search

yeyupiaoling / PPASR

基于PaddlePaddle实现端到端中文语音识别，从入门到实战，超简单的入门案例，超实用的企业项目。支持当前最流行的DeepSpeech2、Conformer、Squeezeformer模型
Apache License 2.0
797 stars 131 forks source link

在linux上 使用本地预测遇到CUBLAS error(13) #156

Closed navy7913 closed 1 year ago

navy7913 commented 1 year ago

環境: gpu 3060ti ubuntu 18.04 Driver version 510 cuda 11.8 cudnn 8.4 下载的环境是paddlepaddle-gpu==2.4.1 cudatoolkit=11.2

在linux上使用infer_path出现CUBLAS error(13)的错误，想请教如何解决这个问题，错误提示请参考下图 image

yeyupiaoling commented 1 year ago

你检查过你的环境能正常运行起来吗？有在这个环境跑过其他的项目吗？我觉得可能是cudnn和cuda问题

navy7913 commented 1 year ago

您好我有重新安装新的cuda和cudnn环境(参考图一)，安装完成后错误提示变成图二所示，想请教您会出现这个错误是因为版本和paddlepaddle-gpu不匹配所导致的吗?

環境: gpu 3060ti ubuntu 18.04 conda install paddlepaddle-gpu==2.4.1 cudatoolkit=11.7

image 图一

image 图二

RuntimeError: (PreconditionNotMet) The third-party dynamic library (libcufft.so) that Paddle depends on is not configured correctly. (error code is libcufft.so: cannot open shared object file: No such file or directory) Suggestions:

  1. Check if the third-party dynamic library (e.g. CUDA, CUDNN) is installed correctly and its version is matched with paddlepaddle you installed.
  2. Configure third-party dynamic library environment variables as follows:
    • Linux: set LD_LIBRARY_PATH by export LD_LIBRARY_PATH=...
    • Windows: set PATH by `set PATH=XXX; (at /paddle/paddle/phi/backends/dynload/dynamic_loader.cc:305)
navy7913 commented 1 year ago

您好，我将paddlepaddle-gpu换成这个版本就可以成功运行了 conda install paddlepaddle-gpu==2.3.2 cudatoolkit=11.6 -c https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/Paddle/ -c conda-forge