Crate ct2rs

This crate provides Rust bindings for OpenNMT/CTranslate2.

This crate provides the following:

• Rust bindings for ctranslate2::Translator, ctranslate2::Generator, and ctranslate2::Whisper provided by CTranslate2, specifically sys::Translator, sys::Generator, and sys::Whisper.
• More user-friendly versions of these, Translator, Generator, and Whisper (whisper feature is required), which incorporate tokenizers for easier handling.

Basic Usage

The following example translates two strings using default settings and outputs each to the standard output.

use ct2rs::{Config, Translator, TranslationOptions};

let sources = vec![
    "Hallo World!",
    "This crate provides Rust bindings for CTranslate2."
];
let translator = Translator::new("/path/to/model", &Default::default())?;
let results = translator.translate_batch(&sources, &Default::default(), None)?;
for (r, _) in results{
    println!("{}", r);
}

Supported Models

The ct2rs crate has been tested and confirmed to work with the following models:

• BART
• BLOOM
• FALCON
• Marian-MT
• MPT
• NLLB
• GPT-2
• GPT-J
• OPT
• T5
• Whisper

Please see the respective examples for each model.

Stream API

This crate also offers a streaming API that utilizes callback closures. Please refer to the example code for more information.

Re-exports

pub use sys::set_log_level;

pub use sys::set_random_seed;

pub use sys::BatchType;

pub use sys::ComputeType;

pub use sys::Config;

pub use sys::Device;

pub use sys::LogLevel;

pub use sys::ScoringOptions;

pub use sys::ScoringResult;

Modules

sys

This module provides raw bindings for CTranslate2.

tokenizers

This module provides tokenizers.

Structs

GenerationOptions

The set of generation options.

GenerationStepResult

The result for a single generation step.

Generator

A text generator with a tokenizer.

TranslationOptions

Options for translation.

Translator

A text translator with a tokenizer.

Whisperwhisper

A speach transcriber using the Whisper speech recognition model published by OpenAI.

WhisperOptionswhisper

Options for whisper generation.

Traits

Tokenizer

Defines the necessary functions for a tokenizer.

Functions

download_modelhub

Downloads a model by its ID and returns the path to the directory where the model files are stored.