nico

paddle的numpy库版本和paddlespeech冲突

8

按照文档，在conda虚拟环境先安装了paddle-gpu版本，再安装了paddlespeech。调用PaddleSpeech/examples/aishell3/vits/local/preprocess.sh预处理shell处理数据时，报错。根据报错应该是paddle的numpy和paddlespeech需要的版本不一致，我也不敢贸然修改numpy的版本，因为可能导致paddle出什么问题。有什么解决方法吗? python:3.9 numpy版本是1.26.2。 paddlepaddle版本: paddlepaddle-gpu 2.5.1.post112 paddlespeech版本:paddlespeech 1.4.1 除了numpy冲突外，还有一个[nltk_data] 错误。 报错如下: File "/home/inspur/ly416100210174/proj/PaddleSpeech/utils/gen_duration_from_textgrid.py", line 18, in import librosa File "/home/inspur/ly416100210174/miniconda3/envs/paddlepaddle/lib/python3.8/site-packages/librosa/__init__.py", line 211, in from . import core File "/home/inspur/ly416100210174/miniconda3/envs/paddlepaddle/lib/python3.8/site-packages/librosa/core/__init__.py", line...

vits训练报错

1

-------------------------------------- C++ Traceback (most recent call last): -------------------------------------- 0 egr::Backward(std::vector const&, std::vector const&, bool) 1 egr::RunBackward(std::vector const&, std::vector const&, bool, bool, std::vector const&, bool, std::vector const&) 2 Conv2dGradNodeFinal::operator()(paddle::small_vector&, bool, bool)...

安装好环境后，运行代码出错

2

outs = net.forward(getOutputsNames(net)) File "F:\刘毅\研究生\机器学习\python\yolov3_with_opencv\object_detection_yolo.py", line 42, in getOutputsNames return [layersNames[i[0] - 1] for i in net.getUnconnectedOutLayers()] File "F:\刘毅\研究生\机器学习\python\yolov3_with_opencv\object_detection_yolo.py", line 42, in return [layersNames[i[0] - 1] for i in net.getUnconnectedOutLayers()]...