Ошибка TypeError: train() got an unexpected keyword argument 'class attribute name'

[romanjkeey]

Здравствуйте. Вроде следую по инструкции. Но на строке
nnet.train(class_attribute_name="class") #nnet.loadModel(MASK_RCNN_MODEL_PATH)

Выходит ошибка TypeError: train() got an unexpected keyword argument 'class attribute name'

`import os
import cv2
import numpy as np
import sys
import json
import matplotlib.image as mpimg
from matplotlib import pyplot as plt
import warnings
warnings.filterwarnings('ignore')

NOMEROFF_NET_DIR = os.path.abspath('C:/Users/ZhuravlevRS/work/datasets/ocr/nomeroff-net-master')
sys.path.append(NOMEROFF_NET_DIR)

specify the path to Mask_RCNN if you placed it outside Nomeroff-net project

MASK_RCNN_DIR = os.path.join(NOMEROFF_NET_DIR, 'Mask_RCNN')
MASK_RCNN_LOG_DIR = os.path.join(NOMEROFF_NET_DIR, "logs")

DATASET_NAME = "mrcnn"
VERSION = "2019_09_18"
MASK_RCNN_FROZEN_PATH = os.path.join(NOMEROFF_NET_DIR, "models/", 'numberplate_{}_{}.pb'.format(DATASET_NAME, VERSION))`

`# Import license plate recognition tools.
from NomeroffNet import Detector
from NomeroffNet.Base import convert_keras_to_freeze_pb

CONFIG = {
"GPU_COUNT": 1,
"IMAGES_PER_GPU": 1,
"WEIGHTS": "coco",
"EPOCHS": 100,
"CLASS_NAMES": ["BG", "numberplate"],
"NAME": "numberplate",
"DATASET_DIR": "C:/Users/ZhuravlevRS/work/datasets/ocr/mrcnn",
"LAYERS": "all",
"NUM_CLASSES": 2
}

Initialize npdetector with default configuration file.

nnet = Detector(MASK_RCNN_DIR, MASK_RCNN_LOG_DIR, CONFIG)`

[dimabendera]

try use

nnet.train() 

instanse

nnet.train(class_attribute_name="class")