Whisper Large v3 Languages of Spain Punctuated
This BSC Whisper adaptation produces punctuated transcripts in Spanish, Catalan, Galician and Basque.
Model and training
It starts from OpenAI Whisper Large v3 and uses 8,110 hours of balanced multilingual speech. Punctuation is part of this trained release, rather than a toggle on the plain LoS checkpoint. The card documents the four-language adaptation and its data sources. Both LoS variants remain available as weights.Whisper Large v3 Languages of Spain Punctuated: official model card
Practical use
Practical guidance. Compare readability and word error separately across all four languages. Punctuation conventions and capitalisation can vary with language and document style. Keep the plain checkpoint as a useful reference when a downstream system needs words and timings without sentence formatting.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | whisper-large-v3-LoS-punctuated |
| License | Apache License 2.0.Whisper Large v3 Languages of Spain Punctuated: official model card |
Related models
- Conformer Transducer Languages of Spain Punctuated
Barcelona Supercomputing Center · BSC LoS Conformer
This BSC LoS Conformer release adds trained punctuation to multilingual speech recognition.
Launch date not confirmedDownloadable Whisper Large v3 LoS is BSC's multilingual adaptation for Spanish, Catalan, Galician and Basque speech.
Launch date not confirmedDownloadable