transformers>=4.41.2,<=4.49.0,!=4.46.0,!=4.46.1,!=4.46.2,!=4.46.3,!=4.47.0,!=4.47.1,!=4.48.0 fire sse-starlette datasets<=3.2.0,>=2.16.0 accelerate<=1.2.1,>=0.34.0 peft<=0.12.0,>=0.11.1 trl>=0.8.6,<=0.9.6 tokenizers>=0.19.0,<=0.21.0 numpy>=1.23.4,<2.0.0 scipy librosa>=0.9.2 sentencepiece einops tiktoken av>=11 numba>=0.56.4 pytorch-lightning>2.0 torchaudio ffmpeg-python vllm==0.6.4.post1 nltk wordsegment g2p_en gradio<=5.21.0,>=4.38.0 x_transformers deepspeed>=0.9.3 matplotlib tyro<0.9.0 modelscope num2words
Make Muyan-TTS avialbe in ComfyUI.
Muyan-TTS is a trainable TTS model designed for podcast applications within a $50,000 budget, which is pre-trained on over 100,000 hours of podcast audio data, enabling zero-shot TTS synthesis with high-quality voice generation. Furthermore, Muyan-TTS supports speaker adaptation with dozens of minutes of target speech, making it highly customizable for individual voices.
cd ComfyUI/custom_nodes
git clone https://github.com/Yuan-ManX/ComfyUI-Muyan-TTS.gitcd ComfyUI-Muyan-TTS
pip install -r requirements.txt| Models | Links |
|——-|——-|
| Muyan-TTS | huggingface \| modelscope |
| Muyan-TTS-SFT | huggingface \| modelscope |
Additionally, you need to download the weights of chinese-hubert-base.