Slovene Conformer CTC BPE E2E Automated Speech Recognition model RSDO-DS2-ASR-E2E 2.0

  Authors

Lebar Bajec, Iztok ; Bajec, Marko ; Bajec, Žan and Rizvič, Mitja

  Item identifier

http://hdl.handle.net/11356/1737

 Project URL

https://rsdo.slovenscina.eu/en/speech-technologies

 Demo URL

https://www.slovenscina.eu/en/razpoznavalnik

 Referenced by

https://github.com/clarinsi/Slovene_ASR_e2e

 Date issued

2022-12-01

 Type

toolService

 Language(s)

Slovenian

 Description

This Conformer CTC BPE E2E Automated Speech Recognition model was trained following the NVIDIA NeMo Conformer-CTC recipe (for details see the official NVIDIA NeMo NMT documentation, https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/stable/asr/intro.html, and NVIDIA NeMo GitHub repository https://github.com/NVIDIA/NeMo). It provides functionality for transcribing Slovene speech to text. The training, development and test datasets were based on the Artur dataset and consisted of 630.38, 16.48 and 15.12 hours of transcribed speech in standardised form, respectively. The model was trained for 200 epochs and reached WER 0.0429 on the development and WER 0.0558 on the test dataset.

 Publisher

Faculty of Computer and Information Science, University of Ljubljana

 Acknowledgement

Ministry of Culture C3340-20-278001 "Development of Slovene in a Digital Environment"

 Subject(s)

speech recognition NeMo model

 Collection(s)

CLARIN.SI data & tools

Show full item record