为何imagefolder传入参数时一直显示“无法识别的参数”

class ImageFolder(Dataset):
    def __init__(self, root, transform=None,patch_size=(256,256), split='train',need_file_name = False):
        splitdir = Path(root) / split  # 相当于osp.join
        if not splitdir.is_dir():
            raise RuntimeError(f'Invalid directory "{root}"')

        splitdir_left = splitdir / "left"
        splitdir_right = splitdir / "right"

        self.left_list = sorted(glob.glob(os.path.join(splitdir_left,"*")))
        self.right_list = sorted(glob.glob(os.path.join(splitdir_right, "*")))

        self.patch_size = patch_size
        #只保留了ToTensor
        self.transform = transform

        ###for homography 单独裁剪 不传参直接设定
        self.homopic_size = 256
        self.homopatch_size = 128
        self.rho = 45
        self.homotransforms = transforms.Compose(
            [
                # ywz
                # transforms.Resize(self.homopic_size),
                # #
                # transforms.CenterCrop(self.homopic_size),
                transforms.ToTensor(),
                transforms.Normalize(mean=MEAN, std=STD),
            ]
        )
        ########################################

        self.need_file_name = need_file_name

    def __getitem__(self, index):
        """
        Args:
            index (int): Index

        Returns:
            img: `PIL.Image.Image` or transformed `PIL.Image.Image`.
        """
        # img1 = Image.open(self.left_list[index]).convert('RGB')
        # img2 = Image.open(self.right_list[index]).convert('RGB')
        if os.path.basename(self.left_list[index]) != os.path.basename(self.right_list[index]):
            print(self.left_list[index])
            raise ValueError("cannot compare pictures.")
        ##
        img1 = cv2.imread(self.left_list[index])
        img1 = cv2.cvtColor(img1, cv2.COLOR_BGR2RGB)
        img2 = cv2.imread(self.right_list[index])
        img2 = cv2.cvtColor(img2, cv2.COLOR_BGR2RGB)
        #random cut for pair
        H, W, _ = img1.shape
        #randint是闭区间
        # print(H)
        # print(W)
        # print(self.patch_size)
        if self.patch_size[0]==H:
            startH = 0
            startW = 0
        else:
            startH = random.randint(0,H-self.patch_size[0]-1)
            startW = random.randint(0,W-self.patch_size[1]-1)

        img1 = img1[startH:(startH + self.patch_size[0]), startW:(startW + self.patch_size[1])]
        img2 = img2[startH:(startH + self.patch_size[0]), startW:(startW + self.patch_size[1])]
        ###
        # print(img1.shape)  #（512，512，3）
        # raise ValueError("stop utils")

        H_list = get_H(img1,img2) #可以忽略 这个是传统单应性获取的方法

        #for homo 在上述patch基础上再进行缩放和裁剪 返回patch以及相应的corners
        homo_img1 = cv2.resize(img1,(self.homopic_size,self.homopic_size))
        homo_img2 = cv2.resize(img2, (self.homopic_size, self.homopic_size))
        homo_img1 = self.homotransforms(homo_img1)
        homo_img2 = self.homotransforms(homo_img2)
        homo_img1 = torch.mean(homo_img1, dim=0, keepdim=True)  # 转灰度
        homo_img2 = torch.mean(homo_img2, dim=0, keepdim=True)  # 转灰度

        # pick top left corner
        if self.homopic_size - self.rho - self.homopatch_size >= self.rho:
            x = random.randint(self.rho, self.homopic_size - self.rho - self.homopatch_size)
            y = random.randint(self.rho, self.homopic_size - self.rho - self.homopatch_size)
        else:
            x = 0
            y = 0
        # print(x,y)
        corners = torch.tensor(
            [
                [x, y],
                [x + self.homopatch_size, y],
                [x + self.homopatch_size, y + self.homopatch_size],
                [x, y + self.homopatch_size],
            ],dtype=torch.float32
        )
        homo_img1 = homo_img1[:, y: y + self.homopatch_size, x: x + self.homopatch_size]
        homo_img2 = homo_img2[:, y: y + self.homopatch_size, x: x + self.homopatch_size]
        ################## [homo_img1,homo_img2,corners]


        ##
        if H_list[0]==None:
            print(self.left_list[index])
            print(self.right_list[index])
            #raise ValueError("None!!H_matrix")
            # 只有ToTensor
            if self.transform:
                return self.transform(img1), self.transform(img2) # ,H_list[1],H_list[2],H_list[3]
            return img1, img2  # ,H_list[1],H_list[2],H_list[3]

        #只有ToTensor
        if self.transform:
            # return self.transform(img1),self.transform(img2),H_list[0] #,H_list[1],H_list[2],H_list[3]
            if self.need_file_name:
                return self.transform(img1), self.transform(img2), H_list[0], os.path.basename(self.left_list[index]),homo_img1,homo_img2,corners  # ,H_list[1],H_list[2],H_list[3]
            else:
                return self.transform(img1), self.transform(img2), H_list[0],homo_img1,homo_img2,corners  # ,H_list[1],H_list[2],H_list[3]

        if self.need_file_name:
            return img1, img2, H_list[0],os.path.basename(self.left_list[index]),homo_img1,homo_img2,corners  # ,H_list[1],H_list[2],H_list[3]
        else:
            return img1,img2,H_list[0],homo_img1,homo_img2,corners #,H_list[1],H_list[2],H_list[3]

    def __len__(self):
        return len(self.left_list)

应该是args.patch-size，你在parser.argument定义的为--patch-size，所以要用短横杠的，除非把parser.argument中的定义改为下划线的--patch_size