huggingface/swift-transformers

Swift Package to implement a transformers-like API in Swift

What it solves

It provides a set of utilities to help developers integrate language models into Swift applications, bringing a familiar API similar to the Python transformers library to the Swift ecosystem.

How it works

The library offers several specialized modules:

  • Tokenizers: Handles text tokenization and chat templating, including native support for tool calling formats.
  • Hub: Enables fast and reliable downloads of models and configuration files directly from the Hugging Face Hub with progress reporting and connection handling.
  • Models & Generation: Provides utilities for working with language models specifically within CoreML for on-device execution.

Who it’s for

Swift developers who want to implement on-device machine learning and language model capabilities in their apps without needing to rewrite the core logic of the Hugging Face ecosystem in Swift.

Highlights

  • Idiomatic Swift API: Designed to feel familiar to users of the Python transformers library.
  • Smarter Downloads: Includes a Hub module for reliable model fetching and an optional Xet trait for parallel downloads.
  • CoreML Integration: Built-in support for loading compiled CoreML models and offline tokenizers to avoid network requests.
  • Tool Calling Support: Native formatting for complex model interactions involving external tools.

Related

  • Dispatch
    Swift Transformers Release: Run On-Device LLMs on Apple Devices

    Hugging Face released the alpha swift-transformers Swift package, swift-chat demo app, updated exporters and transformers-to-coreml tools, and Core ML versions of Llama 2 7B and Falcon 7B, enabling developers to run LLMs on-device in Apple apps.

  • Dispatch
    Swift Transformers 1.0 release notes / what's new

    Hugging Face has released Swift Transformers 1.0, a stable library providing tokenizers, hub access, and model wrappers to simplify local LLM integration for Apple Silicon developers.

  • Dispatch
    swift-huggingface release notes / what's new

    Hugging Face has released swift-huggingface, a comprehensive Swift client for the Hugging Face Hub that introduces reliable model downloads, a Python-compatible cache, and OAuth 2.0 support.

  • Project
    huggingface/huggingface_hub

    The official Python client and CLI for the Hugging Face Hub, enabling users to programmatically download, upload, and manage ML models, datasets, and demos.

  • Project
    eastriverlee/LLM.swift

    A Swift library for running local LLMs on Apple platforms, providing high-level abstractions for structured output, tool use, and chain-of-thought reasoning.