How to use the torchfcn.datasets.voc.VOCClassSegBase function in torchfcn

def __init__(self, root, split='train', transform=False):
        super(VOC2011ClassSeg, self).__init__(
            root, split=split, transform=transform)
        pkg_root = osp.join(osp.dirname(osp.realpath(__file__)), '..')
        imgsets_file = osp.join(
            pkg_root, 'ext/fcn.berkeleyvision.org',
            'data/pascal/seg11valid.txt')
        dataset_dir = osp.join(self.root, 'VOC/VOCdevkit/VOC2012')
        for did in open(imgsets_file):
            did = did.strip()
            img_file = osp.join(dataset_dir, 'JPEGImages/%s.jpg' % did)
            lbl_file = osp.join(dataset_dir, 'SegmentationClass/%s.png' % did)
            self.files['seg11valid'].append({'img': img_file, 'lbl': lbl_file})


class VOC2012ClassSeg(VOCClassSegBase):

    url = 'http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar'  # NOQA

    def __init__(self, root, split='train', transform=False):
        super(VOC2012ClassSeg, self).__init__(
            root, split=split, transform=transform)


class SBDClassSeg(VOCClassSegBase):

    # XXX: It must be renamed to benchmark.tar to be extracted.
    url = 'http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/semantic_contours/benchmark.tgz'  # NOQA

    def __init__(self, root, split='train', transform=False):
        self.root = root
        self.split = split

did = did.strip()
            img_file = osp.join(dataset_dir, 'JPEGImages/%s.jpg' % did)
            lbl_file = osp.join(dataset_dir, 'SegmentationClass/%s.png' % did)
            self.files['seg11valid'].append({'img': img_file, 'lbl': lbl_file})


class VOC2012ClassSeg(VOCClassSegBase):

    url = 'http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar'  # NOQA

    def __init__(self, root, split='train', transform=False):
        super(VOC2012ClassSeg, self).__init__(
            root, split=split, transform=transform)


class SBDClassSeg(VOCClassSegBase):

    # XXX: It must be renamed to benchmark.tar to be extracted.
    url = 'http://www.eecs.berkeley.edu/Research/Projects/CS/vision/grouping/semantic_contours/benchmark.tgz'  # NOQA

    def __init__(self, root, split='train', transform=False):
        self.root = root
        self.split = split
        self._transform = transform

        dataset_dir = osp.join(self.root, 'VOC/benchmark_RELEASE/dataset')
        self.files = collections.defaultdict(list)
        for split in ['train', 'val']:
            imgsets_file = osp.join(dataset_dir, '%s.txt' % split)
            for did in open(imgsets_file):
                did = did.strip()
                img_file = osp.join(dataset_dir, 'img/%s.jpg' % did)

img = img.transpose(2, 0, 1)
        img = torch.from_numpy(img).float()
        lbl = torch.from_numpy(lbl).long()
        return img, lbl

    def untransform(self, img, lbl):
        img = img.numpy()
        img = img.transpose(1, 2, 0)
        img += self.mean_bgr
        img = img.astype(np.uint8)
        img = img[:, :, ::-1]
        lbl = lbl.numpy()
        return img, lbl


class VOC2011ClassSeg(VOCClassSegBase):

    def __init__(self, root, split='train', transform=False):
        super(VOC2011ClassSeg, self).__init__(
            root, split=split, transform=transform)
        pkg_root = osp.join(osp.dirname(osp.realpath(__file__)), '..')
        imgsets_file = osp.join(
            pkg_root, 'ext/fcn.berkeleyvision.org',
            'data/pascal/seg11valid.txt')
        dataset_dir = osp.join(self.root, 'VOC/VOCdevkit/VOC2012')
        for did in open(imgsets_file):
            did = did.strip()
            img_file = osp.join(dataset_dir, 'JPEGImages/%s.jpg' % did)
            lbl_file = osp.join(dataset_dir, 'SegmentationClass/%s.png' % did)
            self.files['seg11valid'].append({'img': img_file, 'lbl': lbl_file})