# High-fidelity speech

A Spanish text-to-speech model fine-tuned based on Microsoft's SpeechT5 architecture, trained on the mabama-v6 dataset

F5-TTS is a fully non-autoregressive zero-shot text-to-speech model that supports high-quality speech synthesis.

This is a Taiwanese accent audio synthesis model based on the Bert-VITS2 architecture, capable of generating natural and fluent Taiwanese accent speech.

A Portuguese text-to-speech model fine-tuned based on Microsoft's SpeechT5 architecture, supporting high-quality speech synthesis

Greek text-to-speech model fine-tuned based on microsoft/speecht5_tts