VoxCPM 2 Reaches TOP 2 on Hugging Face Trending!
Congratulations to VoxCPM 2 on this excellent result~ Thank you to developers and community partners around the world for the enthusiastic support 😻✨!
As a brand-new version of the speech generation model, VoxCPM2 brings some solid upgrades in usability and listening experience:
🌈 More languages:
Native support for 30 languages and 9 Chinese dialects, with no language tags needed, enabling natural cross-lingual synthesis.
🎨 Voice creation:
Supports "create any voice" — not only clone the voice you want, but also faithfully reproduce the emotion and style in it.
🎙️ High-fidelity audio:
Built-in super-resolution: 16 kHz reference audio in, 48 kHz ultra-clear original audio out, with every breath detail faithfully restored.
👍 Blazing inference + truly open source:
RTF as low as 0.13, with real-time streaming generation. Fully open-sourced under Apache-2.0 — take the weights and code~
Thank you to the VoxCPM team for their craftsmanship! We will continue to invest deeply in open source and explore the endless possibilities of voice interaction~ Community members are welcome to download and try it. If you think our model is pretty good, please give us a Like 💖 to help us out! 🙌
➤ Try it
🔗 https://voxcpm.modelbest.cn/
➤ 🤗 Hugging Face
🔗https://huggingface.openbmb.com/model/openbmb/VoxCPM2
➤ 💻 GitHub
🔗https://github.com/OpenBMB/VoxCPM/