关于WoBERT+模型无法加载

[WENGSYX]

config_path = 'WoBERT/bert_config.json'
checkpoint_path = 'WoBERT/bert_model.ckpt'
dict_path = 'WoBERT/vocab.txt'
bert = bert4keras.models.build_transformer_model(
config_path=config_path,
checkpoint_path=checkpoint_path,
with_pool=True,
hierarchical_position=True,
return_keras_model=False,
)
当想加载WoBERT+模型，却会报错Key bert/pooler/dense/kernel not found in checkpoint，能帮帮我吗
Traceback (most recent call last):
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\tensorflow\python\training\py_checkpoint_reader.py", line 70, in get_tensor
self, compat.as_bytes(tensor_str))
RuntimeError: Key bert/pooler/dense/kernel not found in checkpoint
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\IPython\core\interactiveshell.py", line 3343, in run_code
exec(code_obj, self.user_global_ns, self.user_ns)
File "", line 6, in
return_keras_model=False,
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\bert4keras\models.py", line 2338, in build_transformer_model
transformer.load_weights_from_checkpoint(checkpoint_path)
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\bert4keras\models.py", line 297, in load_weights_from_checkpoint
raise e
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\bert4keras\models.py", line 291, in load_weights_from_checkpoint
values.append(self.load_variable(checkpoint, v))
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\bert4keras\models.py", line 691, in load_variable
variable = super(BERT, self).load_variable(checkpoint, name)
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\bert4keras\models.py", line 262, in load_variable
return tf.train.load_variable(checkpoint, name)
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\tensorflow\python\training\checkpoint_utils.py", line 85, in load_variable
return reader.get_tensor(name)
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\tensorflow\python\training\py_checkpoint_reader.py", line 74, in get_tensor
error_translator(e)
File "C:\Users\14301\miniconda3\envs\gluon\lib\site-packages\tensorflow\python\training\py_checkpoint_reader.py", line 35, in error_translator
raise errors_impl.NotFoundError(None, None, error_message)
tensorflow.python.framework.errors_impl.NotFoundError: Key bert/pooler/dense/kernel not found in checkpoint
不过当转换为pytorch版时，却能够正常加载使用。

[bojone]

因为WoBERT没有NSP引入，因而没有pooler那部分的权重。

有两个选择：1、模仿这里直接取出cls用；2、如果是最新版本的bert4keras，可以在build_transformer_model时传入参数ignore_invalid_weights=True