Hello,
thanks for sharing your code.
I am new to BERT and finetuning and I tried your code with costumn data.
I recieved following error on running DocumentBertMaxPool/DocumentBertLinear:
_` batch_predictions = self.bert_doc_classification(batch_document_tensors,
185 batch_document_sequence_lengths,
--> 186 freeze_bert=self.args['freeze_bert'], device=self.args['device'])
187
188 batch_correct_output = correct_output[i:i + self.args['batch_size']].to(device=self.args['device'])
/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py in call(self, *input, **kwargs)
530 result = self._slow_forward(*input, **kwargs)
531 else:
--> 532 result = self.forward(*input, **kwargs)
533 for hook in self._forward_hooks.values():
534 hook_result = hook(self, input, result)
TypeError: forward() got an unexpected keyword argument 'device'_
`
I think the reason for this problem is that there is missing the 'freeze_bert'-argument in the forward-method of the model architecture (document_bert_architectures.py).
Thanks for response.
Hello,
thanks for sharing your code.
I am new to BERT and finetuning and I tried your code with costumn data.
I recieved following error on running DocumentBertMaxPool/DocumentBertLinear:
_` batch_predictions = self.bert_doc_classification(batch_document_tensors,
185 batch_document_sequence_lengths,
--> 186 freeze_bert=self.args['freeze_bert'], device=self.args['device'])
187
188 batch_correct_output = correct_output[i:i + self.args['batch_size']].to(device=self.args['device'])
/usr/local/lib/python3.7/dist-packages/torch/nn/modules/module.py in call(self, *input, **kwargs)
530 result = self._slow_forward(*input, **kwargs)
531 else:
--> 532 result = self.forward(*input, **kwargs)
533 for hook in self._forward_hooks.values():
534 hook_result = hook(self, input, result)
TypeError: forward() got an unexpected keyword argument 'device'_
`
I think the reason for this problem is that there is missing the 'freeze_bert'-argument in the forward-method of the model architecture (document_bert_architectures.py).
Thanks for response.