Whisper Large v3 Languages of Spain
Whisper Large v3 LoS is BSC's multilingual adaptation for Spanish, Catalan, Galician and Basque speech.
Model and training
BSC fine-tuned OpenAI Whisper Large v3 on 8,110 hours balanced across the four languages. Basque material includes speed-based augmentation at 0.9 and 1.1 times the original rate. This release produces plain transcripts. Its shared multilingual training differs from the individual Catalan-accent adaptations.Whisper Large v3 Languages of Spain: official model card
Practical use
Practical guidance. Use one recogniser across the four languages when operational simplicity matters, then check language-specific error rates. Balanced training hours do not guarantee balanced accuracy. Include short recordings where language identification is difficult, and compare with a specialist checkpoint before accepting a decline in a smaller language.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | whisper-large-v3-LoS |
| License | Apache License 2.0.Whisper Large v3 Languages of Spain: official model card |
Related models
- Conformer Transducer Languages of Spain
Barcelona Supercomputing Center · BSC LoS Conformer
BSC's LoS Conformer adapts NVIDIA's Spanish Conformer Transducer to four languages of Spain.
Launch date not confirmedDownloadable This BSC Whisper adaptation produces punctuated transcripts in Spanish, Catalan, Galician and Basque.
Launch date not confirmedDownloadable