Fine-tuned SQuAD v1.1 weights for lucid.models.bert_large_qa.
The official Google whole-word-masking BERT-Large SQuAD checkpoint.
Attributes
SQUAD_V1WeightEntrySQuAD v1.1 fine-tuned checkpoint (F1 93.2 / EM 86.9), sourced
from
transformers/google-bert/bert-large-uncased-whole-word- masking-finetuned-squad.DEFAULTWeightEntryAlias for
SQUAD_V1.Notes
Reference: Devlin, Chang, Lee, Toutanova, "BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding", NAACL 2019 (arXiv:1810.04805). Fine-tuned on SQuAD v1.1.
Examples
>>> from lucid.models import bert_large_qa
>>> model = bert_large_qa(pretrained=True).eval()