XXXXRT666
XXXXRT666
After gradually troubleshooting, I think that this seems to be related to `return self.model.encode(features, to_cpu=to_cpu)`, which is located in `encode` within `faster_whisper.WhisperModel.transcribe` and calls `ctranslate2`. However, the error occurs at...
I haven't encountered this error myself. I am using an Intel Mac from a QQ group member who has experienced this error. I am using a modified `transcribe.py` with added...
> > 不知道有没有人测试过，比如一句30s的音频，是做成1句30s，还是3句10s效果更好？也就是语音切分有没有什么讲究？ > > > > 在3s-10s之间选择，尽量每段音频的语速和情绪是一致的，例如一段悲伤的语音就不要包含转换成欢快情绪的片段了，最好不要有一句话说了一半就被截断的情况。超出10s没有意义，很多样本会直接被弃，根本不会生成频谱模型文件，也就不会被纳入到训练过程中。 你说的这个是参考，训练最长可以用接近一分钟的
使用conda安装faiss-cpu
`python webui.py `
export LD_LIBRARY_PATH
it doesn't work on PyTorch 2.2.1 either, I can see clearly from the Activity Monitor that the CPU and GPU usage is 0
> @XXXXRT666 Can you give me more details? Yes, of course. I run it on python 3.9(anaconda) and python3.11(3.11 requirements installed). After I click the convert button, It does do...
setting index ratio zero or just not using Also, memory not released after unloading the model, I think it is related to PyTorch the program stuck at `score, ix =...