# Encoder-only architecture

OWSM-CTC is an encoder-only speech foundation model based on hierarchical multi-task self-conditioned CTC, supporting multilingual speech recognition, speech translation, and language identification.