AxForge
Sign in Get started

Model reference · open weights

viXTTS

Audio capleaf Text→speech 1 build Its own licence terms 660 dl/mo

viXTTS is an open-weight audio or speech model from capleaf. viXTTS (BF16) weighs 1.9 GB; the smallest configuration that runs it is RTX 3060 12 GB.

Request Rent a GPU

For self-hosting, compare this build's requirements below with the GPU you plan to rent.

  • viXTTS is a text-to-speech model developed by capleaf that clones voices into different languages using a 6-second audio clip.
  • It is fine-tuned from XTTS-v2.0.3 by expanding the tokenizer to Vietnamese and training on the viVoice dataset.
  • The model supports 18 languages, including English, Spanish, French, German, Italian, Portuguese, Polish, Turkish, Russian, Dutch, Czech, Arabic, Chinese, Japanese, Hungarian, Korean, Hindi, and Vietnamese.
  • It is licensed under the Coqui Public Model License.

Summary of the capleaf/viXTTS model card, 2026-10-01

Hugging Face ↗

What it is

Released bycapleaf
Released2024-04-04
VRAM1.9 GB for the weights

What it runs on

Memory and cards for viXTTS (BF16)

1.9 GBweights, file size
1.6 GBruntime overhead
CardOne streamMemory
RTX 3060 12 GBfits11.6 GB
RTX 4060 Ti 16 GBfits15.4 GB
RTX 3090 24 GBfits23.4 GB
RTX 4090 24 GBfits23.4 GB
RTX 5090 32 GBfits31.0 GB
L40S 48 GBfits44.0 GB
A100 80 GBfits78.2 GB
H100 80 GBfits78.1 GB
RTX PRO 6000 Blackwell 96 GBfits93.8 GB
DGX Spark (GB10) 128 GB unifiedfits107 GB
H200 141 GBfits138 GB
B200 180 GBfits176 GB

From the model card

What capleaf says about viXTTS

Read the model card

viⓍTTS là mô hình tạo sinh giọng nói cho phép bạn sao chép giọng nói sang các ngôn ngữ khác nhau chỉ bằng cách sử dụng một đoạn âm thanh nhanh dài 6 giây. Mô hình này được tiếp tục đào tạo từ mô hình XTTS-v2.0.3 bằng cách mở rộng tokenizer sang tiếng Việt và huấn luyện trên tập dữ liệu viVoice.

viⓍTTS is a voice generation model that lets you clone voices into different languages by using just a quick 6-second audio clip. This model is fine-tuned from the XTTS-v2.0.3 model by expanding the tokenizer to Vietnamese and fine-tuning on the viVoice dataset.

Languages

viXTTS supports 18 languages: English (en), Spanish (es), French (fr), German (de), Italian (it), Portuguese (pt), Polish (pl), Turkish (tr), Russian (ru), Dutch (nl), Czech (cs), Arabic (ar), Chinese (zh-cn), Japanese (ja), Hungarian (hu), Korean (ko) Hindi (hi), Vietnamese (vi).

Known Limitations

  • Incompatibility with the original TTS library (a pull request will be made later).
  • Subpar performance for input sentences under 10 words in Vietnamese language (yielding inconsistent output and odd trailing sounds).
  • This model is only fine-tuned in Vietnamese. The model's effectiveness with languages other than Vietnamese hasn't been tested, potentially reducing quality.

Demo

Please checkout this repo

Usage

License

This model is licensed under Coqui Public Model License.

Contact

Fine-tuned by Thinh Le at FPT University HCMC, as a component of Non La's graduation thesis. Contact:

  • You can message me directly on Facebook: (preferred 🤗)
  • GitHub:
  • Email: or

Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.

How it works

How audio & music work

Audio or textinputAudio modelrecognise / synthesiseText or audiooutputSpeech-to-text turns audio into text; text-to-speech and music models turn text into audio.
© 2026 AxForge · EU-hosted AI infrastructure Pricing Docs Trust Privacy Terms