Echo: A crowd-sourced Romanian speech dataset.
Romanian is the seventh most popular European language, with around 30 million speakers worldwide. Despite its popularity, the available speech resources are limited. As a result, there are few models that transcribe Romanian well, most of them being multilingual models that also cover less pop...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
ASLERD
2024-11-01
|
| Series: | Interaction Design and Architecture(s) |
| Online Access: | https://ixdea.org/62_9/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832585185641627648 |
|---|---|
| author | Remus-Dan Ungureanu Mihai Dascalu |
| author_facet | Remus-Dan Ungureanu Mihai Dascalu |
| author_sort | Remus-Dan Ungureanu |
| collection | DOAJ |
| description |
Romanian is the seventh most popular European language, with around 30 million speakers worldwide. Despite its popularity, the available speech resources are limited. As a result, there are few models that transcribe Romanian well, most of them being multilingual models that also cover less popular languages. Echo is a crowd-sourcing platform that has collected more than 300 hours of speech from various contributors. In this study, we document how a large speech dataset enables researchers to train automatic speech recognition, speaker verification, and diarization models to automatically process students’ notes. We publicly release both the dataset and the Whisper-based baseline model as open-source. |
| format | Article |
| id | doaj-art-96eb50bfafa04ddd923dde278887c976 |
| institution | Kabale University |
| issn | 2283-2998 |
| language | English |
| publishDate | 2024-11-01 |
| publisher | ASLERD |
| record_format | Article |
| series | Interaction Design and Architecture(s) |
| spelling | doaj-art-96eb50bfafa04ddd923dde278887c9762025-01-26T18:43:17ZengASLERDInteraction Design and Architecture(s)2283-29982024-11-016214115210.55612/s-5002-062-009Echo: A crowd-sourced Romanian speech dataset.Remus-Dan UngureanuMihai Dascalu Romanian is the seventh most popular European language, with around 30 million speakers worldwide. Despite its popularity, the available speech resources are limited. As a result, there are few models that transcribe Romanian well, most of them being multilingual models that also cover less popular languages. Echo is a crowd-sourcing platform that has collected more than 300 hours of speech from various contributors. In this study, we document how a large speech dataset enables researchers to train automatic speech recognition, speaker verification, and diarization models to automatically process students’ notes. We publicly release both the dataset and the Whisper-based baseline model as open-source.https://ixdea.org/62_9/ |
| spellingShingle | Remus-Dan Ungureanu Mihai Dascalu Echo: A crowd-sourced Romanian speech dataset. Interaction Design and Architecture(s) |
| title | Echo: A crowd-sourced Romanian speech dataset. |
| title_full | Echo: A crowd-sourced Romanian speech dataset. |
| title_fullStr | Echo: A crowd-sourced Romanian speech dataset. |
| title_full_unstemmed | Echo: A crowd-sourced Romanian speech dataset. |
| title_short | Echo: A crowd-sourced Romanian speech dataset. |
| title_sort | echo a crowd sourced romanian speech dataset |
| url | https://ixdea.org/62_9/ |
| work_keys_str_mv | AT remusdanungureanu echoacrowdsourcedromanianspeechdataset AT mihaidascalu echoacrowdsourcedromanianspeechdataset |