Layers
#1
by ryjovsky - opened
Hi,
Have you been adding another layers at the head of BERT or it's just fine-tuned BertForQuestionAnswering model.
No it's just the regular BertForQuestionAnswering model. I used the fine-tuning script provided in the transformers repo (https://wandb.ai/salti/mBERT_QA/runs/wkqzhrp2/code).
salti changed discussion status to closed