TypeError: 'BertTokenizer' object is not callable

[littlelaska]

在将torch代码迁移到MindSpore框架的时候，用到了cybertron替代transformers。
原始代码是：
from transformers import BertTokenizer
model_name_or_path = "dmis-lab/biobert-base-cased-v1.1"
tfm_tokenizer = BertTokenizer.from_pretrained(model_name_or_path)
a = tfm_tokenizer("this is not ok",padding ="max_length", max_length=25, truncation=True, return_tensors="pt")
输出结果如下图：

对应的cybertron代码如下：
import cybertron
model_name_or_path = "dmis-lab/biobert-base-cased-v1.1"
ms_tokenizer = cybertron.BertTokenizer.load(model_name_or_path)
ms_tokenizer("this is not ok", padding="max_length", max_length=self.max_length, truncation=True, return_tensors="pt")
运行报错信息如下图：

看了下cybertron的源代码，tokenizer的方法没有实现完全，想请作者帮忙修复一下这个问题，感谢