OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Exporters

Exporters is a Python library that converts Hugging Face Transformers model checkpoints into Core ML models using ready-made or extendable configuration objects and coremltools. It provides command-line and Python interfaces for preparing models for Core ML deployment.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Exporters — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Exporters when a Transformers checkpoint needs to be converted to Core ML for deployment. For TensorFlow Lite conversion, the README points users to Hugging Face Optimum.

Who it is for

  • Developers preparing supported Transformers models for use in Core ML apps.
  • Python users who need to convert Hub or local model checkpoints to Core ML.
  • Contributors extending the library's ready-made model configurations.

Documented uses

  • Convert a supported Hub or local Transformers checkpoint into a Core ML model package.
  • Choose a task or model-topology configuration for an export.
  • Adjust documented export settings such as quantization or compute units through the Python interface.
  • Try a no-code Transformers-to-Core-ML conversion through the linked Space.
Before you choose

Requirements & limits

The project is a public GitHub repository under the Apache-2.0 license. The README documents installation by cloning the repository and running pip install -e .; macOS is recommended, while export is also possible on Linux.

  • The package is marked work in progress, and ready-made configurations cover a defined set of architectures rather than every Transformers model.
  • The README recommends Optimum for TensorFlow Lite conversion; Exporters' stated conversion focus is Core ML.
  • Linux can perform export, but the documented Core ML model validation requires macOS.
  • The default mlpackage format has documented minimum requirements of iOS 15, macOS 12.0 and Xcode 13.
  • The README cautions that models may fail conversion or produce incorrect outputs despite an apparently successful conversion.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install Exporters and export a Hub checkpoint to Core ML

Set up the Python package and convert a supported checkpoint into an .mlpackage file.

Before you start

  • A Python environment for the package and its dependencies.
  • A checkpoint on the Hugging Face Hub or stored locally.
  • macOS is recommended; Linux can run export, but the documented Core ML validation step requires macOS.

Steps

  1. Clone the repository: git clone https://github.com/huggingface/exporters.git

  2. Enter the repository and install it in editable mode: cd exporters && pip install -e .

  3. Run the documented example: python -m exporters.coreml --model=distilbert-base-uncased exported/

Expected outcome

The example writes a Core ML model named Model.mlpackage to the exported directory.

Guide limitations

  • The README marks the package as a work in progress and lists ready-made configurations for specific architectures; conversion support is not universal.
  • Linux export is possible, but the documented validation step runs only on macOS.
  • The default mlpackage/ML Program format requires iOS 15, macOS 12.0 and Xcode 13 for app use.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 November 2026
Original website title
huggingface/exporters · Official homepage title · retained in its source language
How we review sourcesSuggest a correction

Related options

Hugging Face Optimum — The Exporters README recommends Optimum for TensorFlow Lite conversion and suggests it for inference optimization before mobile deployment.