ChatTTS is a text to speech model specially designed for dialogue scenarios, a high-quality, multi-functional text to speech model that performs well in various conversational applications. It supports Chinese and English speech generation and, trained on approximately 100,000 hours of Chinese and English data, is capable of generating speech quality comparable to human dialogue. ChatTTS is particularly suitable for dialogue tasks of large language model assistants, as well as applications such as creating dialogue-based audio and video introductions. Based on open-source natural language processing and speech synthesis technologies, it provides developers with a powerful and easy-to-use tool.
Why ChatTTS
Make your audio generated by AI sound like a real person with human-like intonations and pauses.
Break the language barrier with our dual language support for English and Chinese.
Rest assured in the efficiency of ChatTTS, designed from over 40,000 hours of pre training
Get access to well-maintained and regularly updated source code on GitHub.
Let's start using this repository locally with basic usage, running in Colab and running in Huggingface & Modelscope. Convert text into natural conversational human voice.
Run your ChatTTS repository locally to achieve a magical text-to-speech experience.
Ten minutes of time, allow you to successfully run ChatTTS on Colab.
Step by step to deploy your ChatTTS Space on HuggingFace.
Navigate to the open-source repository, choose the appropriate folder, and clone the remote repository to your local machine using git commands. Alternatively, you can choose to manually download it from GitHub.
git clone https://github.com/2noise/ChatTTS.git
Enter the folder where you downloaded the files in the terminal or command line, then run the following command to download the dependencies.
pip install omegaconf -q
pip install vocos -q
pip install vector_quantize_pytorch -q
pip install nemo_text_processing -q
pip install WeTextProcessing -q
Import the package, and declare the Python modules and instances we need.
import torch
from ChatTTS.core import Chat
from IPython.display import Audio
chat = ChatTTS.Chat()
chat.load_models()
Determine the text you need to generate speech for, and save it as 'texts'.
texts = ["YOUR_TEXT_TO_GENERATE_AUDIO",]
Generate the speech.
wavs = chat.infer(texts, use_decoder=True)
Play the Audio
Audio(wavs[0], rate=24_000, autoplay=True)
Frequently Asked Questions for the ChatTTS GitHub repository, a generative speech model for daily dialogue.
ChatTTS is a text to speech model designed specifically for dialogue scenarios such as LLM assistants. It supports both English and Chinese languages and is trained with over 100,000 hours composed of Chinese and English. The open-source version on HuggingFace is a 40,000-hour pre-trained model.
Yes, ChatTTS is free to use. You can download the project files from the GitHub repository to your local machine. There are also other developers who have created free versions available on well-known open-source platforms such as Github, HuggingFace and Modelscope.
Installation steps for ChatTTS are outlined in the 'How it works' section. Basicly, you could directly downlownd the text to speech project into your laptop from Github repo and use it in python. You can also follow the prompts on the official GitHub page for downloading and using ChatTTS.
The source code for ChatTTS can be found on its GitHub repository at https://github.com/2noise/ChatTTS.
ChatTTS is currently available in English and Chinese.
To use ChatTTS in your project, you can import it and use the `chat.infer` method with your text. More detailed usage examples might be available in the repository's documentation or example files.
Yes, contributions to the ChatTTS project are welcome in various forms, including issue discussions, GitHub issues submissions, and pull requests. You can also join the QQ group: 808364215 for discussions.
For formal inquiries about the model and its roadmap, you can contact the developers at [email protected]. Joining their QQ group or submitting GitHub issues for support is also encouraged.
ChatTTS is optimized for dialogue-based tasks, enabling natural and expressive speech synthesis with support for multiple speakers. It offers fine-grained control over prosodic features like laughter, pauses, and interjections, and surpasses most open-source TTS models in terms of prosody.
@2024 @ChatTTS.Site all rights reserved. We are not official website of ChatTTS.