Module engine 

Core embedding engine for generating vector embeddings from text.

The EmbeddingEngine provides a high-level interface for:

• Loading ONNX INT8 embedding models from HuggingFace Hub
• Generating embeddings for single texts or batches
• Automatic batching and parallel processing via ONNX Runtime

Example

use inference::{EmbeddingEngine, ModelConfig, EmbeddingModel};

#[tokio::main]
async fn main() {
    let config = ModelConfig::new(EmbeddingModel::MiniLM);
    let engine = EmbeddingEngine::new(config).await.unwrap();

    // Embed a single query
    let embedding = engine.embed_query("What is machine learning?").await.unwrap();
    println!("Embedding dimension: {}", embedding.len());

    // Embed multiple documents
    let docs = vec![
        "Machine learning is a subset of AI.".to_string(),
        "Deep learning uses neural networks.".to_string(),
    ];
    let embeddings = engine.embed_documents(&docs).await.unwrap();
    println!("Generated {} embeddings", embeddings.len());
}

Structs

EmbeddingEngine

The main embedding engine for generating vector embeddings.

EmbeddingEngineBuilder

Builder for creating an EmbeddingEngine with fluent API.