TigreGotico/

tts-train-synthetic-dii_fa-IR

This is a single-speaker synthetic speech dataset for Persian. It carries the Dii voice, a female voice. The audio was synthesized with text-to-speech and adapted to the Dii speaker identity, to train a Dii voice model for Persian. The dataset contains 1997 recordings, with a metadata.csv transcript file and one WAV file per line.

Dataset collection: https://huggingface.co/collections/TigreGotico/synthetic-tts-datasets… See the full description on the dataset page: https://huggingface.co/datasets/TigreGotico/tts-train-synthetic-dii_fa-IR.