[Bug]: 只要一进入调试模式就会出现TypeError: __init__() got multiple values for argument 'self'，正常运行则没有

[HelloBroTan]

软件环境

windows10 X64 anaconda-spyder5.2.2
- paddlepaddle-gpu: 2.3.2
- paddlenlp: 2.0.1

重复问题

• I have searched the existing issues

错误描述

只要一进入调试模式就会出现TypeError: __init__() got multiple values for argument 'self'，正常运行则没有

File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddle\nn\layer\transformer.py", line 653, in <listcomp>
    for i in range(num_layers)])

......
649    def __init__(self, encoder_layer, num_layers, norm=None):
650        super(TransformerEncoder, self).__init__()
651        self.layers = LayerList([(encoder_layer if i == 0 else
652                          type(encoder_layer)(**encoder_layer._config))
653                                 for i in range(num_layers)])
......

稳定复现步骤 & 代码

Spyder 调试模式运行
!continue
[2023-03-29 14:28:06,817] [ INFO] - Found C:\Users\lenovo.paddlenlp\models\bert-base-chinese\bert-base-chinese-vocab.txt
[2023-03-29 14:28:06,963] [ INFO] - Already cached C:\Users\lenovo.paddlenlp\models\bert-base-chinese\bert-base-chinese.pdparams
Traceback (most recent call last):
File "d:\pycharm_proj\nlplearning\nlpbase\exam03_bert\chinese_wordsegmentcrf.py", line 282, in
bert = BertForTokenClassification.from_pretrained(model_name_or_path, num_classes=label_num)
File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddlenlp\transformers\model_utils.py", line 231, in from_pretrained
base_model = cls.base_model_class(*base_args, **base_kwargs)
File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddlenlp\transformers\utils.py", line 83, in impl
init_func(self, *args, **kwargs)
File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddlenlp\transformers\bert\modeling.py", line 303, in init
self.encoder = nn.TransformerEncoder(encoder_layer, num_hidden_layers)
File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddle\nn\layer\transformer.py", line 653, in init
for i in range(num_layers)])
File "D:\anaconda\envs\py37nlppad\lib\site-packages\paddle\nn\layer\transformer.py", line 653, in
for i in range(num_layers)])
TypeError: init() got multiple values for argument 'self'

[w5688414]

请升级paddlenlp和paddle试一下