RecursiveAI.OneNlp.Models.Embeddings
1.0.0
dotnet add package RecursiveAI.OneNlp.Models.Embeddings --version 1.0.0
NuGet\Install-Package RecursiveAI.OneNlp.Models.Embeddings -Version 1.0.0
<PackageReference Include="RecursiveAI.OneNlp.Models.Embeddings" Version="1.0.0" />
<PackageVersion Include="RecursiveAI.OneNlp.Models.Embeddings" Version="1.0.0" />
<PackageReference Include="RecursiveAI.OneNlp.Models.Embeddings" />
paket add RecursiveAI.OneNlp.Models.Embeddings --version 1.0.0
#r "nuget: RecursiveAI.OneNlp.Models.Embeddings, 1.0.0"
#:package RecursiveAI.OneNlp.Models.Embeddings@1.0.0
#addin nuget:?package=RecursiveAI.OneNlp.Models.Embeddings&version=1.0.0
#tool nuget:?package=RecursiveAI.OneNlp.Models.Embeddings&version=1.0.0
RecursiveAI.OneNLP.Models.Embeddings
spaCy-class NLP for .NET. The default English sentence embedder for RecursiveAI.OneNlp: add this package and Nlp.Embed turns text into a 384-value, unit-length vector. No downloads, no native code, NativeAOT and trim clean. Unaffiliated with other projects named OneNLP.
Measured. Vectors match sentence-transformers 6.2.0 to a maximum absolute difference of 4e-7 over 2,103 texts, with no token-ID mismatches, and mteb 2.27.1 scores equal the same-harness sentence-transformers scores on every task: STSBenchmark 85.86 (model card 85.86), TwentyNewsgroupsClustering 48.63 (48.52), TwitterSemEval2015 73.24 (73.33), AskUbuntuDupQuestions 62.58 (62.59), SciFact 71.27 (71.28). Banking77Classification is 81.75 against the model card's 85.74: sentence-transformers itself misses the card by the same amount on that task, so the gap is in the card's number, not in the runtime. Size: 127.6 MiB of weights. Lineage: the authors' weights (MIT); nothing is trained here.
using RecursiveAI.OneNlp;
float[] a = Nlp.Embed("how do I reset my password");
float[] b = Nlp.Embed("I forgot my login credentials");
float cosine = a.Zip(b, (x, y) => x * y).Sum(); // vectors have unit length
The model is BAAI bge-small-en-v1.5 (MIT), exported to ONNX from the authors' published weights and run on OneModel; nothing is trained here. Vectors match sentence-transformers to float tolerance (docs/embeddings.md in the source repository has the parity and MTEB results). PROVENANCE.md states the licence and the authors' training data.
- Measured results against spaCy and others: https://github.com/recursive-ai-oss/onenlp-csharp/blob/main/docs/results-table.md
- Source and documentation: https://github.com/recursive-ai-oss/onenlp-csharp
Copyright (c) Recursive AI, LLC.
| Product | Versions Compatible and additional computed target framework versions. |
|---|---|
| .NET | net10.0 is compatible. net10.0-android was computed. net10.0-browser was computed. net10.0-ios was computed. net10.0-maccatalyst was computed. net10.0-macos was computed. net10.0-tvos was computed. net10.0-windows was computed. |
-
net10.0
- RecursiveAI.OneNlp.Neural (>= 1.0.0)
NuGet packages
This package is not used by any NuGet packages.
GitHub repositories
This package is not used by any popular GitHub repositories.
| Version | Downloads | Last Updated |
|---|---|---|
| 1.0.0 | 0 | 11/8/2026 |
See CHANGELOG.md in the repository.