NovaSearch/stella_en_1.5B_v5
We released a Jasper and Stella model technology report and code.(2025.1)
Benchmarks
CQADupstackPhysicsRetrieval
Duplicate question retrieval from StackExchange Physics
Corpus: 38,314 Queries: 1,039
Quality
ndcg at 10 0.5293
map at 10 0.4592
mrr at 10 0.5136
Performance L4 b1 c16
Corpus TPS 11.7K
Corpus p50 180.9ms
Query TPS 2.0K
Query p50 73.7ms
CosQA
Code search with natural language queries
Corpus: 6,267 Queries: 500
Quality
ndcg at 10 0.3239
map at 10 0.2437
mrr at 10 0.2626
Performance L4 b1 c16
Corpus TPS 8.5K
Corpus p50 100.4ms
Query TPS 1.2K
Query p50 71.8ms
FiQA2018
Financial opinion mining and question answering
Corpus: 57,599 Queries: 648
Quality
ndcg at 10 0.6050
map at 10 0.5203
mrr at 10 0.6769
Performance L4 b1 c16
Corpus TPS 13.0K
Corpus p50 222.0ms
Query TPS 2.2K
Query p50 74.6ms
LegalBenchConsumerContractsQA
Question answering on consumer contracts
Corpus: 153 Queries: 396
Quality
ndcg at 10 0.8108
map at 10 0.7641
mrr at 10 0.7653
Performance L4 b1 c16
Corpus TPS 16.7K
Corpus p50 522.2ms
Query TPS 2.7K
Query p50 84.9ms
NFCorpus
Biomedical literature search from NutritionFacts.org
Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.4170
map at 10 0.1630
mrr at 10 0.6201
Performance L4 b1 c16
Corpus TPS 12.8K
Corpus p50 406.0ms
Query TPS 1.0K
Query p50 70.5ms
NanoFiQA2018Retrieval
Smaller subset of the FiQA financial QA dataset
Quality
ndcg at 10 0.6555
map at 10 0.5859
mrr at 10 0.6764
Performance L4 b1 c16
Corpus TPS 1.2K
Corpus p50 1.5s
Query TPS 1.1K
Query p50 117.8ms
SCIDOCS
Citation prediction, document classification, and recommendation for scientific papers
Corpus: 25,656 Queries: 1,000
Quality
ndcg at 10 0.2666
map at 10 0.1649
mrr at 10 0.4293
Performance L4 b1 c16
Corpus TPS 12.5K
Corpus p50 257.9ms
Query TPS 2.0K
Query p50 77.3ms
SciFact
Scientific claim verification using research literature
Corpus: 5,183 Queries: 300
Quality
ndcg at 10 0.8004
map at 10 0.7557
mrr at 10 0.7662
Performance L4 b1 c16
Corpus TPS 12.8K
Corpus p50 356.8ms
Query TPS 3.0K
Query p50 80.0ms
StackOverflowQA
Programming question answering from Stack Overflow
Corpus: 19,931 Queries: 1,994
Quality
ndcg at 10 0.9134
map at 10 0.8970
mrr at 10 0.8970
Performance L4 b1 c16
Corpus TPS 14.2K
Corpus p50 265.9ms
Query TPS 15.6K
Query p50 329.0ms