mrm8488/bert-small-finetuned-squadv2

BERT-Small created by Google Research and fine-tuned on SQuAD 2.0 for Q&A downstream task . Model is a part of 24 smaller BERT models referenced in Well-Read Students Learn Better: On the Importance of Pre-training Compact Models . The model was trained on a Tesla P100 GPU and 25GB of RAM . The script for fine tuning can be found here . It works with pipelines: "Question-answering" and "Hugginface/transformers" The model is trained with WordPiece masking . It can be used to answer questions when possible, but also determine when no answer is supported by the paragraph and abst,