Preprint
KoSpeech: Open-Source Toolkit for End-to-End Korean Speech Recognition
Soohwan Kim et al., arXiv, 2020
Read the originalRead on October 17, 2026
- What it is
- An open-source toolkit for end-to-end Korean speech recognition. It sets out a baseline on the KsponSpeech corpus of spontaneous Korean dialogue and reports character error rate, the usual metric for Korean in this line of research.
- How much weight it carries
- Peer-reviewed-style preprint from the toolkit's authors, with code released publicly. The result is a baseline from 2020 using an acoustic model alone, so it is not a current state of the art for Korean dictation.
Facts taken from this source
- The KoSpeech baseline reached a character error rate of 10.31 percent on the KsponSpeech corpus of spontaneous Korean, using the acoustic model only.
“By this metric, we achieved 10.31% CER only using acoustic models.” (Section 4, results)