Show simple item record

 
dc.contributor.author Lebar Bajec, Iztok
dc.contributor.author Bajec, Marko
dc.date.accessioned 2025-04-18T08:40:14Z
dc.date.available 2025-04-18T08:40:14Z
dc.date.issued 2025-04-17
dc.identifier.uri http://hdl.handle.net/11356/2024
dc.description This Conformer CTC BPE E2E Automated Speech Recognition model was trained following the NVIDIA NeMo Conformer-CTC fine-tuning 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 starting point was the Conformer CTC BPE E2E Automated Speech Recognition model RSDO-DS2-ASR-E2E 2.0, which was fine-tuned on the Protoverb closed dataset. The model was fine-tuned for 20 epochs, which improved the performance on the Protoverb test dataset for 9.8% relative WER, and for 3.3% relative WER on the Slobench dataset.
dc.language.iso slv
dc.publisher Faculty of Computer and Information Science, University of Ljubljana
dc.rights Apache License 2.0
dc.rights.uri https://opensource.org/licenses/Apache-2.0
dc.rights.label PUB
dc.source.uri https://www.inst-krim.si/project/proteverb/
dc.subject speech recognition
dc.subject NeMo
dc.subject model
dc.title Slovene Conformer CTC BPE E2E Automated Speech Recognition model PROTOVERB-ASR-E2E 1.0
dc.type toolService
metashare.ResourceInfo#ContentInfo.detailedType tool
metashare.ResourceInfo#ResourceComponentType#ToolServiceInfo.languageDependent true
has.files yes
branding CLARIN.SI data & tools
contact.person Marko Bajec marko.bajec@fri.uni-lj.si Faculty of Computer and Information Science, University of Ljubljana
sponsor ARIS in MDP V5-2265 Proteverb – Pravni, etični in tehnološki vidiki obdelave besedilnih in govornih virov podatkov za znanstvene, raziskovalne in razvojne namene Other
files.count 1
files.size 451804284


 Files in this item

This item is
Publicly Available
and licensed under:
Apache License 2.0
Icon
Name
sl-SI_MOL_nemo-1.0.tar.zst
Size
430.87 MB
Format
Unknown
Description
PROTOVERB ASR E2E 1.0
MD5
8b34365c365453901c84ec0b15893b08
 Download file

Show simple item record