🗣️ WavSLM

A single-stream speech language model based on WavLM distillation and FocalCodec.

This repository contains the checkpoint with a codebook size of 65536 trained on Libri-Light, as described in the paper.

📜 Paper: WavSLM: Single-Stream Speech Language Modeling via WavLM Distillation
🌐 Project Page: https://lucadellalib.github.io/wavslm-web/
💾 GitHub: https://github.com/lucadellalib/wavslm

▶️ Quickstart

See the readme at: https://github.com/lucadellalib/wavslm

@ Citing

@inproceedings{dellalibera2026wavslm,
    title     = {{WavSLM}: Single-Stream Speech Language Modeling via {WavLM} Distillation},
    author    = {Luca {Della Libera} and Cem Subakan and Mirco Ravanelli},
    booktitle = {Interspeech},
    year      = {2026},
}