Pl Asr Benchmark: Polish Medical Speech Recognition Evaluation Corpus

Polish ASR Benchmark combines real-world Creative Commons video recordings with synthetic speech from healthcare articles. The dataset is intended for evaluating transcription quality on domain-specific vocabulary like medical terminology and pharmaceuticals. Author akrasi uploaded it to Hugging Face, with a last recorded update in June 2026.

Use Cases

• Benchmark ASR system performance on Polish medical terminology based on the described domain-specific vocabulary.
• Train speech models on healthcare-related speech based on the combination of real and synthetic audio described.
• Evaluate transcription accuracy on anatomical and pharmaceutical terms based on the described content focus.

Strengths

• Combines real-world speech from Creative Commons videos with synthetic speech from articles, as described.
• Focuses on a specific domain (medical/healthcare) and language (Polish) for targeted evaluation.
• Last updated date is explicitly provided as 2026-06-02 21:09:00.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count, file formats, and column-level documentation are absent.
• The dataset's size and specific composition of real vs. synthetic samples are unknown.

Provenance

Source

Hugging Face dataset by author akrasi.

Collection Method

Combines speech from publicly available Creative Commons videos and synthetic speech generated from Polish healthcare articles.

Freshness

Last updated 2026-06-02 21:09:00.