michaelcacioli/Neapolitan-Spoken-Corpus

Neapolitan Spoken Corpus (NSC) A corpus of read Neapolitan speech for ASR evaluation, with a validated Neapolitan–Italian lexicon, LOSO fine-tuning splits, trained LoRA adapters, metric implementations, per-clip results, and error annotations. This release supersedes the earlier 141-clip single-speaker version of this repository. The earlier release corresponds to Speaker S1 of the present corpus; the old audioData/ and transcripts.csv are replaced by data/audio/ and… See the full description on the dataset page: https://huggingface.co/datasets/michaelcacioli/Neapolitan-Spoken-Corpus.