Alibaba-NLP/gte-Qwen2-1.5B-instruct
gte-Qwen2-1.5B-instruct is the latest model in the gte (General Text Embedding) model family. The model is built on Qwen2-1.5B LLM model and use the same training data and strategies as the gte-Qwen2-7B-instruct model.
Benchmarks
CQADupstackPhysicsRetrieval
Duplicate question retrieval from StackExchange Physics
Corpus: 38,314 Queries: 1,039
Performance L4 b1 c16
Corpus TPS 11.6K
Corpus p50 178.5ms
Query TPS 2.2K
Query p50 69.6ms
CosQA
Code search with natural language queries
Corpus: 6,267 Queries: 500
Performance L4 b1 c16
Corpus TPS 9.3K
Corpus p50 96.2ms
Query TPS 1.2K
Query p50 66.8ms
FiQA2018
Financial opinion mining and question answering
Corpus: 57,599 Queries: 648
Performance L4 b1 c16
Corpus TPS 11.8K
Corpus p50 222.9ms
Query TPS 2.1K
Query p50 73.4ms
LegalBenchConsumerContractsQA
Question answering on consumer contracts
Corpus: 153 Queries: 396
Performance L4 b1 c16
Corpus TPS 12.3K
Corpus p50 735.3ms
Query TPS 3.1K
Query p50 71.9ms
NFCorpus
Biomedical literature search from NutritionFacts.org
Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.3925
map at 10 0.1502
mrr at 10 0.6051
Performance L4 b1 c16
Corpus TPS 12.7K
Corpus p50 384.4ms
Query TPS 821
Query p50 90.2ms
NanoFiQA2018Retrieval
Smaller subset of the FiQA financial QA dataset
Quality
ndcg at 10 0.6524
map at 10 0.5848
mrr at 10 0.7032
Performance L4 b1 c16
Corpus TPS 11.3K
Corpus p50 251.5ms
Query TPS 1.9K
Query p50 88.7ms
SCIDOCS
Citation prediction, document classification, and recommendation for scientific papers
Corpus: 25,656 Queries: 1,000
Performance L4 b1 c16
Corpus TPS 12.4K
Corpus p50 261.1ms
Query TPS 2.5K
Query p50 66.4ms
SciFact
Scientific claim verification using research literature
Corpus: 5,183 Queries: 300
Performance L4 b1 c16
Corpus TPS 12.6K
Corpus p50 370.4ms
Query TPS 3.1K
Query p50 74.9ms
StackOverflowQA
Programming question answering from Stack Overflow
Corpus: 19,931 Queries: 1,994
Performance L4 b1 c16
Corpus TPS 12.4K
Corpus p50 299.2ms
Query TPS 11.4K
Query p50 421.4ms