EM scores

llevkova June 6, 2023, 2:46pm 1

This model has much higher F1/EM scores than what is in its card when evaluated on the validation squad2 data. Any ideas why that is?

Topic	Replies	Views
Deepset's bert cased trained on squad2 question Models	145	June 7, 2023
Distribution shift in SQuADv1? 🤗Datasets	261	July 20, 2022
How to reproduce the performance of bert-large-uncased-whole-word-masking-finetuned-squad? Intermediate	300	July 25, 2021
Deepset/roberta-base-squad2 showing overloaded Models	232	May 15, 2023
Dataset confusion for distilroberta trained on squad2 Beginners	206	November 17, 2021