨丨sens: 0.633丨丨spec: 0.562丨\n",
-      "Val loss:0.580215 丨acc: 0.641丨丨iou: 0.593丨丨dice: 0.743丨丨sens: 0.671丨丨spec: 0.526丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [71 / 100]\n",
-      "Train loss:0.582607 丨acc: 0.620丨丨iou: 0.565丨丨dice: 0.720丨丨sens: 0.635丨丨spec: 0.561丨\n",
-      "Val loss:0.580194 丨acc: 0.643丨丨iou: 0.594丨丨dice: 0.744丨丨sens: 0.673丨丨spec: 0.525丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [72 / 100]\n",
-      "Train loss:0.582589 丨acc: 0.622丨丨iou: 0.566丨丨dice: 0.721丨丨sens: 0.637丨丨spec: 0.561丨\n",
-      "Val loss:0.580173 丨acc: 0.644丨丨iou: 0.596丨丨dice: 0.746丨丨sens: 0.675丨丨spec: 0.524丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [73 / 100]\n",
-      "Train loss:0.582570 丨acc: 0.623丨丨iou: 0.568丨丨dice: 0.722丨丨sens: 0.638丨丨spec: 0.561丨\n",
-      "Val loss:0.580150 丨acc: 0.646丨丨iou: 0.598丨丨dice: 0.747丨丨sens: 0.677丨丨spec: 0.524丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [74 / 100]\n",
-      "Train loss:0.582550 丨acc: 0.625丨丨iou: 0.569丨丨dice: 0.723丨丨sens: 0.640丨丨spec: 0.562丨\n",
-      "Val loss:0.580126 丨acc: 0.648丨丨iou: 0.600丨丨dice: 0.749丨丨sens: 0.680丨丨spec: 0.523丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [75 / 100]\n",
-      "Train loss:0.582529 丨acc: 0.626丨丨iou: 0.571丨丨dice: 0.725丨丨sens: 0.642丨丨spec: 0.562丨\n",
-      "Val loss:0.580104 丨acc: 0.649丨丨iou: 0.601丨丨dice: 0.750丨丨sens: 0.681丨丨spec: 0.523丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [76 / 100]\n",
-      "Train loss:0.582508 丨acc: 0.627丨丨iou: 0.572丨丨dice: 0.726丨丨sens: 0.643丨丨spec: 0.563丨\n",
-      "Val loss:0.580080 丨acc: 0.650丨丨iou: 0.603丨丨dice: 0.751丨丨sens: 0.683丨丨spec: 0.523丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [77 / 100]\n",
-      "Train loss:0.582487 丨acc: 0.629丨丨iou: 0.574丨丨dice: 0.727丨丨sens: 0.645丨丨spec: 0.563丨\n",
-      "Val loss:0.580053 丨acc: 0.652丨丨iou: 0.605丨丨dice: 0.753丨丨sens: 0.686丨丨spec: 0.523丨\n",
-      "IoU did not improve from 0.6062 \n",
-      "-------------------------------\n",
-      "Epoch [78 / 100]\n",
-      "Train loss:0.582465 丨acc: 0.630丨丨iou: 0.576丨丨dice: 0.729丨丨sens: 0.647丨丨spec: 0.564丨\n",
-      "Val loss:0.580026 丨acc: 0.654丨丨iou: 0.607丨丨dice: 0.755丨丨sens: 0.688丨丨spec: 0.521丨\n",
-      "IoU improved from 0.6062 to 0.6074\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [79 / 100]\n",
-      "Train loss:0.582443 丨acc: 0.632丨丨iou: 0.577丨丨dice: 0.730丨丨sens: 0.648丨丨spec: 0.565丨\n",
-      "Val loss:0.579994 丨acc: 0.656丨丨iou: 0.609丨丨dice: 0.756丨丨sens: 0.691丨丨spec: 0.521丨\n",
-      "IoU improved from 0.6074 to 0.6094\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [80 / 100]\n",
-      "Train loss:0.582420 丨acc: 0.633丨丨iou: 0.579丨丨dice: 0.731丨丨sens: 0.650丨丨spec: 0.565丨\n",
-      "Val loss:0.579969 丨acc: 0.658丨丨iou: 0.612丨丨dice: 0.758丨丨sens: 0.693丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6094 to 0.6117\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [81 / 100]\n",
-      "Train loss:0.582397 丨acc: 0.635丨丨iou: 0.580丨丨dice: 0.732丨丨sens: 0.652丨丨spec: 0.566丨\n",
-      "Val loss:0.579947 丨acc: 0.659丨丨iou: 0.613丨丨dice: 0.759丨丨sens: 0.695丨丨spec: 0.521丨\n",
-      "IoU improved from 0.6117 to 0.6132\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [82 / 100]\n",
-      "Train loss:0.582374 丨acc: 0.636丨丨iou: 0.582丨丨dice: 0.734丨丨sens: 0.653丨丨spec: 0.567丨\n",
-      "Val loss:0.579921 丨acc: 0.661丨丨iou: 0.615丨丨dice: 0.761丨丨sens: 0.697丨丨spec: 0.521丨\n",
-      "IoU improved from 0.6132 to 0.6149\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [83 / 100]\n",
-      "Train loss:0.582351 丨acc: 0.638丨丨iou: 0.583丨丨dice: 0.735丨丨sens: 0.655丨丨spec: 0.568丨\n",
-      "Val loss:0.579899 丨acc: 0.662丨丨iou: 0.616丨丨dice: 0.762丨丨sens: 0.698丨丨spec: 0.521丨\n",
-      "IoU improved from 0.6149 to 0.6163\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [84 / 100]\n",
-      "Train loss:0.582328 丨acc: 0.639丨丨iou: 0.585丨丨dice: 0.736丨丨sens: 0.657丨丨spec: 0.568丨\n",
-      "Val loss:0.579872 丨acc: 0.664丨丨iou: 0.618丨丨dice: 0.763丨丨sens: 0.700丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6163 to 0.6181\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [85 / 100]\n",
-      "Train loss:0.582305 丨acc: 0.640丨丨iou: 0.586丨丨dice: 0.737丨丨sens: 0.658丨丨spec: 0.569丨\n",
-      "Val loss:0.579849 丨acc: 0.665丨丨iou: 0.620丨丨dice: 0.764丨丨sens: 0.702丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6181 to 0.6196\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [86 / 100]\n",
-      "Train loss:0.582282 丨acc: 0.642丨丨iou: 0.588丨丨dice: 0.739丨丨sens: 0.660丨丨spec: 0.569丨\n",
-      "Val loss:0.579825 丨acc: 0.666丨丨iou: 0.621丨丨dice: 0.765丨丨sens: 0.704丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6196 to 0.6211\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [87 / 100]\n",
-      "Train loss:0.582259 丨acc: 0.643丨丨iou: 0.590丨丨dice: 0.740丨丨sens: 0.661丨丨spec: 0.570丨\n",
-      "Val loss:0.579797 丨acc: 0.668丨丨iou: 0.623丨丨dice: 0.767丨丨sens: 0.706丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6211 to 0.6232\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [88 / 100]\n",
-      "Train loss:0.582235 丨acc: 0.644丨丨iou: 0.591丨丨dice: 0.741丨丨sens: 0.663丨丨spec: 0.570丨\n",
-      "Val loss:0.579773 丨acc: 0.669丨丨iou: 0.625丨丨dice: 0.768丨丨sens: 0.708丨丨spec: 0.522丨\n",
-      "IoU improved from 0.6232 to 0.6246\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [89 / 100]\n",
-      "Train loss:0.582212 丨acc: 0.646丨丨iou: 0.592丨丨dice: 0.742丨丨sens: 0.665丨丨spec: 0.571丨\n",
-      "Val loss:0.579754 丨acc: 0.670丨丨iou: 0.626丨丨dice: 0.769丨丨sens: 0.709丨丨spec: 0.523丨\n",
-      "IoU improved from 0.6246 to 0.6257\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [90 / 100]\n",
-      "Train loss:0.582190 丨acc: 0.647丨丨iou: 0.594丨丨dice: 0.743丨丨sens: 0.666丨丨spec: 0.572丨\n",
-      "Val loss:0.579728 丨acc: 0.672丨丨iou: 0.627丨丨dice: 0.770丨丨sens: 0.710丨丨spec: 0.523丨\n",
-      "IoU improved from 0.6257 to 0.6273\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [91 / 100]\n",
-      "Train loss:0.582166 丨acc: 0.648丨丨iou: 0.595丨丨dice: 0.744丨丨sens: 0.668丨丨spec: 0.572丨\n",
-      "Val loss:0.579705 丨acc: 0.673丨丨iou: 0.628丨丨dice: 0.771丨丨sens: 0.712丨丨spec: 0.523丨\n",
-      "IoU improved from 0.6273 to 0.6285\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [92 / 100]\n",
-      "Train loss:0.582144 丨acc: 0.650丨丨iou: 0.597丨丨dice: 0.746丨丨sens: 0.669丨丨spec: 0.573丨\n",
-      "Val loss:0.579679 丨acc: 0.674丨丨iou: 0.630丨丨dice: 0.772丨丨sens: 0.713丨丨spec: 0.523丨\n",
-      "IoU improved from 0.6285 to 0.6300\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [93 / 100]\n",
-      "Train loss:0.582122 丨acc: 0.651丨丨iou: 0.598丨丨dice: 0.747丨丨sens: 0.671丨丨spec: 0.573丨\n",
-      "Val loss:0.579659 丨acc: 0.675丨丨iou: 0.631丨丨dice: 0.773丨丨sens: 0.715丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6300 to 0.6313\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [94 / 100]\n",
-      "Train loss:0.582101 丨acc: 0.652丨丨iou: 0.600丨丨dice: 0.748丨丨sens: 0.672丨丨spec: 0.574丨\n",
-      "Val loss:0.579631 丨acc: 0.677丨丨iou: 0.633丨丨dice: 0.774丨丨sens: 0.717丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6313 to 0.6329\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [95 / 100]\n",
-      "Train loss:0.582080 丨acc: 0.653丨丨iou: 0.601丨丨dice: 0.749丨丨sens: 0.674丨丨spec: 0.574丨\n",
-      "Val loss:0.579606 丨acc: 0.678丨丨iou: 0.634丨丨dice: 0.775丨丨sens: 0.718丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6329 to 0.6344\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [96 / 100]\n",
-      "Train loss:0.582059 丨acc: 0.655丨丨iou: 0.602丨丨dice: 0.750丨丨sens: 0.675丨丨spec: 0.574丨\n",
-      "Val loss:0.579588 丨acc: 0.679丨丨iou: 0.635丨丨dice: 0.776丨丨sens: 0.719丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6344 to 0.6353\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [97 / 100]\n",
-      "Train loss:0.582039 丨acc: 0.656丨丨iou: 0.604丨丨dice: 0.751丨丨sens: 0.676丨丨spec: 0.574丨\n",
-      "Val loss:0.579570 丨acc: 0.680丨丨iou: 0.636丨丨dice: 0.777丨丨sens: 0.720丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6353 to 0.6363\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [98 / 100]\n",
-      "Train loss:0.582019 丨acc: 0.657丨丨iou: 0.605丨丨dice: 0.752丨丨sens: 0.678丨丨spec: 0.575丨\n",
-      "Val loss:0.579550 丨acc: 0.681丨丨iou: 0.638丨丨dice: 0.778丨丨sens: 0.722丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6363 to 0.6375\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [99 / 100]\n",
-      "Train loss:0.582000 丨acc: 0.658丨丨iou: 0.606丨丨dice: 0.753丨丨sens: 0.679丨丨spec: 0.575丨\n",
-      "Val loss:0.579531 丨acc: 0.682丨丨iou: 0.639丨丨dice: 0.779丨丨sens: 0.723丨丨spec: 0.524丨\n",
-      "IoU improved from 0.6375 to 0.6389\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Epoch [100 / 100]\n",
-      "Train loss:0.581981 丨acc: 0.659丨丨iou: 0.607丨丨dice: 0.754丨丨sens: 0.681丨丨spec: 0.575丨\n",
-      "Val loss:0.579514 丨acc: 0.683丨丨iou: 0.640丨丨dice: 0.779丨丨sens: 0.724丨丨spec: 0.525丨\n",
-      "IoU improved from 0.6389 to 0.6397\n",
-      "saving best checkpoint at: checkpoint/best_UNet.ckpt \n",
-      "Done!\n"
-     ]
-    }
-   ],
-   "source": [
-    "import mindspore.nn as nn\n",
-    "from mindspore import ops\n",
-    "import mindspore\n",
-    "from mindspore import ms_function\n",
-    "os.system(\"pip install ml_collections\" )\n",
-    "import ml_collections\n",
-    "\n",
-    "def get_config():\n",
-    "    \"\"\"configuration \"\"\"\n",
-    "    config = ml_collections.ConfigDict()\n",
-    "    config.epochs = 100\n",
-    "    config.train_data_path = \"src/datasets/ISBI/train/\"\n",
-    "    config.val_data_path = \"src/datasets/ISBI/val/\"\n",
-    "    config.imgsize = 224\n",
-    "    config.batch_size = 4\n",
-    "    config.pretrained_path = None\n",
-    "    config.in_channel = 3\n",
-    "    config.n_classes = 1\n",
-    "    config.lr = 0.0001\n",
-    "    return config\n",
-    "\n",
-    "cfg = get_config()\n",
-    "\n",
-    "\n",
-    "train_dataset = create_dataset(cfg.train_data_path, img_size=cfg.imgsize, batch_size= cfg.batch_size, augment=True, shuffle = True)\n",
-    "val_dataset = create_dataset(cfg.val_data_path, img_size=cfg.imgsize, batch_size= cfg.batch_size, augment=False, shuffle = False)\n",
-    "\n",
-    "\n",
-    "def train(model, dataset, loss_fn, optimizer, met):\n",
-    "    # Define forward function\n",
-    "    def forward_fn(data, label):\n",
-    "        logits = model(data)\n",
-    "        loss = loss_fn(logits, label)\n",
-    "        return loss, logits\n",
-    "    # Get gradient function\n",
-    "    grad_fn = ops.value_and_grad(forward_fn, None, optimizer.parameters, has_aux=True)\n",
-    "    # Define function of one-step training\n",
-    "    @ms_function\n",
-    "    def train_step(data, label):\n",
-    "        (loss, logits), grads = grad_fn(data, label)\n",
-    "        loss = ops.depend(loss, optimizer(grads))\n",
-    "        return loss, logits\n",
-    "\n",
-    "    size = dataset.get_dataset_size()\n",
-    "    model.set_train(True)\n",
-    "    train_loss = 0\n",
-    "    train_pred = []\n",
-    "    train_label = []\n",
-    "    for batch, (data, label) in enumerate(dataset.create_tuple_iterator()):\n",
-    "        loss, logits = train_step(data, label)\n",
-    "        train_loss += loss.asnumpy()\n",
-    "        train_pred.extend(logits.asnumpy())\n",
-    "        train_label.extend(label.asnumpy())\n",
-    "\n",
-    "    train_loss /= size\n",
-    "    metric = metrics_(met, smooth=1e-5)\n",
-    "    metric.clear()\n",
-    "    metric.update(train_pred, train_label)\n",
-    "    res = metric.eval()\n",
-    "    print(f'Train loss:{train_loss:>4f}','丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % (res[0], res[1], res[2], res[3], res[4]))\n",
-    "\n",
-    "\n",
-    "def val(model, dataset, loss_fn, met):\n",
-    "    size = dataset.get_dataset_size()\n",
-    "    model.set_train(False)\n",
-    "    val_loss = 0\n",
-    "    val_pred = []\n",
-    "    val_label = []\n",
-    "    for batch, (data, label) in enumerate(dataset.create_tuple_iterator()):\n",
-    "        pred = model(data)\n",
-    "        val_loss += loss_fn(pred, label).asnumpy()\n",
-    "        val_pred.extend(pred.asnumpy())\n",
-    "        val_label.extend(label.asnumpy())\n",
-    "\n",
-    "    val_loss /= size\n",
-    "    metric = metrics_(met, smooth=1e-5)\n",
-    "    metric.clear()\n",
-    "    metric.update(val_pred, val_label)\n",
-    "    res = metric.eval()\n",
-    "\n",
-    "    print(f'Val loss:{val_loss:>4f}','丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % (res[0], res[1], res[2], res[3], res[4]))\n",
-    "\n",
-    "    checkpoint = res[1]\n",
-    "    return checkpoint, res[4]\n",
-    "\n",
-    "\n",
-    "net = UNet(cfg.in_channel, cfg.n_classes)\n",
-    "\n",
-    "criterion = nn.BCEWithLogitsLoss()\n",
-    "optimizer = nn.SGD(params=net.trainable_params(), learning_rate=cfg.lr)\n",
-    "\n",
-    "iters_per_epoch = train_dataset.get_dataset_size()\n",
-    "total_train_steps = iters_per_epoch * cfg.epochs\n",
-    "print('iters_per_epoch: ', iters_per_epoch)\n",
-    "print('total_train_steps: ', total_train_steps)\n",
-    "\n",
-    "metrics_name = [\"acc\", \"iou\", \"dice\", \"sens\", \"spec\"]\n",
-    "\n",
-    "best_iou = 0\n",
-    "ckpt_path = 'checkpoint/best_UNet.ckpt'\n",
-    "for epoch in range(cfg.epochs):\n",
-    "    print(f\"Epoch [{epoch+1} / {cfg.epochs}]\")\n",
-    "    train(net, train_dataset, criterion, optimizer, metrics_name)\n",
-    "    checkpoint_best, spec = val(net, val_dataset, criterion, metrics_name)\n",
-    "    if epoch > 2 and spec > 0.2:\n",
-    "        if checkpoint_best > best_iou:\n",
-    "            print('IoU improved from %0.4f to %0.4f' % (best_iou, checkpoint_best))\n",
-    "            best_iou = checkpoint_best\n",
-    "            mindspore.save_checkpoint(net, ckpt_path)\n",
-    "            print(\"saving best checkpoint at: {} \".format(ckpt_path))\n",
-    "        else:\n",
-    "            print('IoU did not improve from %0.4f' % (best_iou),\"\\n-------------------------------\")\n",
-    "print(\"Done!\")\n"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "12d545bd",
-   "metadata": {},
-   "source": [
-    "#### 2.6 模型预测\n",
-    "代码如下："
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 8,
-   "id": "324016be",
-   "metadata": {
-    "scrolled": true
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "30\n",
-      "Evaluation metrics cannot be calculated without Mask\n"
-     ]
-    }
-   ],
-   "source": [
-    "import os\n",
-    "import cv2\n",
-    "import mindspore.dataset as ds\n",
-    "import glob\n",
-    "import mindspore.dataset.vision as vision_C\n",
-    "import mindspore.dataset.transforms as C_transforms\n",
-    "import random\n",
-    "import mindspore\n",
-    "from mindspore.dataset.vision import Inter\n",
-    "import numpy as np\n",
-    "from tqdm import tqdm\n",
-    "# import skimage.io as io\n",
-    "\n",
-    "\n",
-    "def val_transforms(img_size):\n",
-    "    return C_transforms.Compose([\n",
-    "    vision_C.Resize(img_size, interpolation=Inter.NEAREST),\n",
-    "    vision_C.Rescale(1/255., 0),\n",
-    "    vision_C.HWC2CHW()\n",
-    "    ])\n",
-    "\n",
-    "class Data_Loader:\n",
-    "    def __init__(self, data_path, have_mask):\n",
-    "        # 初始化函数，读取所有data_path下的图片\n",
-    "        self.data_path = data_path\n",
-    "        self.have_mask = have_mask\n",
-    "        self.imgs_path = glob.glob(os.path.join(data_path, 'image/*.png'))\n",
-    "        if self.have_mask:\n",
-    "            self.label_path = glob.glob(os.path.join(data_path, 'mask/*.png'))\n",
-    "\n",
-    "    def __getitem__(self, index):\n",
-    "        # 根据index读取图片\n",
-    "        image = cv2.imread(self.imgs_path[index])\n",
-    "        if self.have_mask:\n",
-    "            label = cv2.imread(self.label_path[index], cv2.IMREAD_GRAYSCALE)\n",
-    "            label = label.reshape((label.shape[0], label.shape[1], 1))\n",
-    "        else:\n",
-    "            label = image\n",
-    "        return image, label\n",
-    "\n",
-    "    @property\n",
-    "    def column_names(self):\n",
-    "        column_names = ['image', 'label']\n",
-    "        return column_names\n",
-    "\n",
-    "    def __len__(self):\n",
-    "        return len(self.imgs_path)\n",
-    "\n",
-    "\n",
-    "def create_dataset(data_dir, img_size, batch_size, shuffle, have_mask = False):\n",
-    "    mc_dataset = Data_Loader(data_path=data_dir, have_mask = have_mask)\n",
-    "    print(len(mc_dataset))\n",
-    "    dataset = ds.GeneratorDataset(mc_dataset, mc_dataset.column_names, shuffle=shuffle)\n",
-    "    transform_img = val_transforms(img_size)\n",
-    "    seed = random.randint(1, 1000)\n",
-    "    mindspore.set_seed(seed)\n",
-    "    dataset = dataset.map(input_columns='image', num_parallel_workers=1, operations=transform_img)\n",
-    "    mindspore.set_seed(seed)\n",
-    "    dataset = dataset.map(input_columns=\"label\", num_parallel_workers=1, operations=transform_img)\n",
-    "    dataset = dataset.batch(batch_size, num_parallel_workers=1)\n",
-    "    return dataset\n",
-    "\n",
-    "def model_pred(model, test_loader, result_path, have_mask):\n",
-    "    model.set_train(False)\n",
-    "    test_pred = []\n",
-    "    test_label = []\n",
-    "    for batch, (data, label) in enumerate(test_loader.create_tuple_iterator()):\n",
-    " \n",
-    "        pred = model(data)\n",
-    "\n",
-    "        pred[pred > 0.5] = float(1)\n",
-    "        pred[pred <= 0.5] = float(0)\n",
-    "\n",
-    "        preds = np.squeeze(pred, axis=0)\n",
-    "        img = np.transpose(preds,(1, 2, 0))\n",
-    "\n",
-    "        if not os.path.exists(result_path):\n",
-    "            os.makedirs(result_path)\n",
-    "        # io.imsave(os.path.join(result_path, \"%05d.png\" % batch), img.asnumpy())\n",
-    "        cv2.imwrite(os.path.join(result_path, \"%05d.png\" % batch), img.asnumpy()*255.)\n",
-    "\n",
-    "        test_pred.extend(pred.asnumpy())\n",
-    "        test_label.extend(label.asnumpy())\n",
-    "\n",
-    "    if have_mask:\n",
-    "        mtr = ['acc', 'iou', 'dice', 'sens', 'spec']\n",
-    "        metric = metrics_(mtr, smooth=1e-5)\n",
-    "        metric.clear()\n",
-    "        metric.update(test_pred, test_label)\n",
-    "        res = metric.eval()\n",
-    "        print(f'丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % (res[0], res[1], res[2], res[3], res[4]))\n",
-    "    else:\n",
-    "        print(\"Evaluation metrics cannot be calculated without Mask\")\n",
-    "\n",
-    "if __name__ == '__main__':\n",
-    "    net = UNet(3, 1)\n",
-    "    mindspore.load_checkpoint(\"checkpoint/best_UNet.ckpt\", net=net)\n",
-    "    result_path = \"predict\"\n",
-    "    test_dataset = create_dataset(\"src/datasets/ISBI/test/\", 224, 1, shuffle=False, have_mask=False)\n",
-    "    model_pred(net, test_dataset, result_path, have_mask=False)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "f72c470c",
-   "metadata": {},
-   "source": [
-    "#### 2.7 可视化预测结果"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 9,
-   "id": "bd765a3c",
-   "metadata": {
-    "scrolled": true
-   },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 12 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "data": {
-      "image/png": "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\n",
-      "text/plain": [
-       "<Figure size 432x288 with 12 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    }
-   ],
-   "source": [
-    "image_path = \"src/datasets/ISBI/test/image/\"\n",
-    "pred_path = \"predict/\"\n",
-    "\n",
-    "image_list = os.listdir(image_path)\n",
-    "pred_list = os.listdir(pred_path)[1:]\n",
-    "# print(image_list)\n",
-    "# print(pred_list)\n",
-    "test_image = np.array([cv2.imread(image_path + image_list[p], -1) for p in range(len(image_list))])\n",
-    "pred_masks = np.array([cv2.imread(pred_path + pred_list[p], -1) for p in range(len(pred_list))])\n",
-    "\n",
-    "show_image(test_image, num = 12)\n",
-    "show_image(pred_masks, num = 12)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "28df2dc3",
-   "metadata": {},
-   "source": [
-    "```\n",
-    "\n",
-    "|                | Acc  | IoU  | Dice | Sens | Spec |\n",
-    "| :------------: | :--: | :--: | :--: | ---: | ---- |\n",
-    "|   Train set    | 0.90 | 0.88 | 0.94 | 0.94 | 0.76 |\n",
-    "| Validation set | 0.92 | 0.90 | 0.95 | 0.96 | 0.77 |\n",
-    "```\n",
-    "\n",
-    "根据表1评价指标结构，本案例构建的网络模型具有较好的性能，能够实现对测试集进行较为准确的预测，针对测试集的部分预测结果如图4所示。\n",
-    "\n",
-    "<center>\n",
-    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwigmy47j20l2064dgy.jpg\" alt=\"image-20220819101847606\" style=\"zoom:50%;\" />\n",
-    "    <br>\n",
-    "    <div style=\"color:orange;\n",
-    "    display: inline-block;\n",
-    "    color: #999;\n",
-    "    padding: 2px;\">图4 模型预测结果</div>\n",
-    "</center>"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "b4e8303d",
-   "metadata": {},
-   "source": [
-    "### 3 总结\n",
-    "\n",
-    "本案例基于MindSpore框架针对ISBI数据集，完成了数据读取、数据集创建、Unet模型构建，并根据特定需求自定义了评估指标和回调函数，进行了模型训练和评估，顺利完成了预测结果的输出。通过此案例进一步加深了对Unet模型结构和特性的理解，并结合MindSpore框架提供的文档和教程，掌握了利用Mindspore框架实现特定案例的流程，以及多种API的使用方法，为以后在实际场景中应用MindSpore框架提供支持。"
-   ]
-  }
- ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "MindSpore",
-   "language": "python",
-   "name": "mindspore"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.7.6"
-  }
- },
- "nbformat": 4,
- "nbformat_minor": 5
-}
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "0",
+   "metadata": {},
+   "source": [
+    "## 基于MindSpore框架的UNet-2D案例实现\n",
+    "\n",
+    "### 1 模型简介\n",
+    "\n",
+    "Unet模型于2015年在论文《U-Net: Convolutional Networks for Biomedical Image Segmentation》中被提出，最初的提出是为了解决医学图像分割问题，用于细胞层面的图像分割任务。\n",
+    "\n",
+    "Unet模型是在FCN网络的基础上构建的，但由于FCN无法获取上下文信息以及位置信息，导致准确性较低，Unet模型由此引入了U型结构获取上述两种信息，并且模型结构简单高效、容易构建，在较小的数据集上也能实现较高的准确率。\n",
+    "\n",
+    "#### 1.1 模型结构\n",
+    "Unet模型的整体结构由两部分组成，即特征提取网络和特征融合网络，其结构也被称为“编码器-解码器结构”，并且由于网络整体结构类似于大写的英文字母“U”，故得名Unet，在其原始论文中定义的网络结构如图1所示。\n",
+    "\n",
+    "<center>\n",
+    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bw5io8tqj20n10fa3zo.jpg\" alt=\"image-20220819101847606\" style=\"zoom:67%;\" />\n",
+    "    <br>\n",
+    "    <div style=\"color:orange;\n",
+    "    display: inline-block;\n",
+    "    color: #999;\n",
+    "    padding: 2px;\">图1 Unet网络结构图</div>\n",
+    "</center>\n",
+    "\n",
+    "整个模型结构就是在原始图像输入后，首先进行特征提取，再进行特征融合：\n",
+    "\n",
+    "a) 左半部分负责特征提取的网络结构（即编码器结构）需要利用两个3x3的卷积核与2x2的池化层组成一个“下采样模块”，每一个下采样模块首先会对特征图进行两次valid卷积，再进行一次池化操作。由此经过4个下采样模块后，原始尺寸为572x572大小、通道数为1的原始图像，转换为了大小为28x28、通道数为1024的特征图。\n",
+    "\n",
+    "b) 右半部分负责进行上采样的网络结构（即解码器结构）需要利用1次反卷积操作、特征拼接操作以及两个3x3的卷积核作为一个“上采样模块”，每一个上采样模块首先会对特征图通过反卷积操作使图像尺寸增加1倍，再通过拼接编码器结构中的特征图使得通道数增加，最后经过两次valid卷积。由此经过4个上采样模块后，经过下采样模块的、大小为28x28、通道数为1024的特征图，转换为了大小为388x388、通道数为64的特征图。\n",
+    "\n",
+    "c) 网络结构的最后一部分是通过两个1x1的卷积核将经过上采样得到的通道数为64的特征图，转换为了通道数为2的图像作为预测结果输出。\n",
+    "#### 1.2 模型特点\n",
+    "\n",
+    "a) 利用拼接操作将低级特征图与高级特征图进行特征融合\n",
+    "\n",
+    "b) 完全对称的U型结构使得高分辨率信息和低分辨率信息在目标图片中增加，前后特征融合更为彻底。\n",
+    "\n",
+    "c) 结合了下采样时的低分辨率信息（提供物体类别识别依据）和上采样时的高分辨率信息（提供精准分割定位依据），此外还通过融合操作填补底层信息以提高分割精度。\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1",
+   "metadata": {},
+   "source": [
+    "### 2 案例实现\n",
+    "\n",
+    "#### 2.1 **环境准备与数据读取**\n",
+    "\n",
+    "本案例基于MindSpore-Ascend版本实现，在Ascend上完成模型训练。\n",
+    "\n",
+    "案例实现所使用的数据即ISBI果蝇电镜图数据集，直接从本地加载数据，在data目录下载好的数据集包括3个tif文件，分别对应测试集样本、训练集标签、训练集样本，文件路径结构如下：\n",
+    "\n",
+    "```\n",
+    ".datasets/\n",
+    "└── ISBI\n",
+    "    ├── test-volume.tif\n",
+    "    ├── train-labels.tif\n",
+    "    └── train-volume.tif\n",
+    "```\n",
+    "\n",
+    "其中每个tif文件都由30副图片压缩而成，所以接下来需要获取每个tif文件中所存储的所有图片，将其转换为png格式存储，得到训练集样本对应的30张png图片、训练集标签对应的30张png图片以及测试集样本对应的30张png图片。"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 需要Ascend版本的mindspore，因为cpu版本目前不支持mint.nn.conv2d\n",
+    "! pip install ml_collections\n",
+    "! pip install numpy\n",
+    "! pip install matplotlib"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import sys\n",
+    "import mindspore\n",
+    "print(sys.executable)\n",
+    "mindspore.set_device('Ascend')\n",
+    "print(mindspore.run_check())\n",
+    "from PIL import Image, ImageSequence\n",
+    "import math\n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "#显示下载好的数据\n",
+    "train_image_path = \"data/train-volume.tif\"\n",
+    "train_masks_path = \"data/train-labels.tif\"\n",
+    "image = np.array([np.array(p) for p in ImageSequence.Iterator(Image.open(train_image_path))])\n",
+    "masks = np.array([np.array(p) for p in ImageSequence.Iterator(Image.open(train_masks_path))])\n",
+    "\n",
+    "def show_image(image_list,num = 6):\n",
+    "    img_titles = []\n",
+    "    img_draws = []\n",
+    "    for ind,img in enumerate(image_list):\n",
+    "        if ind == num:\n",
+    "            break\n",
+    "        img_titles.append(ind)\n",
+    "        img_draws.append(img)\n",
+    "\n",
+    "    for i in range(len(img_titles)):\n",
+    "        if len(img_titles) > 6:\n",
+    "            row = 3\n",
+    "        elif 3<len(img_titles)<=6:\n",
+    "            row = 2\n",
+    "        else:\n",
+    "            row = 1\n",
+    "        col = math.ceil(len(img_titles)/row)\n",
+    "        plt.subplot(row,col,i+1),plt.imshow(img_draws[i],'gray')\n",
+    "        plt.title(img_titles[i])\n",
+    "        plt.xticks([]),plt.yticks([])\n",
+    "    plt.show()\n",
+    "    \n",
+    "show_image(image,num = 12)\n",
+    "show_image(masks,num = 12)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4",
+   "metadata": {},
+   "source": [
+    "具体的实现方式首先是将tif文件转换为数组形式，之后通过io操作将每张图片对应的数组存储为png图像，处理过后的训练集样本及其对应的标签图像如图2所示。将3个tif文件转换为png格式后，针对训练集的样本与标签，将其以2:1的比例，重新划分为了训练集与验证集，划分完成后的文件路径结构如下：\n",
+    "\n",
+    "```\n",
+    ".datasets/\n",
+    "└── ISBI\n",
+    "    ├── test_imgs\n",
+    "    │   ├── 00000.png\n",
+    "    │   ├── 00001.png\n",
+    "    │   └── . . . . .\n",
+    "    ├── train\n",
+    "    │   ├── image\n",
+    "    │   │   ├── 00001.png\n",
+    "    │   │   ├── 00002.png\n",
+    "    │   │   └── . . . . .\n",
+    "    │   └── mask\n",
+    "    │       ├── 00001.png\n",
+    "    │       ├── 00002.png\n",
+    "    │       └── . . . . .\n",
+    "    └── val\n",
+    "        ├── image\n",
+    "        │   ├── 00000.png\n",
+    "        │   ├── 00003.png\n",
+    "        │   └── . . . . .\n",
+    "        └── mask\n",
+    "            ├── 00000.png\n",
+    "            ├── 00003.png\n",
+    "            └── . . . . .\n",
+    "```\n",
+    "\n",
+    "<center>\n",
+    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwhda4g2j20fz0chabb.jpg\" alt=\"image-20220819101847606\" style=\"zoom:50%;\" />\n",
+    "    <br>\n",
+    "    <div style=\"color:orange;\n",
+    "    display: inline-block;\n",
+    "    color: #999;\n",
+    "    padding: 2px;\">图2 训练集样本及其对应标签</div>\n",
+    "</center>\n",
+    "\n",
+    "#### 2.2 数据集创建\n",
+    "\n",
+    "在进行上述tif文件格式转换，以及测试集和验证集的进一步划分后，就完成了数据读取所需的所有工作，接下来就需要利用处理好的图像数据，通过一定的图像变换来进行数据增强，并完成数据集的创建。\n",
+    "\n",
+    "数据增强部分是引入了mindspore.dataset.vision，针对训练集样本和标签，首先通过A.resize()方法将图像尺寸重新调整为统一大小，之后再进行转置以及水平翻转、垂直翻转，完成针对训练集样本和标签的数据增强。针对验证集的样本和标签，仅通过resize()方法将图像尺寸重新调整为统一大小。\n",
+    "\n",
+    "其次数据集的创建部分，首先是定义了Data_Loader类，在该类的__init__函数中，根据传入的data_path参数，确定在数据读取阶段设置好的、训练集和验证集的存储路径，再设置对应的样本和标签路径，并针对训练集和验证集的不同数据增强方法。在该类的__getitem__函数中，通过传入索引值读取训练集或验证集存储路径下的样本和标签图像，并对图像进行对应的数据增强操作，之后再对样本和标签的形状进行转置，就完成了__getitem__函数对样本和标签图像的读取。最后通过定义create_dataset函数，传入data_dir、batch_size等参数，在函数中实例化Data_Loader类获取data_dir，也就是训练集或验证集对应路径下的样本和标签元组对，再通过mindspore.dataset中的GeneratorDataset将元组转换为Tensor，最后通过设定好的batch_size将样本和标签按照batch_size大小分组，由此完成数据集的创建，上述流程对应代码如下："
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 划分数据集 tif->image\n",
+    "import os\n",
+    "import cv2\n",
+    "import numpy as np\n",
+    "\n",
+    "# 1. 定义保存路径 \n",
+    "root_path = 'datasets/ISBI'\n",
+    "train_img_dir = os.path.join(root_path, 'train', 'image')\n",
+    "train_mask_dir = os.path.join(root_path, 'train', 'mask')\n",
+    "val_img_dir = os.path.join(root_path, 'val', 'image')\n",
+    "val_mask_dir = os.path.join(root_path, 'val', 'mask')\n",
+    "\n",
+    "# 2. 创建文件夹\n",
+    "for d in [train_img_dir, train_mask_dir, val_img_dir, val_mask_dir]:\n",
+    "    os.makedirs(d, exist_ok=True)\n",
+    "\n",
+    "print(f\"正在转换数据... 总帧数: {len(image)}\")\n",
+    "\n",
+    "# 3. 开始切分并保存\n",
+    "# 2:1的方式\n",
+    "split_point = 20\n",
+    "\n",
+    "for i in range(len(image)):\n",
+    "    # 获取单帧数据\n",
+    "    img_frame = image[i]\n",
+    "    mask_frame = masks[i]\n",
+    "    \n",
+    "    # 决定存放在 训练集 还是 验证集\n",
+    "    if i < split_point:\n",
+    "        save_img_path = os.path.join(train_img_dir, f\"{i}.png\")\n",
+    "        save_mask_path = os.path.join(train_mask_dir, f\"{i}.png\")\n",
+    "    else:\n",
+    "        save_img_path = os.path.join(val_img_dir, f\"{i}.png\")\n",
+    "        save_mask_path = os.path.join(val_mask_dir, f\"{i}.png\")\n",
+    "    \n",
+    "    # 保存图片 \n",
+    "    cv2.imwrite(save_img_path, img_frame)\n",
+    "    cv2.imwrite(save_mask_path, mask_frame)\n",
+    "\n",
+    "print(\"转换完成！\")\n",
+    "print(f\"训练集保存路径: {train_img_dir}\")\n",
+    "print(f\"验证集保存路径: {val_img_dir}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "6",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import cv2\n",
+    "import mindspore.dataset as ds\n",
+    "import glob\n",
+    "import mindspore.dataset.vision as vision_C  #.c_transforms\n",
+    "import mindspore.dataset.transforms as C_transforms #.c_transform\n",
+    "import random\n",
+    "import mindspore\n",
+    "from mindspore.dataset.vision import Inter\n",
+    "\n",
+    "def train_transforms(img_size):\n",
+    "    return [\n",
+    "    vision_C.Resize(img_size, interpolation=Inter.NEAREST),\n",
+    "    vision_C.Rescale(1./255., 0.0),\n",
+    "    vision_C.RandomHorizontalFlip(prob=0.5),\n",
+    "    vision_C.RandomVerticalFlip(prob=0.5),\n",
+    "    vision_C.HWC2CHW()\n",
+    "    ]\n",
+    "\n",
+    "\n",
+    "def val_transforms(img_size):\n",
+    "    return [\n",
+    "    vision_C.Resize(img_size, interpolation=Inter.NEAREST),\n",
+    "    vision_C.Rescale(1/255., 0),\n",
+    "    vision_C.HWC2CHW()\n",
+    "    ]\n",
+    "\n",
+    "\n",
+    "\n",
+    "class Data_Loader:\n",
+    "    def __init__(self, data_path):\n",
+    "        # 初始化函数，读取所有data_path下的图片\n",
+    "        self.data_path = data_path\n",
+    "        self.imgs_path = glob.glob(os.path.join(data_path, 'image/*.png'))\n",
+    "        self.label_path = glob.glob(os.path.join(data_path, 'mask/*.png'))\n",
+    "\n",
+    "    def __getitem__(self, index):\n",
+    "        # 根据index读取图片\n",
+    "        image = cv2.imread(self.imgs_path[index])\n",
+    "        label = cv2.imread(self.label_path[index], cv2.IMREAD_GRAYSCALE)\n",
+    "        label = label.reshape((label.shape[0], label.shape[1], 1))\n",
+    "    \n",
+    "        return image, label\n",
+    "\n",
+    "    @property\n",
+    "    def column_names(self):\n",
+    "        column_names = ['image', 'label']\n",
+    "        return column_names\n",
+    "\n",
+    "    def __len__(self):\n",
+    "        # 返回训练集大小\n",
+    "        return len(self.imgs_path)\n",
+    "\n",
+    "\n",
+    "def create_dataset(data_dir, img_size, batch_size, augment, shuffle):\n",
+    "    mc_dataset = Data_Loader(data_path=data_dir)\n",
+    "    dataset = ds.GeneratorDataset(mc_dataset, mc_dataset.column_names, shuffle=shuffle)\n",
+    "\n",
+    "    if augment:\n",
+    "        transform_img = train_transforms(img_size)\n",
+    "    else:\n",
+    "        transform_img = val_transforms(img_size)\n",
+    "\n",
+    "    seed = random.randint(1,1000)\n",
+    "    mindspore.set_seed(seed)\n",
+    "    dataset = dataset.map(input_columns='image', num_parallel_workers=1, operations=transform_img)\n",
+    "    mindspore.set_seed(seed)\n",
+    "    dataset = dataset.map(input_columns=\"label\", num_parallel_workers=1, operations=transform_img)\n",
+    "\n",
+    "    if shuffle:\n",
+    "        dataset = dataset.shuffle(buffer_size=10000)\n",
+    "    dataset = dataset.batch(batch_size, num_parallel_workers=1)\n",
+    "    if augment == True and shuffle == True:\n",
+    "        print(\"训练集数据量：\", len(mc_dataset))\n",
+    "    elif augment == False and shuffle == False:\n",
+    "        print(\"验证集数据量：\", len(mc_dataset))\n",
+    "    else:\n",
+    "        pass\n",
+    "    return dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if __name__ == '__main__':\n",
+    "    train_dataset = create_dataset('datasets/ISBI/val', img_size=224, batch_size=3, augment=False, shuffle=False)\n",
+    "    for item, (image, label) in enumerate(train_dataset):\n",
+    "        if item < 5:\n",
+    "            print(f\"Shape of image [N, C, H, W]: {image.shape} {image.dtype}\",'---',f\"Shape of label [N, C, H, W]: {label.shape} {label.dtype}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "8",
+   "metadata": {},
+   "source": [
+    "#### 2.3 模型构建\n",
+    "\n",
+    "本案例实现中所构建的Unet模型结构与2015年论文中提出的Unet结构大致相同，但本案例中Unet网络模型的“下采样模块”与“上采样模块”使用的卷积类型都为Same卷积，而原论文中使用的是Valid卷积。此外，原论文的网络模型最终使用两个1x1的卷积核，输出了通道数2的预测图像，而本案例的网络模型最终使用的是1个1x1的卷积核，输出通道数为1的灰度图，和标签图像格式保持一致。实际构建的Unet模型结构如图3所示。\n",
+    "\n",
+    "<center>\n",
+    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwhpesqpj20no0g7jso.jpg\" alt=\"image-20220819101847606\" style=\"zoom:67%;\" />\n",
+    "    <br>\n",
+    "    <div style=\"color:orange;\n",
+    "    display: inline-block;\n",
+    "    color: #999;\n",
+    "    padding: 2px;\">图3 实际构建的Unet模型结构</div>\n",
+    "</center>\n",
+    "\n",
+    "MindSpore框架构建网络的流程与PyTorch类似，在定义模型类时需要继承Cell类，并重写__init__和construct方法。具体的实现方式首先是定义了一个DoubleConv模型类，在类中重写__init__方法，通过使用mint.nn.Conv2d层定义“下采样模块”与“上采样模块”中都使用到的两个卷积函数，并且在每个卷积层后加入mint.nn.BatchNorm2d层来对每次卷积后的特征图进行标准化，防止过拟合，以及使用mint.nn.ReLU层加入非线性的激活函数。之后在construct方法中使用定义好的运算构建前向网络。\n",
+    "\n",
+    "在DoubleConv模型类定义好之后，接下来就是通过定义UNet模型类来完成整个UNet网络的构建。在UNet模型类的__init__方法中实例化double_conv类来表示两个连续的卷积层，接着使用mint.nn.MaxPool2d来进行最大池化，由此完成了1个“下采样模块”的构建，重复4次即可完成网络中的编码器部分。针对解码器部分，使用了mint.nn.ResizeBilinear层来表示反卷积层，接着实例化了DoubleConv类来表示两个卷积层，由此完成了1个“上采样模块”的构建，重复4次即完成网络中解码器部分的搭建。之后通过1个mint.nn.Conv2d层来完成预测图像的输出。最后在construct方法中使用定义好的运算构建前向网络，由此完成整个Unet网络模型的构建。上述构建流程的对应代码如下所示：\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "9",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import mindspore as ms\n",
+    "from mindspore import nn, mint\n",
+    "\n",
+    "# DoubleConv \n",
+    "class DoubleConv(nn.Cell):\n",
+    "    def __init__(self, in_ch, out_ch):\n",
+    "        super().__init__()\n",
+    "        self.seq = nn.SequentialCell(\n",
+    "            mint.nn.Conv2d(in_ch, out_ch, kernel_size=3, padding=1, bias=False),\n",
+    "            mint.nn.BatchNorm2d(out_ch),\n",
+    "            mint.nn.ReLU(),\n",
+    "            mint.nn.Conv2d(out_ch, out_ch, kernel_size=3, padding=1, bias=False),\n",
+    "            mint.nn.BatchNorm2d(out_ch),\n",
+    "            mint.nn.ReLU()\n",
+    "        )\n",
+    "\n",
+    "    def construct(self, x):\n",
+    "        return self.seq(x)\n",
+    "\n",
+    "# 修改 UNet 类\n",
+    "class UNet(nn.Cell):\n",
+    "    def __init__(self, in_ch=3, n_classes=1):\n",
+    "        super(UNet, self).__init__()\n",
+    "        \n",
+    "        # --- 下采样部分 ---\n",
+    "        self.double_conv1 = DoubleConv(in_ch, 64)\n",
+    "        self.maxpool1 = mint.nn.MaxPool2d(kernel_size=2, stride=2)\n",
+    "        \n",
+    "        self.double_conv2 = DoubleConv(64, 128)\n",
+    "        self.maxpool2 = mint.nn.MaxPool2d(kernel_size=2, stride=2)\n",
+    "        \n",
+    "        self.double_conv3 = DoubleConv(128, 256)\n",
+    "        self.maxpool3 = mint.nn.MaxPool2d(kernel_size=2, stride=2)\n",
+    "        \n",
+    "        self.double_conv4 = DoubleConv(256, 512)\n",
+    "        self.maxpool4 = mint.nn.MaxPool2d(kernel_size=2, stride=2)\n",
+    "        \n",
+    "        self.double_conv5 = DoubleConv(512, 1024)\n",
+    "\n",
+    "        # --- 上采样部分 ---\n",
+    "        \n",
+    "        self.double_conv6 = DoubleConv(1024 + 512, 512)\n",
+    "        self.double_conv7 = DoubleConv(512 + 256, 256)\n",
+    "        self.double_conv8 = DoubleConv(256 + 128, 128)\n",
+    "        self.double_conv9 = DoubleConv(128 + 64, 64)\n",
+    "\n",
+    "        self.final = mint.nn.Conv2d(64, n_classes, kernel_size=1)\n",
+    "\n",
+    "    def construct(self, x):\n",
+    "        # 编码器 (Encoder)\n",
+    "        feature1 = self.double_conv1(x)\n",
+    "        tmp = self.maxpool1(feature1)\n",
+    "        \n",
+    "        feature2 = self.double_conv2(tmp)\n",
+    "        tmp = self.maxpool2(feature2)\n",
+    "        \n",
+    "        feature3 = self.double_conv3(tmp)\n",
+    "        tmp = self.maxpool3(feature3)\n",
+    "        \n",
+    "        feature4 = self.double_conv4(tmp)\n",
+    "        tmp = self.maxpool4(feature4)\n",
+    "        \n",
+    "        feature5 = self.double_conv5(tmp)\n",
+    "\n",
+    "        # 解码器 (Decoder)\n",
+    "        \n",
+    "        # Block 1\n",
+    "\n",
+    "        up_feature1 = mint.nn.functional.interpolate(\n",
+    "            feature5, \n",
+    "            size=feature4.shape[2:],  # 自动获取 feature4 的高和宽\n",
+    "            mode='bilinear', \n",
+    "            align_corners=True\n",
+    "        )\n",
+    "        tmp = mint.cat((feature4, up_feature1), dim=1) \n",
+    "        tmp = self.double_conv6(tmp)\n",
+    "        \n",
+    "        # Block 2\n",
+    "        up_feature2 = mint.nn.functional.interpolate(\n",
+    "            tmp, \n",
+    "            size=feature3.shape[2:], \n",
+    "            mode='bilinear', \n",
+    "            align_corners=True\n",
+    "        )\n",
+    "        tmp = mint.cat((feature3, up_feature2), dim=1)\n",
+    "        tmp = self.double_conv7(tmp)\n",
+    "        \n",
+    "        # Block 3\n",
+    "        up_feature3 = mint.nn.functional.interpolate(\n",
+    "            tmp, \n",
+    "            size=feature2.shape[2:], \n",
+    "            mode='bilinear', \n",
+    "            align_corners=True\n",
+    "        )\n",
+    "        tmp = mint.cat((feature2, up_feature3), dim=1)\n",
+    "        tmp = self.double_conv8(tmp)\n",
+    "        \n",
+    "        # Block 4\n",
+    "        up_feature4 = mint.nn.functional.interpolate(\n",
+    "            tmp, \n",
+    "            size=feature1.shape[2:], \n",
+    "            mode='bilinear', \n",
+    "            align_corners=True\n",
+    "        )\n",
+    "        tmp = mint.cat((feature1, up_feature4), dim=1)\n",
+    "        tmp = self.double_conv9(tmp)\n",
+    "        \n",
+    "        # Output\n",
+    "        # 使用  mint.sigmoid\n",
+    "        # output = mint.sigmoid(self.final(tmp))\n",
+    "        \n",
+    "        # return output\n",
+    "        return self.final(tmp)\n",
+    "\n",
+    "# 实例化模型\n",
+    "print(\"Re-initializing UNet...\")\n",
+    "net = UNet()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "10",
+   "metadata": {},
+   "source": [
+    "#### 2.4 自定义评估指标\n",
+    "\n",
+    "为了能够更加全面和直观的观察网络模型训练效果，本案例实现中还使用了MindSpore框架来自定义Metrics，在自定义的metrics类中使用了多种评价函数来评估模型的好坏，分别为准确率Acc、交并比IoU、Dice系数、灵敏度Sens、特异性Spec。\n",
+    "\n",
+    "a) 其中准确率Acc是图像中正确分类的像素百分比。即分类正确的像素占总像素的比例，用公式可表示为：\n",
+    "$$\n",
+    "A c c=\\frac{T P+T N}{T P+T N+F P+F N}\n",
+    "$$\n",
+    "其中：\n",
+    "\n",
+    "- TP：真阳性数，在label中为阳性，在预测值中也为阳性的个数。\n",
+    "- TN：真阴性数，在label中为阴性，在预测值中也为阴性的个数。\n",
+    "- FP：假阳性数，在label中为阴性，在预测值中为阳性的个数。\n",
+    "- FN：假阴性数，在label中为阳性，在预测值中为阴性的个数。\n",
+    "\n",
+    "b) 交并比IoU是预测分割和标签之间的重叠区域除以预测分割和标签之间的联合区域（两者的交集/两者的并集），是语义分割中最常用的指标之一，其计算公式为：\n",
+    "$$\n",
+    "I o U=\\frac{|A \\cap B|}{|A \\cup B|}=\\frac{T P}{T P+F P+F N}\n",
+    "$$\n",
+    "c) Dice系数定义为两倍的交集除以像素和，也叫F1 score，与IoU呈正相关关系，其计算公式为：\n",
+    "$$\n",
+    "\\text { Dice }=\\frac{2|A \\cap B|}{|A|+|B|}=\\frac{2 T P}{2 T P+F P+F N}\n",
+    "$$\n",
+    "d) 敏感度Sens和特异性Spec分别是描述识别出的阳性占所有阳性的比例，以及描述识别出的负例占所有负例的比例，计算公式分别为：\n",
+    "$$\n",
+    "\\text { Sens }=\\frac{T P}{T P+F N}\n",
+    "$$\n",
+    "\n",
+    "$$\n",
+    "\\text { Spec }=\\frac{T N}{F P+T N}\n",
+    "$$\n",
+    "\n",
+    "具体的实现方法首先是自定义metrics_类，并按照MindSpore官方文档继承nn.Metric父类，接着根据上述5个评价指标的计算公式，在类中定义5个指标的计算方法，之后通过重新实现clear方法来初始化相关参数；重新实现update方法来传入模型预测值和标签，通过上述定义的各评价指标计算方法，计算每个指标的值并存入一个列表；最后通过重新实现eval方法来讲存储各评估指标值的列表返回。上述流程对应的代码如下：\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "11",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import mindspore as ms\n",
+    "from mindspore import nn, mint, Tensor\n",
+    "\n",
+    "class metrics_(nn.Metric):\n",
+    "    def __init__(self, metrics, smooth=1e-5):\n",
+    "        \"\"\"\n",
+    "        初始化\n",
+    "        metrics: list, e.g. [\"acc\", \"iou\", \"dice\", \"sens\", \"spec\"]\n",
+    "        \"\"\"\n",
+    "        super(metrics_, self).__init__()\n",
+    "        self.metrics = metrics\n",
+    "        self.smooth = float(smooth) # 简单类型转换，替代 Validator\n",
+    "        self.metrics_list = [0. for i in range(len(self.metrics))]\n",
+    "        self._samples_num = 0\n",
+    "        self.clear()\n",
+    "\n",
+    "    def clear(self):\n",
+    "        \"\"\"清除内部评估结果\"\"\"\n",
+    "        self.metrics_list = [0. for i in range(len(self.metrics))]\n",
+    "        self._samples_num = 0\n",
+    "\n",
+    "    # 以下所有 Metrics 方法均接收 Tensor 输入，使用 mint 算子计算\n",
+    "    # 使用 .item() 将单步结果转为 Python float 累加，保持原逻辑的数值精度\n",
+    "    \n",
+    "    def Acc_metrics(self, y_pred, y):\n",
+    "        # 展平\n",
+    "        y_pred_f = y_pred.view(-1)\n",
+    "        y_f = y.view(-1)\n",
+    "        \n",
+    "        # 计算相等元素的个数\n",
+    "        tp = mint.sum(mint.eq(y_pred_f, y_f)).float()\n",
+    "        total = float(len(y_pred_f))\n",
+    "        \n",
+    "        single_acc = tp / total\n",
+    "        return single_acc.item()\n",
+    "\n",
+    "    def IoU_metrics(self, y_pred, y):\n",
+    "        y_pred_f = y_pred.view(-1)\n",
+    "        y_f = y.view(-1)\n",
+    "        \n",
+    "        intersection = mint.sum(y_pred_f * y_f)\n",
+    "        # Union = A + B - Intersection\n",
+    "        unionset = mint.sum(y_pred_f) + mint.sum(y_f) - intersection\n",
+    "        \n",
+    "        single_iou = intersection / (unionset + self.smooth)\n",
+    "        return single_iou.item()\n",
+    "\n",
+    "    def Dice_metrics(self, y_pred, y):\n",
+    "        y_pred_f = y_pred.view(-1)\n",
+    "        y_f = y.view(-1)\n",
+    "        \n",
+    "        intersection = mint.sum(y_pred_f * y_f)\n",
+    "        unionset = mint.sum(y_pred_f) + mint.sum(y_f)\n",
+    "        \n",
+    "        single_dice = 2 * intersection / (unionset + self.smooth)\n",
+    "        return single_dice.item()\n",
+    "\n",
+    "    def Sens_metrics(self, y_pred, y):\n",
+    "        y_pred_f = y_pred.view(-1)\n",
+    "        y_f = y.view(-1)\n",
+    "        \n",
+    "        tp = mint.sum(y_pred_f * y_f)\n",
+    "        actual_positives = mint.sum(y_f)\n",
+    "        \n",
+    "        single_sens = tp / (actual_positives + self.smooth)\n",
+    "        return single_sens.item()\n",
+    "\n",
+    "    def Spec_metrics(self, y_pred, y):\n",
+    "        y_pred_f = y_pred.view(-1)\n",
+    "        y_f = y.view(-1)\n",
+    "        \n",
+    "        # TN: (1-pred) * (1-y)\n",
+    "        true_neg = mint.sum((1 - y_f) * (1 - y_pred_f))\n",
+    "        total_neg = mint.sum(1 - y_f)\n",
+    "        \n",
+    "        single_spec = true_neg / (total_neg + self.smooth)\n",
+    "        return single_spec.item()\n",
+    "\n",
+    "    def update(self, *inputs):\n",
+    "        if len(inputs) != 2:\n",
+    "            raise ValueError(\"For 'update', it needs 2 inputs (predicted value, true value), \"\n",
+    "                             \"but got {}.\".format(len(inputs)))\n",
+    "        \n",
+    "        y_pred_raw = inputs[0]\n",
+    "        y_true = inputs[1]\n",
+    "\n",
+    "        # 确保数据类型一致，通常转为 float32 进行计算\n",
+    "        y_true = y_true.float()\n",
+    "        \n",
+    "        # 使用 mint 算子进行二值化\n",
+    "        y_pred = (y_pred_raw > 0.5).float()\n",
+    "\n",
+    "        if y_pred.shape != y_true.shape:\n",
+    "            raise ValueError(f\"For 'update', shapes must match. \"\n",
+    "                             f\"Pred: {y_pred.shape}, True: {y_true.shape}.\")\n",
+    "\n",
+    "        batch_size = y_true.shape[0]\n",
+    "        self._samples_num += batch_size\n",
+    "\n",
+    "        # 逐样本计算指标\n",
+    "        for i in range(batch_size):\n",
+    "            if \"acc\" in self.metrics:\n",
+    "                self.metrics_list[0] += self.Acc_metrics(y_pred[i], y_true[i])\n",
+    "            if \"iou\" in self.metrics:\n",
+    "                self.metrics_list[1] += self.IoU_metrics(y_pred[i], y_true[i])\n",
+    "            if \"dice\" in self.metrics:\n",
+    "                self.metrics_list[2] += self.Dice_metrics(y_pred[i], y_true[i])\n",
+    "            if \"sens\" in self.metrics:\n",
+    "                self.metrics_list[3] += self.Sens_metrics(y_pred[i], y_true[i])\n",
+    "            if \"spec\" in self.metrics:\n",
+    "                self.metrics_list[4] += self.Spec_metrics(y_pred[i], y_true[i])\n",
+    "\n",
+    "    def eval(self):\n",
+    "        if self._samples_num == 0:\n",
+    "            raise RuntimeError(\"Samples number is 0, please call update before eval.\")\n",
+    "            \n",
+    "        # 计算平均值\n",
+    "        return [val / float(self._samples_num) for val in self.metrics_list]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "12",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "x = Tensor(np.array([[[[0.2, 0.5, 0.7], [0.3, 0.1, 0.2], [0.9, 0.6, 0.8]]]]))\n",
+    "y = Tensor(np.array([[[[0, 1, 1], [1, 0, 0], [0, 1, 1]]]]))\n",
+    "metric = metrics_([\"acc\", \"iou\", \"dice\", \"sens\", \"spec\"],smooth=1e-5)\n",
+    "metric.clear()\n",
+    "metric.update(x, y)\n",
+    "res = metric.eval()\n",
+    "print( '丨acc: %.4f丨丨iou: %.4f丨丨dice: %.4f丨丨sens: %.4f丨丨spec: %.4f丨' % (res[0], res[1], res[2], res[3],res[4]), flush=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "13",
+   "metadata": {},
+   "source": [
+    "#### 2.5 模型训练及评估\n",
+    "\n",
+    "在模型训练时，首先是设置模型训练的epoch次数为50，再通过2.1节中自定义的create_dataset方法创建了训练集和验证集，其中训练集batch_size大小为4，验证集batch_size大小为2，图像尺寸统一调整为224x224；损失函数使用nn.BCELoss，优化器使用nn.Adam，并设置学习率为0.01。回调函数方面使用了LossMonitor和TimeMonitor来监控训练过程中每个epoch结束后，损失值Loss的变化情况以及每个epoch、每个step的运行时间，还实例化了2.5节中自定义的回调类EvalCallBack，实现计算每个epoch结束后，在2.4节中定义的5个评估指标，并保存当前最优模型。在50个epcoh结束后，模型在训练集和验证集上的评估指标如表1所示：\n",
+    "\n",
+    "模型训练部分的代码如下："
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "14",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import mindspore as ms\n",
+    "from mindspore import nn, mint, value_and_grad\n",
+    "import ml_collections\n",
+    "\n",
+    "# 1. 设置动态图模式  一般默认情况下是动态图模式，因此此次可以不需要设置动态图模式\n",
+    "# ms.set_context(mode=ms.PYNATIVE_MODE)\n",
+    "\n",
+    "def get_config():\n",
+    "    \"\"\"configuration \"\"\"\n",
+    "    config = ml_collections.ConfigDict()\n",
+    "    config.epochs = 100\n",
+    "    # 请确保路径正确\n",
+    "    config.train_data_path = \"datasets/ISBI/train/\"\n",
+    "    config.val_data_path = \"datasets/ISBI/val/\"\n",
+    "    config.imgsize = 224\n",
+    "    config.batch_size = 4\n",
+    "    config.pretrained_path = None\n",
+    "    config.in_channel = 3\n",
+    "    config.n_classes = 1\n",
+    "    config.lr = 0.001\n",
+    "    return config\n",
+    "\n",
+    "cfg = get_config()\n",
+    "\n",
+    "train_dataset = create_dataset(cfg.train_data_path, img_size=cfg.imgsize, batch_size= cfg.batch_size, augment=True, shuffle = True)\n",
+    "val_dataset = create_dataset(cfg.val_data_path, img_size=cfg.imgsize, batch_size= cfg.batch_size, augment=False, shuffle = False)\n",
+    "\n",
+    "def train(model, dataset, loss_fn, optimizer, met):\n",
+    "    # 定义正向计算函数\n",
+    "    def forward_fn(data, label):\n",
+    "        logits = model(data)\n",
+    "        loss = loss_fn(logits, label)\n",
+    "        return loss, logits\n",
+    "\n",
+    "    # 获取梯度计算函数 \n",
+    "    grad_fn = value_and_grad(forward_fn, None, optimizer.parameters, has_aux=True)\n",
+    "\n",
+    "    size = dataset.get_dataset_size()\n",
+    "    model.set_train(True)\n",
+    "    \n",
+    "    # 实例化指标计算器 \n",
+    "    metric = metrics_(met, smooth=1e-5)\n",
+    "    metric.clear()\n",
+    "    \n",
+    "    train_loss = 0\n",
+    "\n",
+    "    # 使用迭代器\n",
+    "    iterator = dataset.create_tuple_iterator()\n",
+    "    \n",
+    "    for batch, (data, label) in enumerate(iterator):\n",
+    "        # 1. 计算梯度\n",
+    "        (loss, logits), grads = grad_fn(data, label)\n",
+    "        current_loss = loss.item()\n",
+    "        \n",
+    "        # 2. 更新指标\n",
+    "        metric.update(logits, label)\n",
+    "\n",
+    "        # 3. 优化器更新参数 \n",
+    "        optimizer(grads)\n",
+    "        \n",
+    "        # 4. 累计 Loss (用于显示)\n",
+    "        train_loss += current_loss\n",
+    "        \n",
+    "\n",
+    "    train_loss /= size\n",
+    "    \n",
+    "    # 5. 计算最终指标\n",
+    "    res = metric.eval()\n",
+    "    print(f'Train loss:{train_loss:>4f}','丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % (res[0], res[1], res[2], res[3], res[4]))\n",
+    "\n",
+    "\n",
+    "def val(model, dataset, loss_fn, met):\n",
+    "    size = dataset.get_dataset_size()\n",
+    "    model.set_train(False)\n",
+    "    \n",
+    "    # 实例化指标计算器\n",
+    "    metric = metrics_(met, smooth=1e-5)\n",
+    "    metric.clear()\n",
+    "    \n",
+    "    val_loss = 0\n",
+    "\n",
+    "    iterator = dataset.create_tuple_iterator()\n",
+    "    \n",
+    "    for batch, (data, label) in enumerate(iterator):\n",
+    "        # 1. 前向推理\n",
+    "        logits = model(data)\n",
+    "        \n",
+    "        # 2. 计算 Loss\n",
+    "        loss = loss_fn(logits, label)\n",
+    "        val_loss += loss.item()\n",
+    "        \n",
+    "        # 3. 更新指标 (Batch-wise update)\n",
+    "        metric.update(logits, label)\n",
+    "\n",
+    "    val_loss /= size\n",
+    "    \n",
+    "    # 4. 计算最终指标\n",
+    "    res = metric.eval()\n",
+    "\n",
+    "    print(f'Val loss:{val_loss:>4f}','丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % (res[0], res[1], res[2], res[3], res[4]))\n",
+    "\n",
+    "    checkpoint = res[1] \n",
+    "    return checkpoint, res[4]\n",
+    "\n",
+    "# --- 主程序部分 ---\n",
+    "\n",
+    "# 1. 实例化模型\n",
+    "net = UNet(in_ch=cfg.in_channel, n_classes=cfg.n_classes)\n",
+    "\n",
+    "# 2. 定义 Loss \n",
+    "criterion = mint.nn.BCEWithLogitsLoss()\n",
+    "\n",
+    "# 3. 定义优化器 \n",
+    "optimizer = mint.optim.SGD(params=net.trainable_params(), lr=cfg.lr)\n",
+    "\n",
+    "# 获取步数信息\n",
+    "iters_per_epoch = train_dataset.get_dataset_size()\n",
+    "total_train_steps = iters_per_epoch * cfg.epochs\n",
+    "print('iters_per_epoch: ', iters_per_epoch)\n",
+    "print('total_train_steps: ', total_train_steps)\n",
+    "\n",
+    "metrics_name = [\"acc\", \"iou\", \"dice\", \"sens\", \"spec\"]\n",
+    "\n",
+    "best_iou = 0\n",
+    "ckpt_path = 'checkpoint/best_UNet.ckpt'\n",
+    "\n",
+    "# 创建保存目录\n",
+    "if not os.path.exists(\"checkpoint\"):\n",
+    "    os.makedirs(\"checkpoint\")\n",
+    "\n",
+    "for epoch in range(cfg.epochs):\n",
+    "    print(f\"Epoch [{epoch+1} / {cfg.epochs}]\")\n",
+    "    \n",
+    "    # 训练\n",
+    "    train(net, train_dataset, criterion, optimizer, metrics_name)\n",
+    "    \n",
+    "    # 验证\n",
+    "    checkpoint_best, spec = val(net, val_dataset, criterion, metrics_name)\n",
+    "    \n",
+    "    # 保存\n",
+    "    if epoch > 2 and spec > 0.2:\n",
+    "        if checkpoint_best > best_iou:\n",
+    "            print('IoU improved from %0.4f to %0.4f' % (best_iou, checkpoint_best))\n",
+    "            best_iou = checkpoint_best\n",
+    "            ms.save_checkpoint(net, ckpt_path)\n",
+    "            print(\"saving best checkpoint at: {} \".format(ckpt_path))\n",
+    "        else:\n",
+    "            print('IoU did not improve from %0.4f' % (best_iou),\"\\n-------------------------------\")\n",
+    "            \n",
+    "print(\"Done!\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "15",
+   "metadata": {},
+   "source": [
+    "#### 2.6 模型预测\n",
+    "在预测部分需要创建一个测试数据集，放在datasets/ISBI/test/文件夹下，代码如下："
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "16",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import os\n",
+    "import cv2\n",
+    "import glob\n",
+    "import random\n",
+    "import numpy as np\n",
+    "from tqdm import tqdm\n",
+    "\n",
+    "import mindspore\n",
+    "import mindspore.dataset as ds\n",
+    "import mindspore.dataset.vision as vision\n",
+    "import mindspore.dataset.transforms as transforms\n",
+    "from mindspore import mint, ops, Tensor\n",
+    "\n",
+    "\n",
+    "\n",
+    "def val_transforms(img_size):\n",
+    "    \"\"\"\n",
+    "    使用标准的 mindspore.dataset.vision 接口\n",
+    "    \"\"\"\n",
+    "    return transforms.Compose([\n",
+    "        vision.Resize(img_size, interpolation=vision.Inter.NEAREST),\n",
+    "        vision.Rescale(1/255., 0),\n",
+    "        vision.HWC2CHW()\n",
+    "    ])\n",
+    "\n",
+    "class Data_Loader:\n",
+    "    def __init__(self, data_path, have_mask):\n",
+    "        self.data_path = data_path\n",
+    "        self.have_mask = have_mask\n",
+    "        self.imgs_path = glob.glob(os.path.join(data_path, 'image/*.png'))\n",
+    "        if self.have_mask:\n",
+    "            self.label_path = glob.glob(os.path.join(data_path, 'mask/*.png'))\n",
+    "\n",
+    "    def __getitem__(self, index):\n",
+    "        image = cv2.imread(self.imgs_path[index])\n",
+    "        if self.have_mask:\n",
+    "            label = cv2.imread(self.label_path[index], cv2.IMREAD_GRAYSCALE)\n",
+    "            label = label.reshape((label.shape[0], label.shape[1], 1))\n",
+    "        else:\n",
+    "\n",
+    "            label = image \n",
+    "        return image, label\n",
+    "\n",
+    "    @property\n",
+    "    def column_names(self):\n",
+    "        return ['image', 'label']\n",
+    "\n",
+    "    def __len__(self):\n",
+    "        return len(self.imgs_path)\n",
+    "\n",
+    "def create_dataset(data_dir, img_size, batch_size, shuffle, have_mask=False):\n",
+    "    mc_dataset = Data_Loader(data_path=data_dir, have_mask=have_mask)\n",
+    "    print(f\"Dataset size: {len(mc_dataset)}\")\n",
+    "    \n",
+    "    # 定义数据集\n",
+    "    dataset = ds.GeneratorDataset(mc_dataset, mc_dataset.column_names, shuffle=shuffle)\n",
+    "    \n",
+    "    # 预处理操作\n",
+    "    transform_img = val_transforms(img_size)\n",
+    "    \n",
+    "    # 设置随机种子\n",
+    "    seed = random.randint(1, 1000)\n",
+    "    ds.config.set_seed(seed) # 更新为推荐的设置种子方式\n",
+    "    \n",
+    "    # Map 操作\n",
+    "    dataset = dataset.map(input_columns='image', num_parallel_workers=1, operations=transform_img)\n",
+    "    dataset = dataset.map(input_columns=\"label\", num_parallel_workers=1, operations=transform_img)\n",
+    "    \n",
+    "    dataset = dataset.batch(batch_size, num_parallel_workers=1)\n",
+    "    return dataset\n",
+    "\n",
+    "def model_pred(model, test_loader, result_path, have_mask):\n",
+    "    \"\"\"\n",
+    "    使用 MindSpore 2.7.1+ 的 mint 接口进行预测后处理\n",
+    "    \"\"\"\n",
+    "    model.set_train(False)\n",
+    "    test_pred = []\n",
+    "    test_label = []\n",
+    "    \n",
+    "    if not os.path.exists(result_path):\n",
+    "        os.makedirs(result_path)\n",
+    "\n",
+    "    print(\"Start Prediction...\")\n",
+    "    # 使用 create_tuple_iterator 获取数据\n",
+    "    for batch_idx, (data, label) in enumerate(tqdm(test_loader.create_tuple_iterator())):\n",
+    "        \n",
+    "        # 1. 模型推理 (输出通常为 Logits 或 Probabilities)\n",
+    "        logits = model(data)\n",
+    "        \n",
+    "        # 2. 使用 mint 接口进行二值化处理\n",
+    "\n",
+    "        pred_binary = (logits > 0.5).float() \n",
+    "\n",
+    "        # 3. 维度变换 (Tensor操作)\n",
+    "\n",
+    "        img_tensor = mint.squeeze(pred_binary, dim=0)\n",
+    "        \n",
+    "        # Mint: permute 调整通道顺序 (C, H, W) -> (H, W, C)\n",
+    "        img_tensor = img_tensor.permute(1, 2, 0)\n",
+    "\n",
+    "        # 4. 转为 Numpy 进行保存 (cv2 需要 numpy)\n",
+    "\n",
+    "        img_np = img_tensor.asnumpy() * 255.0\n",
+    "        \n",
+    "        # 保存结果\n",
+    "        cv2.imwrite(os.path.join(result_path, \"%05d.png\" % batch_idx), img_np)\n",
+    "\n",
+    "        # 5. 收集结果用于指标计算 (转为扁平列表)\n",
+    " \n",
+    "        test_pred.extend(pred_binary.asnumpy().flatten())\n",
+    "        test_label.extend(label.asnumpy().flatten())\n",
+    "\n",
+    "    if have_mask:\n",
+    "\n",
+    "        mtr = ['acc', 'iou', 'dice', 'sens', 'spec']\n",
+    "        try:\n",
+    "\n",
+    "            metric = metrics_(mtr, smooth=1e-5) \n",
+    "            metric.clear()\n",
+    "            metric.update(test_pred, test_label)\n",
+    "            res = metric.eval()\n",
+    "            print(f'丨acc: %.3f丨丨iou: %.3f丨丨dice: %.3f丨丨sens: %.3f丨丨spec: %.3f丨' % \n",
+    "                  (res[0], res[1], res[2], res[3], res[4]))\n",
+    "        except NameError:\n",
+    "            print(\"Warning: 'metrics_' class is not defined. Skipping metric evaluation.\")\n",
+    "    else:\n",
+    "        print(\"Evaluation metrics cannot be calculated without Mask\")\n",
+    "\n",
+    "\n",
+    "net = UNet(3, 1)\n",
+    "# 加载权重\n",
+    "param_dict = mindspore.load_checkpoint(\"checkpoint/best_UNet.ckpt\")\n",
+    "mindspore.load_param_into_net(net, param_dict)\n",
+    "        \n",
+    "result_path = \"predict\"\n",
+    "# 创建测试集\n",
+    "test_dataset = create_dataset(\"datasets/ISBI/test/\", 224, 1, shuffle=False, have_mask=False)\n",
+    "        \n",
+    "# 执行预测\n",
+    "model_pred(net, test_dataset, result_path, have_mask=False)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "17",
+   "metadata": {},
+   "source": [
+    "#### 2.7 可视化预测结果"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "18",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "image_path = \"datasets/ISBI/test/image/\"\n",
+    "pred_path = \"predict/\"\n",
+    "\n",
+    "image_list = os.listdir(image_path)\n",
+    "pred_list = os.listdir(pred_path)[1:]\n",
+    "# print(image_list)\n",
+    "# print(pred_list)\n",
+    "test_image = np.array([cv2.imread(image_path + image_list[p], -1) for p in range(len(image_list))])\n",
+    "pred_masks = np.array([cv2.imread(pred_path + pred_list[p], -1) for p in range(len(pred_list))])\n",
+    "\n",
+    "show_image(test_image, num = 12)\n",
+    "show_image(pred_masks, num = 12)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "19",
+   "metadata": {},
+   "source": [
+    "```\n",
+    "\n",
+    "|                | Acc  | IoU  | Dice | Sens | Spec |\n",
+    "| :------------: | :--: | :--: | :--: | ---: | ---- |\n",
+    "|   Train set    | 0.90 | 0.88 | 0.94 | 0.94 | 0.76 |\n",
+    "| Validation set | 0.92 | 0.90 | 0.95 | 0.96 | 0.77 |\n",
+    "```\n",
+    "\n",
+    "根据表1评价指标结构，本案例构建的网络模型具有较好的性能，能够实现对测试集进行较为准确的预测，针对测试集的部分预测结果如图4所示。\n",
+    "\n",
+    "<center>\n",
+    "    <img src=\"https://tva1.sinaimg.cn/large/e6c9d24ely1h5bwigmy47j20l2064dgy.jpg\" alt=\"image-20220819101847606\" style=\"zoom:50%;\" />\n",
+    "    <br>\n",
+    "    <div style=\"color:orange;\n",
+    "    display: inline-block;\n",
+    "    color: #999;\n",
+    "    padding: 2px;\">图4 模型预测结果</div>\n",
+    "</center>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "20",
+   "metadata": {},
+   "source": [
+    "### 3 总结\n",
+    "\n",
+    "本案例基于MindSpore框架针对ISBI数据集，完成了数据读取、数据集创建、Unet模型构建，并根据特定需求自定义了评估指标和回调函数，进行了模型训练和评估，顺利完成了预测结果的输出。通过此案例进一步加深了对Unet模型结构和特性的理解，并结合MindSpore框架提供的文档和教程，掌握了利用Mindspore框架实现特定案例的流程，以及多种API的使用方法，为以后在实际场景中应用MindSpore框架提供支持。"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "MindSpore (Ascend)",
+   "language": "python",
+   "name": "mindspore"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.10.19"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}