IN THE WORKS
Current version: Alpha (not released yet)
This repository houses the data accompanying the below cited masters thesis. The ASR model included is the Russian-based one (for which the greatest WER was acheived). It should be downloadable and usedable from here, or on HuggingFace (found in the .gitmodules file).
The originator of the dataset and main author is Enzo Gamboni. The second author and administrator of this project is Michael Rießler.
The paper by Gamboni (2025, to appear) includes a complete description. If you use the data, please cite this paper (see the bibtex code snippet below).
@mastersthesis{gamboni2025a,
address = {Joensuu},
author = {Gamboni, Enzo},
school = {University of Eastern Finland, Philosophical faculty, School of Humanities},
title = {sjd-ASR},
year = {2025, to appear}}
The code and data in this repository are free and open and licensed under CC-BY, see LICENSE. Other parts of the data – under bound licenses – are found in sjd-fair and sjd-bound, which are private repositories and visible only to project collaborators.
If you are interested in data use or collaboration, contact Michael Rießler.