Why did we open-source our inference engine? Read the post

← Catalog

naver/splade-cocondenser-selfdistil

Open comparison →

Primitive: /encode · Encode · BERT

SPLADE model for passage retrieval. For additional details, please visit: * paper: https://arxiv.org/abs/2205.04733 * code: https://github.com/naver/splade

Sparse

Overview

Hardware: — drives latency, throughput & cost

Size110M params
Tasks /encode
Licensecc-by-nc-sa-4.0
Languagesen
Latency72 ms
Throughput40.0K tok/s
Cost$0.0056 /1M tok

Cost is approximate — computed from list GPU prices; your actual price depends on the provider you deploy SIE with.

Embedding

Output typesSparse
Dimensionssparse: 30,522
Max sequence length512
Inputstext

Benchmarks

CQADupstackPhysicsRetrieval

scientific retrieval en

Duplicate question retrieval from StackExchange Physics

Corpus: 38,314 Queries: 1,039
Performance L4 b1 c16
Corpus 32.8K tok/s
Corpus p50 60.2ms
Query 3.0K tok/s
Query p50 56.3ms
Reference →

CosQA

technology retrieval en

Code search with natural language queries

Corpus: 6,267 Queries: 500
Performance L4 b1 c16
Corpus 14.9K tok/s
Corpus p50 53.6ms
Query 1.6K tok/s
Query p50 55.0ms
Reference →

FiQA2018

finance retrieval en

Financial opinion mining and question answering

Corpus: 57,599 Queries: 648
Performance L4 b1 c16
Corpus 34.6K tok/s
Corpus p50 69.2ms
Query 3.2K tok/s
Query p50 56.2ms
Reference →

LegalBenchConsumerContractsQA

legal retrieval en

Question answering on consumer contracts

Corpus: 153 Queries: 396
Performance L4 b1 c16
Corpus 84.4K tok/s
Corpus p50 88.5ms
Query 5.2K tok/s
Query p50 49.9ms
Reference →

NFCorpus

medical retrieval en

Biomedical literature search from NutritionFacts.org

Corpus: 3,593 Queries: 323
Quality
ndcg at 10 0.3403
map at 10 0.1282
mrr at 10 0.5447
Performance L4 b1 c16
Corpus 55.9K tok/s
Corpus p50 79.2ms
Query 1.5K tok/s
Query p50 52.1ms
Reference →

SCIDOCS

scientific retrieval en

Citation prediction, document classification, and recommendation for scientific papers

Corpus: 25,656 Queries: 1,000
Performance L4 b1 c16
Corpus 42.1K tok/s
Corpus p50 68.2ms
Query 3.1K tok/s
Query p50 55.1ms
Reference →

SciFact

scientific retrieval en

Scientific claim verification using research literature

Corpus: 5,183 Queries: 300
Performance L4 b1 c16
Corpus 53.5K tok/s
Corpus p50 75.5ms
Query 4.8K tok/s
Query p50 54.3ms
Reference →

StackOverflowQA

technology retrieval en

Programming question answering from Stack Overflow

Corpus: 19,931 Queries: 1,994
Performance L4 b1 c16
Corpus 37.9K tok/s
Corpus p50 76.9ms
Query 60.7K tok/s
Query p50 76.7ms
Reference →

Open source inference for agents

Open-source inference for the models behind your agents. Run it yourself, or let us run it for you.

Github 2.1K

Contact us

Tell us about your use case and we'll get back to you shortly.

Apply for an inference grant

Free capacity on our hosted cluster for selected projects. Tell us what you run and we reply by email.