xflash

build under python3.11 success

env: echo %TORCH_CUDA_ARCH_LIST% "12.0" Remove --no-build-isolation facing: 

Also me facing same issue.

File "D:\Software\CosyVoice\testcosyvoice.txt", line 8, in cosyvoice = CosyVoice2('pretrained_models/CosyVoice2-0.5B', load_jit=False, load_trt=False, fp16=False, use_flow_cache=False) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "D:\Software\CosyVoice\cosyvoice\cli\cosyvoice.py", line 143, in __init__ configs = load_hyperpyyaml(f, overrides={'qwen_pretrain_path': os.path.join(model_dir, 'CosyVoice-BlankEN')}) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Program Files\Python311\Lib\site-packages\hyperpyyaml\core.py", line...

简单cpu跑下试试吧，50卡就别想了。目前为止就分享了docker，关键代码，模型都没有开源。用开源引流的快餐型项目而已。

和Python支持没关系吧，从项目名称开始就是走歪了 ，开源了客户端代码而已，放出来的服务端都是硅基已经要淘汰的部分.并且关键代码都是pyc so. 这种公司你想期待啥