(PreconditionNotMet) Cannot load cudnn shared library. Cannot invoke method cudnnGetVersion.

[BUAACY]

【处理中】开启提取视频关键帧...
【处理中】开始提取字幕信息，此步骤可能花费较长时间，请耐心等待...

(PreconditionNotMet) Cannot load cudnn shared library. Cannot invoke method cudnnGetVersion. | 0/204 [00:00<?, ?it/s]
[Hint: cudnn_dso_handle should not be null.] (at ../paddle/phi/backends/dynload/cudnn.cc:60)

0%|▋ | 1/204 [00:01<05:25, 1.60s/it]

[VeteranChauffeur]

too

[VeteranChauffeur]

python ./backend/main.py --use_gpu False --show_log True

请输入视频完整路径：/home/video-subtitle-extractor-main/1.mp4
请指定字幕区域 (ymin ymax xmin xmax)：832 923 183 1280
识别字幕语言：ch
识别模式：fast
使用GPU进行加速
帧数：1200.0，帧率：30.0
V4-ch_det_fast
V4-ch_rec_fast
【处理中】开启提取视频关键帧...
16:20:29: Logging wxScroll1
16:20:29: Starting program...
16:20:29: CVideoSubFinderApp::OnInitCmdLine...
m_VC.isOpened(): 1 path: /home/video-subtitle-extractor-main/1.mp4 No such file or directory
【处理中】开始提取字幕信息，此步骤可能花费较长时间，请耐心等待...
(PreconditionNotMet) Cannot load cudnn shared library. Cannot invoke method cudnnGetVersion.
[Hint: cudnn_dso_handle should not be null.] (at ../paddle/phi/backends/dynload/cudnn.cc:64)

4%|█▊ | 53/1200 [00:04<01:38, 11.63it/s]

都指定不要GPU了，还是会走到GPU相关代码

[YaoFANGUK]

你安装的cuda和cudnn版本可以说下吗

[VeteranChauffeur]

wget https://developer.nvidia.com/downloads/compute/cudnn/secure/8.9.7/local_installers/12.x/cudnn-local-repo-rhel7-8.9.7.29-1.0-1.x86_64.rpm
yum localinstall cudnn-local-repo-rhel7-8.9.7.29-1.0-1.x86_64.rpm

rpm -ql cudnn-local-repo-rhel7-8.9.7.29.x86_64
ls /var/cudnn-local-repo-rhel7-8.9.7.29/

yum localinstall -y /var/cudnn-local-repo-rhel7-8.9.7.29/libcudnn8-8.9.7.29-1.cuda12.2.x86_64.rpm
yum localinstall -y /var/cudnn-local-repo-rhel7-8.9.7.29/libcudnn8-devel-8.9.7.29-1.cuda12.2.x86_64.rpm
yum localinstall -y /var/cudnn-local-repo-rhel7-8.9.7.29/libcudnn8-samples-8.9.7.29-1.cuda12.2.x86_64.rpm

ls /usr/local/cuda/include/ | grep cudnn
ls /usr/local/cuda/lib64/ | grep libcudnn

mkdir ~/cudnn_temp
cd ~/cudnn_temp
rpm2cpio /var/cudnn-local-repo-rhel7-8.9.7.29/libcudnn8-8.9.7.29-1.cuda12.2.x86_64.rpm | cpio -idmv
rpm2cpio /var/cudnn-local-repo-rhel7-8.9.7.29/libcudnn8-devel-8.9.7.29-1.cuda12.2.x86_64.rpm | cpio -idmv

cp usr/include/cudnn*.h /usr/local/cuda/include/
cp usr/lib64/libcudnn* /usr/local/cuda/lib64/
chmod a+r /usr/local/cuda/include/cudnn*.h /usr/local/cuda/lib64/libcudnn*

ls /usr/local/cuda/include/ | grep cudnn
ls /usr/local/cuda/lib64/ | grep libcudnn

echo 'export PATH=/usr/local/cuda/bin:$PATH' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
source ~/.bashrc

重新装cudnn就可以了

不过请教一下，大佬，后面可能需要部署cpu的版本，传--use_gpu False 是不是就不应该报Cannot invoke method cudnnGetVersion错呀