Vietnamese Text-To-Speech Audio Dataset

A 2022 dataset containing Vietnamese text and corresponding audio samples for speech synthesis tasks. Created by author 'duongmle' and hosted on Hugging Face, it is categorized as containing approximately 1,000 samples based on platform size tags.

Use Cases

• Training a Vietnamese TTS model using the paired text and audio samples.
• Fine-tuning a multilingual speech synthesis model by incorporating Vietnamese audio data.
• Benchmarking speech synthesis quality for Vietnamese using the provided audio-text pairs.
• Analyzing phonetic or prosodic features specific to Vietnamese from the audio recordings.

Strengths

• Provides paired text and audio data, essential for speech synthesis model training.
• Focuses on Vietnamese, a lower-resource language in speech AI.
• Hosted in a structured format (WebDataset) suitable for machine learning pipelines.

Limitations

• Small scale with approximately 1,000 samples, limiting model training potential.
• Limited metadata; speaker demographics, recording conditions, and audio quality are unknown.
• Last updated in June 2022, potentially lacking recent linguistic trends or model architectures.

Provenance

Source

Hugging Face, uploaded by user duongmle.

Collection Method

Collection method unknown; likely involves recording or synthesizing Vietnamese speech paired with text transcripts.

Time Range

Creation date circa 2022.

Freshness

Last updated in 2022.

Geography

Primarily Vietnamese language content, suggesting origin in Vietnam or Vietnamese-speaking regions.