WildElder: Chinese Elderly Speech Dataset with Fine-Grained Annotations

WildElder is a speech dataset focused on elderly scenarios, containing raw audio and corresponding text annotations. The data was collected and cleaned from real-world environments to preserve diversity. The dataset is authored by Hui519 and was last updated on 2026-05-02.

Use Cases

• Train or evaluate Automatic Speech Recognition (ASR) models based on the raw audio and text annotations.
• Conduct speaker-related research, such as identification or diarization, based on the elderly speech data.
• Develop front-end speech processing techniques, like enhancement or feature extraction, using the real-world audio.
• Research back-end speech processing, such as language modeling or intent recognition, using the annotated transcripts.

Strengths

• Contains fine-grained manual annotations for the audio data.
• Data was collected from real-world environments, likely preserving natural diversity.
• Specifically focuses on the elderly demographic and Chinese language.

Limitations

• Row count, file formats, and column-level documentation are unknown.
• License information is not provided, which may restrict usage.
• The description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Hui519 on Hugging Face.

Collection Method

Collected and cleaned from real-world environments.

Time Range

null

Freshness

Last updated 2026 05 02 02:45:34; freshness should be verified.

Geography

null