AllMpnetBaseV2Sharp 0.1.3

AllMpnetBaseV2Sharp

C# implementation of Sentence Transformers all-mpnet-base-v2.

Built as a modern .NET 8 class library using ONNX Runtime and HuggingFace tokenizers for correctness and parity with the original Python model.

📦 NuGet

AllMpnetBaseV2Sharp

❌ The NuGet package does not include the ONNX model or tokenizer.json. You will need to download them manually from all-mpnet-base-v2 on Hugging Face.

By default, the embedder looks for these files in the ./model folder: ./model/model.onnx ./model/tokenizer.json

You can also specify custom file locations by setting the paths in MpnetOptions.

🚀 How to Use

Single sentence

using AllMpnetBaseV2Sharp;

var sentence = "This is an example sentence";

using var embedder = new AllMpnetBaseV2Embedder();
var embedding = embedder.Encode(sentence);

Multiple sentences

using AllMpnetBaseV2Sharp;

string[] sentences =
{
 "This is an example sentence",
 "Here is another"
};

using var embedder = new AllMpnetBaseV2Embedder();
var embeddings = embedder.Encode(sentences);

Custom ONNX model (advanced)

var options = new MpnetOptions
{
 ModelPath = "path/to/model.onnx",
 TokenizerPath = "path/to/tokenizer.json"
};

using var embedder = new AllMpnetBaseV2Embedder(options);
var embedding = embedder.Encode("This is an example sentence");