-
在服务器部署后,推理速度比较慢,GPU没有吃满,吃满了一个CPU核心,是不是我的部署方法有问题
-
(venv) D:\talkingface\V-Express>python inference.py --reference_image_path "./test_samples/short_case/10/ref.jpg" --audio_path "./test_samples/short_case/10/aud.mp3" --kps_path "./test_samples/short_c…
-
略显生硬
-
我在A100上运行的chattts,生成长文本需要很长时间,后面尝试去做文本分割,一段段的生成,速度也还是不行,之前实现了基于paddlespeech的语音合成,流式的效果很好,但是音色比较拉跨,有没有合成效果不那么拉跨,而且速度还可以的tts推荐
-
Env: Python 3.10
Issue summary:
```
P: Error #15: Initializing libomp.dylib, but found libiomp5.dylib already initialized.
OMP: Hint This means that multiple copies of the OpenMP runtime have be…
-
hi,我想咨询下data 路径下speaker增加的方法。是使用了VQ Encoder 将speaker的语音转换成了embedding并保存为pt文件吗?还是有其他别的方法?谢谢~
-
torch._dynamo.exc.BackendCompilerFailed: backend='inductor' raised:
InvalidCxxCompiler: No working C++ compiler found in torch._inductor.config.cpp.cxx: (None, 'g++')
Python版本:3.10.4
操作系统:win10
…
-
训练了几个克隆声音的模型,可以根据参考音频生成ChatTTS使用的音色嵌入。
模型Demo页面:
http://region-9.autodl.pro:41137
欢迎大家反馈一下测试效果,加入QQ群474529811给建议或者讨论。
-----
用法:下载声音模型(.pt结尾),记住temperature要设置得非常低,否则声音会不准确。
`
rand_spk = to…
-
Hi
Where can I download the model for the project?
-
##报错步骤
安装requirements.txt环境后运行webui.py出现,配置环境有教程吗
![image](https://github.com/2noise/ChatTTS/assets/100887031/c6809a9f-2585-429e-b9e4-6fc2747fd75b)
huggingface_hub.utils._errors.LocalEntryNotFoundE…