adapter-transformers

⚠️ IMPORTANT NOTE ⚠️

This is the legacy adapter-transformers library, which has been replaced by the new Adapters library, found here: https://github.com/adapter-hub/adapters.

• ⚠️ Beginning with version 4.0.0, the adapter-transformers package will automatically install the latest adapters package version instead. ⚠️
  From this version on, the adapter-transformers package does not contain any own functionality.
• Older versions of adapter-transformers are kept for archival purposes, and should not be used for active projects.

Install the new library directly via: pip install adapters.

The documentation of adapter-transformers can be found at https://docs-legacy.adapterhub.ml. The documentation of the new Adapters library can be found at https://docs.adapterhub.ml. For transitioning, please read: https://docs.adapterhub.ml/transitioning.html.

Adapters

A Unified Library for Parameter-Efficient and Modular Transfer Learning

🌍 Website   •   💻 GitHub   •   📚 Docs   •   📜 Paper   •   🧪 Tutorials

Adapters is an add-on library to HuggingFace's Transformers, integrating various adapter methods into state-of-the-art pre-trained language models with minimal coding overhead for training and inference.

pip install adapters