← Catalog
sentence-transformers/all-MiniLM-L6-v2 Add to compare Open comparison →
Primitive: /encode · Encode ·
BERT
This is a sentence-transformers model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search.
Dense
View on Hugging Face → Fine-tuned from nreimers/MiniLM-L6-H384-uncased
Overview
Hardware:
L4 RTX-PRO-6000 — drives latency, throughput & cost
Size 23M params Tasks /encode License apache-2.0 Languages en Latency 53 ms Throughput 55.3K tok/s Cost $0.0040 /1M tok
Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.
Embedding Output types Dense Dimensions dense: 384 Max sequence length 256 Inputs text
Benchmarks Duplicate question retrieval from StackExchange Physics
Corpus: 38,314 Queries: 1,039
Quality
ndcg at 10 0.4698
map at 10 0.4074
mrr at 10 0.4632
Performance A10G b1 c16
Corpus 1.7K tok/s
Corpus p50 1.1s
Query 512 tok/s
Query p50 327.0ms
Performance L4 b1 c16
Corpus 37.0K tok/s
Corpus p50 53.3ms
Query 2.8K tok/s
Query p50 49.8ms
Reference →
Code search with natural language queries
Corpus: 6,267 Queries: 500
Quality
ndcg at 10 0.3361
map at 10 0.2667
mrr at 10 0.2843
Performance A10G b1 c16
Corpus 1.1K tok/s
Corpus p50 750.0ms
Query 299 tok/s
Query p50 296.9ms
Performance L4 b1 c16
Corpus 17.1K tok/s
Corpus p50 50.1ms
Query 1.8K tok/s
Query p50 45.5ms
Reference →
Financial opinion mining and question answering
Corpus: 57,599 Queries: 648
Quality
ndcg at 10 0.1503
map at 10 0.1084
mrr at 10 0.1771
Performance A10G b1 c16
Corpus 1.9K tok/s
Corpus p50 1.3s
Query 587 tok/s
Query p50 345.4ms
Performance L4 b1 c16
Corpus 46.3K tok/s
Corpus p50 52.4ms
Query 3.7K tok/s
Query p50 43.8ms
Reference →
Question answering on consumer contracts
Corpus: 153 Queries: 396
Quality
ndcg at 10 0.6561
map at 10 0.5883
mrr at 10 0.5883
Performance L4 b1 c16
Corpus 128.8K tok/s
Corpus p50 58.9ms
Query 4.3K tok/s
Query p50 59.1ms
Reference →
Biomedical literature search from NutritionFacts.org
Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.2324
map at 10 0.0703
mrr at 10 0.3935
Performance L4 b1 c16
Corpus 130.0K tok/s
Corpus p50 20.0ms
Query 2.6K tok/s
Query p50 17.1ms
Reference →
Smaller subset of the FiQA financial QA dataset
Quality
ndcg at 10 0.4774
map at 10 0.3931
mrr at 10 0.5476
Performance L4 b1 c16
Corpus 44.2K tok/s
Corpus p50 56.1ms
Query 2.8K tok/s
Query p50 49.9ms
Reference →
Citation prediction, document classification, and recommendation for scientific papers
Corpus: 25,656 Queries: 1,000
Quality
ndcg at 10 0.0648
map at 10 0.0363
mrr at 10 0.1075
Performance L4 b1 c16
Corpus 55.3K tok/s
Corpus p50 51.8ms
Query 4.1K tok/s
Query p50 43.1ms
Reference →
Scientific claim verification using research literature
Corpus: 5,183 Queries: 300
Quality
ndcg at 10 0.6112
map at 10 0.5565
mrr at 10 0.5620
Performance L4 b1 c16
Corpus 75.7K tok/s
Corpus p50 52.8ms
Query 5.8K tok/s
Query p50 44.5ms
Reference →
Programming question answering from Stack Overflow
Corpus: 19,931 Queries: 1,994
Quality
ndcg at 10 0.8396
map at 10 0.8117
mrr at 10 0.8117
Performance L4 b1 c16
Corpus 58.9K tok/s
Corpus p50 56.0ms
Query 65.7K tok/s
Query p50 61.5ms
Reference →
