A pre-training method that is designed to better represent and predict spans of text.
SQuAD 1.1 contains 100,000+ question-answer pairs on 500+ articles.
The model was trained on a Tesla P100 GPU and 25GB of RAM.
The script for fine tuning can be found
here1{
2 "exact": 85.49668874172185,
3 "f1": 91.9845699540379,
4 "total": 10570,
5 "HasAns_exact": 85.49668874172185,
6 "HasAns_f1": 91.9845699540379,
7 "HasAns_total": 10570,
8 "best_exact": 85.49668874172185,
9 "best_exact_thresh": 0.0,
10 "best_f1": 91.9845699540379,
11 "best_f1_thresh": 0.0
12}
1from transformers import pipeline
2
3qa_pipeline = pipeline(
4 "question-answering",
5 model="mrm8488/spanbert-finetuned-squadv1",
6 tokenizer="mrm8488/spanbert-finetuned-squadv1"
7)
8
9qa_pipeline({
10 'context': "Manuel Romero has been working hardly in the repository hugginface/transformers lately",
11 'question': "Who has been working hard for hugginface/transformers lately?"
12
13})
