Reproduce “Developing a Speech Language Model for Basque” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Andoni Sudupe, Inma Hernaez, Ander Arriandiaga, Christoforos Souganidis, Ibon Saratxaga, Eva Navas
The authors adapt Latxa with speech components to build an end-to-end spoken language model for Basque, training it on translated instructions and synthesized audio. On 4,419 synthetic test instructions, it scores 3.86 out of 5 for content, with 9.94% word error rate and 370 milliseconds streaming latency under specified conditions. This demonstrates feasibility on synthetic data, not real user speech.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.