This dataset contains Brazilian Portuguese speech recorded in multiple regions of Brazil, including scripted speech and spontaneous speech on everyday topics. The evaluation sample includes the source text for scripted utterances and transcripts for spontaneous speech.
Use Cases
- Fine-tuning Brazilian Portuguese ASR models: Speech recordings paired with their source scripts or transcripts can be used to fine-tune automatic speech recognition (ASR) models for Brazilian Portuguese.
- Regional accent adaptation: Speech collected across different regions of Brazil can be used to build speech recognition models that account for regional differences in pronunciation and accent.