This is a single-speaker synthetic speech dataset for European Portuguese. It carries the Miro voice, a male voice. The audio was synthesized with text-to-speech and adapted to the Miro speaker identity, to train a Miro voice model for European Portuguese. The dataset contains 1150 recordings, with a metadata.csv transcript file and one WAV file per line.
Models trained on this dataset:… See the full description on the dataset page:
https://huggingface.co/datasets/TigreGotico/tts-train-synthetic-miro_pt-PT_v2.
