1git clone --recursive https://github.com/FunAudioLLM/CosyVoice.git
2# If you failed to clone submodule due to network failures, please run following command until success
3cd CosyVoice
4git submodule update --init --recursive
We strongly recommand that you download our pretrained CosyVoice-300MCosyVoice-300M-SFTCosyVoice-300M-Instruct model and speech_kantts_ttsfrd resource.
If you are expert in this field, and you are only interested in training your own CosyVoice model from scratch, you can skip this step.
For zero_shot/cross_lingual inference, please use CosyVoice-300M model.
For sft inference, please use CosyVoice-300M-SFT model.
For instruct inference, please use CosyVoice-300M-Instruct model.
First, add third_party/AcademiCodec and third_party/Matcha-TTS to your PYTHONPATH.
1from cosyvoice.cli.cosyvoice import CosyVoice
2from cosyvoice.utils.file_utils import load_wav
3import torchaudio
45cosyvoice = CosyVoice('speech_tts/CosyVoice-300M-SFT')6# sft usage7print(cosyvoice.list_avaliable_spks())8output = cosyvoice.inference_sft('你好，我是通义生成式语音大模型，请问有什么可以帮您的吗？','中文女')9torchaudio.save('sft.wav', output['tts_speech'],22050)1011cosyvoice = CosyVoice('speech_tts/CosyVoice-300M')12# zero_shot usage13prompt_speech_16k = load_wav('zero_shot_prompt.wav',16000)14output = cosyvoice.inference_zero_shot('收到好友从远方寄来的生日礼物，那份意外的惊喜与深深的祝福让我心中充满了甜蜜的快乐，笑容如花儿般绽放。','希望你以后能够做的比我还好呦。', prompt_speech_16k)15torchaudio.save('zero_shot.wav', output['tts_speech'],22050)16# cross_lingual usage17prompt_speech_16k = load_wav('cross_lingual_prompt.wav',16000)18output = cosyvoice.inference_cross_lingual('<|en|>And then later on, fully acquiring that company. So keeping management in line, interest in line with the asset that\'s coming into the family is a reason why sometimes we don\'t buy the whole thing.', prompt_speech_16k)19torchaudio.save('cross_lingual.wav', output['tts_speech'],22050)2021cosyvoice = CosyVoice('speech_tts/CosyVoice-300M-Instruct')22# instruct usage23output = cosyvoice.inference_instruct('在面对挑战时，他展现了非凡的<strong>勇气</strong>与<strong>智慧</strong>。','中文男','Theo \'Crimson\', is a fiery, passionate rebel leader. Fights with fervor for justice, but struggles with impulsiveness.')24torchaudio.save('instruct.wav', output['tts_speech'],22050)
Start web demo
You can use our web demo page to get familiar with CosyVoice quickly.
We support sft/zero_shot/cross_lingual/instruct inference in web demo.
Please see the demo website for details.
python
1# change speech_tts/CosyVoice-300M-SFT for sft inference, or speech_tts/CosyVoice-300M-Instruct for instruct inference2python3 webui.py --port 50000--model_dir speech_tts/CosyVoice-300M
Advanced Usage
For advanced user, we have provided train and inference scripts in examples/libritts/cosyvoice/run.sh.
You can get familiar with CosyVoice following this recipie.
Build for deployment
Optionally, if you want to use grpc for service deployment,
you can run following steps. Otherwise, you can just ignore this step.
sh
1cd runtime/python
2docker build -t cosyvoice:v1.0 .
3# change speech_tts/CosyVoice-300M to speech_tts/CosyVoice-300M-Instruct if you want to use instruct inference
4docker run -d --runtime=nvidia -p 50000:50000 cosyvoice:v1.0 /bin/bash -c "cd /opt/CosyVoice/CosyVoice/runtime/python && python3 server.py --port 50000 --max_conc 4 --model_dir speech_tts/CosyVoice-300M && sleep infinity"
5python3 client.py --port 50000 --mode <sft|zero_shot|cross_lingual|instruct>
The content provided above is for academic purposes only and is intended to demonstrate technical capabilities. Some examples are sourced from the internet. If any content infringes on your rights, please contact us to request its removal.
