RuntimeError: The size of tensor a (108) must match the size of tensor b (212) at non-singleton dimension 0

[demongolem-biz]

Out of the box, I try to run run_dee_task.py and I get the above error. The only thing I changed was when I got this:

Traceback (most recent call last): File "run_dee_task.py", line 10, in from dee.dee_task import DEETask, DEETaskSetting File "/home/gregory.werner/Doc2EDAG/dee/dee_task.py", line 11, in from .dee_helper import logger, DEEExample, DEEExampleLoader, DEEFeatureConverter,
File "/home/gregory.werner/Doc2EDAG/dee/dee_helper.py", line 14, in from .ner_task import NERExample, NERFeatureConverter File "/home/gregory.werner/Doc2EDAG/dee/ner_task.py", line 14, in from .ner_model import BertForBasicNER, judge_ner_prediction File "/home/gregory.werner/Doc2EDAG/dee/ner_model.py", line 15, in class BertForBasicNER(PreTrainedBertModel): NameError: name 'PreTrainedBertModel' is not defined

I exchange BertPreTrainedModel for PreTrainedBertModel

[WangYao-GoGoGo]

I have the similar problem.

The size of tensor a (11) must match the size of tensor b (22) at non-singleton dimension 0

loss = self.model( doc_batch_dict, features, use_gold_span=use_gold_span, train_flag=True, teacher_prob=teacher_prob )