README.md
| 1 | --- |
| 2 | license: other |
| 3 | license_name: coqui-public-model-license |
| 4 | license_link: https://coqui.ai/cpml |
| 5 | library_name: coqui |
| 6 | pipeline_tag: text-to-speech |
| 7 | widget: |
| 8 | - text: "Once when I was six years old I saw a magnificent picture" |
| 9 | --- |
| 10 | |
| 11 | # ⓍTTS |
| 12 | ⓍTTS is a Voice generation model that lets you clone voices into different languages by using just a quick 6-second audio clip. There is no need for an excessive amount of training data that spans countless hours. |
| 13 | |
| 14 | This is the same or similar model to what powers [Coqui Studio](https://coqui.ai/) and [Coqui API](https://docs.coqui.ai/docs). |
| 15 | |
| 16 | ### Features |
| 17 | - Supports 17 languages. |
| 18 | - Voice cloning with just a 6-second audio clip. |
| 19 | - Emotion and style transfer by cloning. |
| 20 | - Cross-language voice cloning. |
| 21 | - Multi-lingual speech generation. |
| 22 | - 24khz sampling rate. |
| 23 | |
| 24 | ### Updates over XTTS-v1 |
| 25 | - 2 new languages; Hungarian and Korean |
| 26 | - Architectural improvements for speaker conditioning. |
| 27 | - Enables the use of multiple speaker references and interpolation between speakers. |
| 28 | - Stability improvements. |
| 29 | - Better prosody and audio quality across the board. |
| 30 | |
| 31 | ### Languages |
| 32 | XTTS-v2 supports 17 languages: **English (en), Spanish (es), French (fr), German (de), Italian (it), Portuguese (pt), |
| 33 | Polish (pl), Turkish (tr), Russian (ru), Dutch (nl), Czech (cs), Arabic (ar), Chinese (zh-cn), Japanese (ja), Hungarian (hu), Korean (ko) |
| 34 | Hindi (hi)**. |
| 35 | |
| 36 | Stay tuned as we continue to add support for more languages. If you have any language requests, feel free to reach out! |
| 37 | |
| 38 | ### Code |
| 39 | The [code-base](https://github.com/coqui-ai/TTS) supports inference and [fine-tuning](https://tts.readthedocs.io/en/latest/models/xtts.html#training). |
| 40 | |
| 41 | ### Demo Spaces |
| 42 | - [XTTS Space](https://huggingface.co/spaces/coqui/xtts) : You can see how model performs on supported languages, and try with your own reference or microphone input |
| 43 | - [XTTS Voice Chat with Mistral or Zephyr](https://huggingface.co/spaces/coqui/voice-chat-with-mistral) : You can experience streaming voice chat with Mistral 7B Instruct or Zephyr 7B Beta |
| 44 | |
| 45 | | | | |
| 46 | | ------------------------------- | --------------------------------------- | |
| 47 | | 🐸💬 **CoquiTTS** | [coqui/TTS on Github](https://github.com/coqui-ai/TTS)| |
| 48 | | 💼 **Documentation** | [ReadTheDocs](https://tts.readthedocs.io/en/latest/) |
| 49 | | 👩💻 **Questions** | [GitHub Discussions](https://github.com/coqui-ai/TTS/discussions) | |
| 50 | | 🗯 **Community** | [Discord](https://discord.gg/5eXr5seRrv) | |
| 51 | |
| 52 | |
| 53 | ### License |
| 54 | This model is licensed under [Coqui Public Model License](https://coqui.ai/cpml). There's a lot that goes into a license for generative models, and you can read more of [the origin story of CPML here](https://coqui.ai/blog/tts/cpml). |
| 55 | |
| 56 | ### Contact |
| 57 | Come and join in our 🐸Community. We're active on [Discord](https://discord.gg/fBC58unbKE) and [Twitter](https://twitter.com/coqui_ai). |
| 58 | You can also mail us at info@coqui.ai. |
| 59 | |
| 60 | Using 🐸TTS API: |
| 61 | |
| 62 | ```python |
| 63 | from TTS.api import TTS |
| 64 | tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2", gpu=True) |
| 65 | |
| 66 | # generate speech by cloning a voice using default settings |
| 67 | tts.tts_to_file(text="It took me quite a long time to develop a voice, and now that I have it I'm not going to be silent.", |
| 68 | file_path="output.wav", |
| 69 | speaker_wav="/path/to/target/speaker.wav", |
| 70 | language="en") |
| 71 | |
| 72 | ``` |
| 73 | |
| 74 | Using 🐸TTS Command line: |
| 75 | |
| 76 | ```console |
| 77 | tts --model_name tts_models/multilingual/multi-dataset/xtts_v2 \ |
| 78 | --text "Bugün okula gitmek istemiyorum." \ |
| 79 | --speaker_wav /path/to/target/speaker.wav \ |
| 80 | --language_idx tr \ |
| 81 | --use_cuda true |
| 82 | ``` |
| 83 | |
| 84 | Using the model directly: |
| 85 | |
| 86 | ```python |
| 87 | from TTS.tts.configs.xtts_config import XttsConfig |
| 88 | from TTS.tts.models.xtts import Xtts |
| 89 | |
| 90 | config = XttsConfig() |
| 91 | config.load_json("/path/to/xtts/config.json") |
| 92 | model = Xtts.init_from_config(config) |
| 93 | model.load_checkpoint(config, checkpoint_dir="/path/to/xtts/", eval=True) |
| 94 | model.cuda() |
| 95 | |
| 96 | outputs = model.synthesize( |
| 97 | "It took me quite a long time to develop a voice and now that I have it I am not going to be silent.", |
| 98 | config, |
| 99 | speaker_wav="/data/TTS-public/_refclips/3.wav", |
| 100 | gpt_cond_len=3, |
| 101 | language="en", |
| 102 | ) |
| 103 | ``` |
| 104 | |