opensearch-project/opensearch-neural-sparse-encoding-doc-v3-distill
The model should be selected considering search relevance, model inference and retrieval efficiency(FLOPS). We benchmark models' performance on a subset of BEIR benchmark: TrecCovid,NFCorpus,NQ,HotpotQA,FiQA,ArguAna,Touche,DBPedia,SCIDOCS,FEVER,Climate FEVER,SciFact,Quora.
Benchmarks
CQADupstackPhysicsRetrieval
Duplicate question retrieval from StackExchange Physics
Corpus: 38,314 Queries: 1,039
Performance L4 b1 c16
Corpus TPS 33.5K
Corpus p50 58.4ms
Query TPS 4.1K
Query p50 41.1ms
CosQA
Code search with natural language queries
Corpus: 6,267 Queries: 500
Performance L4 b1 c16
Corpus TPS 15.6K
Corpus p50 51.5ms
Query TPS 1.9K
Query p50 47.9ms
FiQA2018
Financial opinion mining and question answering
Corpus: 57,599 Queries: 648
Performance L4 b1 c16
Corpus TPS 40.5K
Corpus p50 59.2ms
Query TPS 4.6K
Query p50 39.5ms
LegalBenchConsumerContractsQA
Question answering on consumer contracts
Corpus: 153 Queries: 396
Performance L4 b1 c16
Corpus TPS 97.4K
Corpus p50 76.3ms
Query TPS 6.3K
Query p50 40.7ms
NFCorpus
Biomedical literature search from NutritionFacts.org
Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.3399
map at 10 0.1314
mrr at 10 0.5457
Performance L4 b1 c16
Corpus TPS 71.1K
Corpus p50 65.9ms
Query TPS 1.9K
Query p50 41.1ms
SCIDOCS
Citation prediction, document classification, and recommendation for scientific papers
Corpus: 25,656 Queries: 1,000
Performance L4 b1 c16
Corpus TPS 47.1K
Corpus p50 59.3ms
Query TPS 4.4K
Query p50 39.9ms
SciFact
Scientific claim verification using research literature
Corpus: 5,183 Queries: 300
Performance L4 b1 c16
Corpus TPS 59.3K
Corpus p50 67.2ms
Query TPS 6.1K
Query p50 42.4ms
StackOverflowQA
Programming question answering from Stack Overflow
Corpus: 19,931 Queries: 1,994
Performance L4 b1 c16
Corpus TPS 53.1K
Corpus p50 62.1ms
Query TPS 97.0K
Query p50 46.1ms