A Unified Library for Parameter-Efficient and Modular Transfer Learning

The Adapter-Hub website

ARCHIVED. Please use https://docs.adapterhub.ml/huggingface_hub.html || 🔌 A central repository collecting pre-trained adapter modules

ARCHIVED. Go to https://github.com/adapter-hub/adapters for the latest version.

Research code for "What to Pre-Train on? Efficient Intermediate Task Selection", EMNLP 2021

Research code for the paper "How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models"

UNKs Everywhere: Adapting Multilingual Language Models to New Scripts