ktrapeznikov/biobert_v1.1_pubmed_squad_v2

The model was fine-tuned on SQuAD V2 using run_squad.py . It was trained on 4 NVIDIA GeForce GeForce GTX 2080 Ti 11Gb . The model scores if a paragraph contains an answer: start_scores.softmax(dim=1).log()[:,,.None] + end_score = span_score[:,0,0] #no answer scores. I did not sweep for best threshold. The model is cased. It is now running on the dev set. It will be used to test the accuracy of the model’s ability to predict an answer from a given paragraph. It has been used in the past for,