ViMedCSS: Vietnamese Medical Speech with English Code-Switching Terms

ViMedCSS is a Vietnamese medical speech dataset designed for code-switching automatic speech recognition. It contains 11,832 training utterances totaling 24.30 hours, with each utterance embedding at least one non-Vietnamese medical term, primarily English. The dataset was created by tensorxt and is associated with the LREC 2026 conference.

Use Cases

• Train code-switching ASR models based on Vietnamese speech with embedded English medical terms.
• Benchmark speech recognition performance in a specialized medical domain.
• Study linguistic patterns of medical terminology mixing in Vietnamese speech.
• Develop language models for handling multilingual medical conversations.

Strengths

• Contains 11,832 training utterances.
• Provides 24.30 hours of total audio duration.
• Specifically designed for code-switching with at least one non-Vietnamese term per utterance.
• Associated with a peer-reviewed conference (LREC 2026).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The description metadata is limited; actual data quality requires manual inspection after download.
• The dataset may reflect bias inherent to its specific collection source and medical domain.

Provenance

Source

huggingface

Collection Method

Likely contains recorded medical speech utterances with manual or automated transcription and annotation.

Time Range

null

Freshness

Last updated 2026-02-20 03:33:18; freshness should be verified.

Geography

null