0 accuracy when i run Mask RCNN with LabelMe software

[SaraBallkoci]

Hi guys, can someone please help me with my thesis, I have labeled my images with LabelMe software. I have used 80 images for train and 20 for val. I train the model with no errors, but i get 0 accuracy when i test it afterwards. Can someone please give me any ideas how to fix this, or please if you have worked with LabelMe can you share the code with me, I would be very thankful. Thank you!

class CustomDataset(utils.Dataset):

def load_custom(self, dataset_dir, subset):
    # Add the class
    self.add_class("object", 1, "tooth")

    # Make sure the subset is either 'train' or 'val'
    assert subset in ["train", "val"]
    dataset_dir = os.path.join(dataset_dir, subset)

    # Iterate over all JSON files in the directory
    for filename in os.listdir(dataset_dir):
        if filename.endswith(".json"):
            json_file = os.path.join(dataset_dir, filename)
            # print(f"Processing {json_file}...")  # Print the current JSON file being processed

            annotations_json = json.load(open(json_file))
            shapes = annotations_json.get('shapes', [])

            print("shapes:", len(shapes))

            # If there are no annotations in the file, skip it
            if not shapes:
                continue

            # Extract image filename and load the image
            image_filename = annotations_json['imagePath'].split('\\')[-1]
            image_path = os.path.join(dataset_dir, image_filename)
            # print(f"Loading image {image_path}...")  # Print the image file being loaded

            num_ids = [1] * len(shapes)
            
            print("numids",num_ids)
            
            image = skimage.io.imread(image_path)
            height, width = image.shape[:2]

            # Add the image to the dataset
            self.add_image(
                "object",
                image_id=image_filename,
                path=image_path,
                width=width, height=height,
                polygons=[{
                    'name': 'polygon',
                    'all_points_x': [point[0] for point in shape['points']],
                    'all_points_y': [point[1] for point in shape['points']]
                } for shape in shapes],
                num_ids=num_ids
            )

def load_mask(self, image_id):
    """Generate instance masks for an image.
    Returns:
    masks: A bool array of shape [height, width, instance count] with
        one mask per instance.
    class_ids: a 1D array of class IDs of the instance masks.
    """
    # If not an object dataset image, delegate to parent class.
    image_info = self.image_info[image_id]
    if image_info["source"] != "object":
        return super(self.__class__, self).load_mask(image_id)

    # Initialize the mask array
    mask = np.zeros([image_info["height"], image_info["width"], len(image_info["polygons"])],
                    dtype=np.uint8)

    for i, p in enumerate(image_info["polygons"]):
        # Clipping the coordinates to ensure they are within the image dimensions
        all_points_x = np.clip(p['all_points_x'], 0, image_info["width"] - 1)
        all_points_y = np.clip(p['all_points_y'], 0, image_info["height"] - 1)

        # Get indexes of pixels inside the polygon and set them to 1
        rr, cc = skimage.draw.polygon(all_points_y, all_points_x)
        mask[rr, cc, i] = 1

    # Return mask, and array of class IDs of each instance. Since we have
    # one class ID only, we return an array of 1s
    return mask.astype(np.bool), np.ones([mask.shape[-1]], dtype=np.int32)
def image_reference(self, image_id):
    """Return the path of the image."""
    info = self.image_info[image_id]
    if info["source"] == "object":
        return info["path"]
    else:
        super(self.__class__, self).image_reference(image_id)