Model card for ald (x-low)
- Language: es_MX (Spanish, Mexico)
- Speakers: 1
- Quality: x-low
- Samplerate: 22,050Hz
Dataset
Synthetic speech dataset generated using the original es_MX-ald-medium Piper TTS model.
The dataset contains approximately 8 hours of audio generated from Spanish sentences obtained from the Tatoeba Project: https://tatoeba.org/en/downloads
Training
This model was trained from scratch using the synthetic dataset generated with the es_MX-ald-medium voice.
The goal of this model is to provide a smaller and faster alternative with fewer parameters.