Crate fastembed

FastEmbed - Fast, light, accurate library built for retrieval embedding generation.

The library provides the TextEmbedding struct to interface with text embedding models.

Instantiating TextEmbedding

use fastembed::{TextEmbedding, InitOptions, EmbeddingModel};

// With default InitOptions
let model = TextEmbedding::try_new(Default::default())?;

// List all supported models
dbg!(TextEmbedding::list_supported_models());

// With custom InitOptions
let model = TextEmbedding::try_new(
       InitOptions::new(EmbeddingModel::AllMiniLML6V2).with_show_download_progress(true),
)?;

Find more info about the available options in the InitOptions documentation.

Embeddings generation

 let documents = vec![
    "passage: Hello, World!",
    "query: Hello, World!",
    "passage: This is an example passage.",
    // You can leave out the prefix but it's recommended
    "fastembed-rs is licensed under MIT"
    ];

 // Generate embeddings with the default batch size, 256
 let embeddings = model.embed(documents, None)?;

 println!("Embeddings length: {}", embeddings.len()); // -> Embeddings length: 4

Re-exports

• pub use crate::output::EmbeddingOutput;
• pub use crate::output::OutputKey;
• pub use crate::output::OutputPrecedence;
• pub use crate::output::SingleBatchOutput;

Modules

• output
  Utilities to help with the embeddings output.

Structs

• ExecutionProviderDispatch
  Dynamic execution provider container, used to provide a list of multiple types of execution providers when configuring execution providers for a SessionBuilder or EnvironmentBuilder.
• ImageEmbedding
  Rust representation of the ImageEmbedding model
• ImageInitOptions
  Options for initializing the ImageEmbedding model
• ImageInitOptionsUserDefined
  Options for initializing UserDefinedImageEmbeddingModel
• InitOptions
  Options for initializing the TextEmbedding model
• InitOptionsUserDefined
  Options for initializing UserDefinedEmbeddingModel
• ModelInfo
  Data struct about the available models
• RerankInitOptions
  Options for initializing the reranking model
• RerankInitOptionsUserDefined
  Options for initializing UserDefinedRerankerModel
• RerankResult
  Rerank result.
• RerankerModelInfo
  Data struct about the available reanker models
• SparseEmbedding
• SparseInitOptions
  Options for initializing the SparseTextEmbedding model
• SparseTextEmbedding
  Rust representation of the SparseTextEmbedding model
• TextEmbedding
  Rust representation of the TextEmbedding model
• TextRerank
• TokenizerFiles
• UserDefinedEmbeddingModel
  Struct for “bring your own” embedding models
• UserDefinedImageEmbeddingModel
  Struct for “bring your own” embedding models
• UserDefinedRerankingModel
  Struct for “bring your own” reranking models
• UserDefinedSparseModel
  Struct for “bring your own” embedding models

Enums

• EmbeddingModel
• ImageEmbeddingModel
• OnnxSource
  Enum for the source of the onnx file
• Pooling
• QuantizationMode
  Enum for quantization mode.
• RerankerModel

Constants

• DEFAULT_CACHE_DIR

Functions

• read_file_to_bytes
  Public function to read a file to bytes. To be used when loading local model files.

Type Aliases

• Embedding
  Type alias for the embedding vector
• Error
  Type alias for the error type