Loaded runtime CuDNN library: 7500 (compatibility version 7500) but source was compiled with 7004

wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ python video_test.py ./data/demo_data/video.mp4
2019-07-26 23:20:59.087061: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-07-26 23:20:59.211459: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:898] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-07-26 23:20:59.211931: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212] Found device 0 with properties:
name: GeForce 930M major: 5 minor: 0 memoryClockRate(GHz): 0.941
pciBusID: 0000:04:00.0
totalMemory: 1.96GiB freeMemory: 1.79GiB
2019-07-26 23:20:59.211955: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1312] Adding visible gpu devices: 0
2019-07-26 23:21:00.990866: I tensorflow/core/common_runtime/gpu/gpu_device.cc:993] Creating TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1558 MB memory) -> physical GPU (device: 0, name: GeForce 930M, pci bus id: 0000:04:00.0, compute capability: 5.0)
2019-07-26 23:21:11.568484: E tensorflow/stream_executor/cuda/cuda_dnn.cc:378] Loaded runtime CuDNN library: 7500 (compatibility version 7500) but source was compiled with 7004 (compatibility version 7000).  If using a binary install, upgrade your CuDNN library to match.  If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration.
2019-07-26 23:21:11.569469: F tensorflow/core/kernels/conv_ops.cc:717] Check failed: stream->parent()->GetConvolveAlgorithms( conv_parameters.ShouldIncludeWinogradNonfusedAlgo<T>(), &algorithms)
Aborted (core dumped)

wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ python video_test.py ./data/demo_data/video.mp4
2019-07-26 23:20:59.087061: I tensorflow/core/platform/cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-07-26 23:20:59.211459: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:898] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-07-26 23:20:59.211931: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1212] Found device 0 with properties:
name: GeForce 930M major: 5 minor: 0 memoryClockRate(GHz): 0.941
pciBusID: 0000:04:00.0
totalMemory: 1.96GiB freeMemory: 1.79GiB
2019-07-26 23:20:59.211955: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1312] Adding visible gpu devices: 0
2019-07-26 23:21:00.990866: I tensorflow/core/common_runtime/gpu/gpu_device.cc:993] Creating TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 1558 MB memory) -> physical GPU (device: 0, name: GeForce 930M, pci bus id: 0000:04:00.0, compute capability: 5.0)
2019-07-26 23:21:11.568484: E tensorflow/stream_executor/cuda/cuda_dnn.cc:378] Loaded runtime CuDNN library: 7500 (compatibility version 7500) but source was compiled with 7004 (compatibility version 7000).  If using a binary install, upgrade your CuDNN library to match.  If building from sources, make sure the library loaded at runtime matches a compatible version specified during compile configuration.
2019-07-26 23:21:11.569469: F tensorflow/core/kernels/conv_ops.cc:717] Check failed: stream->parent()->GetConvolveAlgorithms( conv_parameters.ShouldIncludeWinogradNonfusedAlgo<T>(), &algorithms)
Aborted (core dumped)

该错误为 cuDNN的版本号不匹配，导致命令中止。本程序需要 cuDNN的版本号 7000版本，而本机安装的cuDNN的版本号7500，所以需要卸载cuDNN 7500，安装cuDNN 7000

相关命令：

查看cuda版本 ：      nvcc -V

查看位置  ：            which nvcc

查看NVIDIA动态使用情况：  watch -n 1 nvidia-smi  

cuda 版本    ：    cat /usr/local/cuda/version.txt

cudnn 版本  ：    cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2

NVIDIA 驱动版本  ：  cat /proc/driver/nvidia/version

查看环境变量  :           env

LD_DEBUG=all cat

卸载cuda ：                 sudo  /usr/local/cuda-8.0/bin/uninstall_cuda_8.0.pl

卸载NVIDIA Driver ：   sudo  /usr/bin/nvidia-uninstall

wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ cat /usr/local/cuda/version.txt
CUDA Version 9.0.176
wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 5
#define CUDNN_PATCHLEVEL 0
--
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)

#include "driver_types.h"

CuDNN版本为 7.5.0

卸载cudnn 7.5.0:

wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ sudo rm -rf /usr/local/cuda/include/cudnn.h
wu@wu-X555LF:~/YOLOv3_TensorFlow-master$ sudo rm -rf /usr/local/cuda/lib64/libcudnn*
wu@wu-X555LF:~/YOLOv3_TensorFlow-master$

执行安装命令cudnn 7.0.5 :

 sudo dpkg -i libcudnn7_7.0.5.15-1+cuda9.0_amd64.deb

sudo dpkg -i libcudnn7-dev_7.0.5.15-1+cuda9.0_amd64.deb

sudo dpkg -i libcudnn7-doc_7.0.5.1