# High-Fidelity Audio

Llasa is a text-to-speech (TTS) system based on LLaMA, which extends the capabilities of the language model by integrating speech tokens, supporting Chinese and English speech generation.

Bark is a Transformer-based text-to-audio model created by Suno, capable of generating highly realistic multilingual speech, music, background noise, and sound effects.

Stable Audio Tools is a text-to-audio model capable of generating high-quality audio content based on text descriptions.

F5-TTS is a German speech synthesis model based on flow matching technology, focusing on generating smooth and faithful speech output.

An English text-to-speech model based on the VITS architecture, trained by Kakao Enterprise, supporting high-quality speech synthesis

MusicGen is a text-to-music generation model developed by Meta AI, capable of producing high-quality music samples based on text descriptions or audio prompts.

This is a voice conversion model based on RVC (Retrieval-based Voice Conversion) technology, capable of transforming input audio into Harry Styles' distinctive vocal style.

This is a voice conversion model based on RVC (Retrieval-based Voice Conversion) technology, capable of transforming input audio into Taylor Swift-style speech.

This is a voice conversion model based on RVC (Retrieval-based Voice Conversion) technology, capable of transforming input audio into Michael Jackson-style speech.

This is a voice conversion model based on RVC (Retrieval-Based Voice Conversion) technology, capable of transforming input audio into speech with a specific style.

This is a voice conversion model based on RVC (Retrieval-based Voice Conversion) technology, specifically designed to transform input audio into the vocal style of BLACKPINK member JISOO.

MusicGen is a text-to-music model that generates high-quality music samples based on text descriptions or audio prompts, utilizing a 1.5-billion-parameter autoregressive Transformer architecture.

Bark is a Transformer-based text-to-audio model created by Suno, capable of generating highly realistic multilingual speech, music, background noise, and simple sound effects.

A SpeechT5 speech synthesis (text-to-speech) model fine-tuned on the LibriTTS dataset, supporting high-quality text-to-speech conversion.

This is a text-to-speech (TTS) model based on ESPnet2, trained using the LJSpeech dataset, combining Conformer, FastSpeech2, and HiFi-GAN architectures.

This is a ConvTasNet model trained based on the Asteroid framework, specifically designed for audio separation tasks, trained on the sep_clean task of the Libri2Mix dataset.