A high-performance TTS model with voice cloning and emotion control for natural speech generation.
更新时间：2025-02-11 20:34:15
Zonos-v0.1是一款领先的开放权重文本转语音（TTS）模型，训练数据超过20万小时的多语言语音，致力于提供媲美甚至超越行业顶级TTS服务商的语音合成效果。该模型能够从文本提示生成自然流畅的语音，并通过提供说话人嵌入或音频前缀进行语音克隆，且支持情感表达（如快乐、悲伤、愤怒等）及语速、音高等控制。Zonos-v0.1的语音输出原生为44kHz的高质量音频。
使用Zonos-v0.1时，只需通过Python代码或Gradio界面进行简单操作。首先加载预训练模型，并通过提供音频文件生成说话人嵌入，然后输入文本并设置条件字典进行语音生成。如果需要进行批量生成，推荐使用Gradio界面以提高效率，避免每次都加载模型。
Zonos-v0.1是开源项目，支持在Linux系统（特别是Ubuntu 22.04/24.04）上安装，推荐使用具备NVIDIA 3000系列及以上显卡的计算机。可以通过Docker文件进行安装，并支持快速部署。
Zyphra
联系邮箱：[email protected]
社交媒体：Twitter：@ZyphraAI，Instagram：@zyphra.ai
A customizable AI companion for Twitch and YouTube streams, allowing for unique personalities and text-to-speech interactions.
A free online tool that generates various AI voices for TikTok, including character voices, language accents, and more.
kokoro-onnx is a lightweight Text-to-Speech (TTS) system based on the Kokoro model and ONNX runtime, offering fast, high-quality speech synthesis with multiple voices and languages. It’s optimized for macOS M1 devices and provides easy setup.
Kokoro TTS is a cutting-edge AI text-to-speech model with 82 million parameters, delivering high-quality, multilingual, and natural-sounding speech synthesis. Perfect for creating audiobooks, podcasts, and more.
LLaSA是一种基于LLaMA模型的文本到语音（TTS）合成系统，结合了XCodec2语音编码器，支持从文本或语音提示生成语音，已在25万个小时的中英双语数据集上训练。
LLaSA is an advanced system designed to scale both training and inference for LLaMA-based speech synthesis. It optimizes computational efficiency, leveraging large-scale datasets and cutting-edge machine learning frameworks to enhance text-to-speech performance.
IndexTTS is an advanced text-to-speech model that enhances pronunciation correction, pause control, and sound quality, setting new standards in TTS technology with a state-of-the-art zero-shot approach.
