Datasets:
audio audioduration (s) 0.49 49.1 |
|---|
YAML Metadata Warning:The task_ids "text-to-speech" is not in the official list: acceptability-classification, entity-linking-classification, fact-checking, intent-classification, language-identification, multi-class-classification, multi-label-classification, multi-input-text-classification, natural-language-inference, semantic-similarity-classification, sentiment-classification, topic-classification, semantic-similarity-scoring, sentiment-scoring, sentiment-analysis, hate-speech-detection, text-scoring, named-entity-recognition, part-of-speech, parsing, lemmatization, word-sense-disambiguation, coreference-resolution, extractive-qa, open-domain-qa, closed-domain-qa, news-articles-summarization, news-articles-headline-generation, dialogue-modeling, dialogue-generation, conversational, language-modeling, text-simplification, explanation-generation, abstractive-qa, open-domain-abstractive-qa, closed-domain-qa, open-book-qa, closed-book-qa, text2text-generation, slot-filling, masked-language-modeling, keyword-spotting, speaker-identification, audio-intent-classification, audio-emotion-recognition, audio-language-identification, multi-label-image-classification, multi-class-image-classification, face-detection, vehicle-detection, instance-segmentation, semantic-segmentation, panoptic-segmentation, image-captioning, image-inpainting, image-colorization, super-resolution, grasping, task-planning, tabular-multi-class-classification, tabular-multi-label-classification, tabular-single-column-regression, rdf-to-text, multiple-choice-qa, multiple-choice-coreference-resolution, document-retrieval, utterance-retrieval, entity-linking-retrieval, fact-checking-retrieval, univariate-time-series-forecasting, multivariate-time-series-forecasting, visual-question-answering, document-question-answering, pose-estimation
🇷🇸 Serbian Common Voice Style TTS Dataset
A Serbian single-speaker speech dataset prepared for Text-to-Speech (TTS) training using the Common Voice dataset formatting style.
📌 Dataset Overview
This dataset was created for training Serbian TTS models using frameworks such as Coqui TTS and VITS.
The dataset contains high-quality Serbian speech recordings paired with manually created text transcripts written in Cyrillic script.
Features
- Language: Serbian
- Script: Cyrillic
- Format: Common Voice compatible
- Use Case: Text-to-Speech
- Audio Format: WAV
- Sampling Rate: 22050 Hz
- Speaker Type: Single-speaker female voice
⚠️ Speaker & Content Disclaimer
The voice recordings contained in this dataset are not the author's personal voice.
The speaker voluntarily provided explicit permission for the recordings to be used, processed, and publicly released for research and open-source TTS purposes.
The dataset follows the same licensing approach as Common Voice and is released under the CC-BY-4.0 license.
All transcripts included in this dataset were independently and manually written specifically for this dataset.
No text content was copied, extracted, or reused from:
- existing proprietary datasets
- audiobooks
- books
- subtitles
- copyrighted speech corpora
- commercial TTS datasets
The text material was created solely for the purpose of building an open Serbian TTS dataset.
📚 Dataset Structure
Example structure:
clips/
├── 0001.wav
├── 0002.wav
├── 0003.wav
clean.tsv
Example clean.tsv format:
client_id path sentence
Dragana 0001 Здраво, како си?
Dragana 0002 Добродошли у систем.
📊 Dataset Information
- Approximately 8 hours of speech
- Single female speaker
- Serbian Cyrillic text only
- Cleaned and normalized transcripts
- Numbers removed from transcripts
- Manually verified alignments
⚙️ Compatibility
This dataset is compatible with:
- Coqui TTS
- VITS
- XTTS preprocessing pipelines
- Common Voice dataset formatter
- Custom PyTorch TTS pipelines
📜 License
This dataset is released under the CC-BY-4.0 license.
The speaker has provided permission for public distribution and use of the recordings for open-source speech synthesis research and development.
👤 Author
Darko Milošević
🤝 Acknowledgements
- Mozilla Common Voice formatting approach
- Coqui TTS
- VITS Architecture
⚠️ Limitations
- Cyrillic script only
- Numbers are not included
- Single-speaker dataset
- May not generalize well to rare Serbian words or dialects
- Downloads last month
- 1,878