phanxuanquang/EmbeddingGemma.NET
🏗️ Frameworkphanxuanquang
A lightweight .NET library for integrating Google's EmbeddingGemma-300m model using ONNX Runtime.
EmbeddingGemma.NET bridges the gap between Google's advanced embedding models and the .NET ecosystem. The library is built on top of Microsoft.ML.OnnxRuntime, ensuring efficient execution of the EmbeddingGemma-300m model on various hardware configurations. It simplifies the complex process of model loading, input tokenization, and output vector processing into a developer-friendly API. Key features include support for high-dimensional vector generation suitable for semantic search, compatibility with Semantic Kernel for RAG workflows, and optimized performance for local inference. By abstracting the underlying ONNX graph, it allows C# developers to focus on building intelligent features like document clustering, recommendation engines, and context-aware chatbots without the overhead of managing external AI services or complex model conversion pipelines.
💡Highlights
- ├─Powered by ONNX Runtime
- ├─Optimized for .NET ecosystem
- └─Enables local RAG workflows
🎯For
- ├─C# Developers
- └─AI Engineers