请问使用simbert生成的时候报错是什么原因

[YYYYYYHP]

报错如下：
2021-07-20 20:21:22.706441: W tensorflow/core/common_runtime/base_collective_executor.cc:217] BaseCollectiveExecutor::StartAbort Invalid argument: Incompatible shapes: [1,512,128] vs. [2,513,128]
[[{{node Embedding-Position/add}}]]
Traceback (most recent call last):
File "/Users/yanghp/PycharmProjects/Data_Augmentation/classification.py", line 45, in
aug_data = units.simbert_augment(train_data,create_num=3b)
File "/Users/yanghp/PycharmProjects/Data_Augmentation/units.py", line 32, in simbert_augment
aug_sentences = simbert.replace(sent=line[0], create_num=create_num)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/nlpcda/tools/Simbert.py", line 28, in replace
synonyms = self.model.gen_synonyms(text=sent, n=n, k=create_num)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/nlpcda/tools/simbert/generator.py", line 64, in gen_synonyms
r = self.generate(text, n)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/nlpcda/tools/simbert/generator.py", line 57, in generate
output_ids = self.random_sample([token_ids, segment_ids], n, topk)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/bert4keras/snippets.py", line 454, in random_sample
probas = self.predict(inputs, output_ids, step, 'probas') # 计算当前概率
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/bert4keras/snippets.py", line 386, in new_predict
result = predict(self, inputs, output_ids, step)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/nlpcda/tools/simbert/generator.py", line 52, in predict
return self.seq2seq.predict([token_ids, segment_ids])[:, -1]
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/keras/engine/training.py", line 1462, in predict
callbacks=callbacks)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/keras/engine/training_arrays.py", line 324, in predict_loop
batch_outs = f(ins_batch)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/keras/backend.py", line 3727, in call
outputs = self._graph_fn(*converted_inputs)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1551, in call
return self._call_impl(args, kwargs)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1591, in _call_impl
return self._call_flat(args, self.captured_inputs, cancellation_manager)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 1692, in _call_flat
ctx, args, cancellation_manager=cancellation_manager))
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/eager/function.py", line 545, in call
ctx=ctx)
File "/Users/yanghp/opt/anaconda3/envs/py36/lib/python3.6/site-packages/tensorflow_core/python/eager/execute.py", line 67, in quick_execute
six.raise_from(core._status_to_exception(e.code, message), None)
File "", line 3, in raise_from
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [1,512,128] vs. [2,513,128]
[[node Embedding-Position/add (defined at /opt/anaconda3/envs/py36/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3009) ]] [Op:__inference_keras_scratch_graph_4612]

Function call stack:
keras_scratch_graph

更改较小的create_num会减小报错的概率，但还是会出现。请问怎么解决？

[425776024]

create_num=3b ?