Question Answering On Natural Questions Long

评测结果

各个模型在此基准测试上的表现结果

模型名称	EM	Paper Title	Repository
FiE	58.4	0.8% Nyquist computational ghost imaging via non-experimental deep learning	-
DensePhrases	71.9	Learning Dense Representations of Phrases at Scale	-
R2-D2 w HN-DPR	55.9	R2-D2: A Modular Baseline for Open-Domain Question Answering	-
UnitedQA (Hybrid)	54.7	UnitedQA: A Hybrid Approach for Open Domain Question Answering	-
BERTwwm + SQuAD 2	-	Frustratingly Easy Natural Question Answering	-
Cluster-Former (#C=512)	-	Cluster-Former: Clustering-based Sparse Transformer for Long-Range Dependency Encoding	-
DrQA	-	Reading Wikipedia to Answer Open-Domain Questions	-
Locality-Sensitive Hashing	-	Reformer: The Efficient Transformer	-
UniK-QA	54.9	UniK-QA: Unified Representations of Structured and Unstructured Knowledge for Open-Domain Question Answering	-
BERTjoint	-	A BERT Baseline for the Natural Questions	-
Sparse Attention	-	Generating Long Sequences with Sparse Transformers	-
BPR (linear scan; l=1000)	41.6	Efficient Passage Retrieval with Hashing for Open-domain Question Answering	-
DecAtt + DocReader	-	Natural Questions: a Benchmark for Question Answering Research

0 of 13 row(s) selected.