When the installation tensorflow, tried many methods, and finally get a command.
Let me talk about issues now arise:
- First installed TensorFlow CPU version
appeared the following issues, mainly due to the CPU is an early result, only install an earlier version tensorflow.
Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX AVX2
- GPU version installed
consideration to a server on the block P100 graphics card, do not it is a throwaway, or installed directly GPU version of it.
Ever since the emergence of the following issues: The main problem is the inconsistency tensorflow-GPU version of CUDA and CUDNN version.
2019-10-15 19:19:41.440285: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2019-10-15 19:19:41.465433: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1006] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-10-15 19:19:41.465758: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1618] Found device 0 with properties:
Using the following command to find CUDA version and CUDNN
cat /usr/local/cuda/version.txt
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
- Later, with anaconda, a python entered the current environment, direct mounting conda, a successful step up ...
conda install tensorflow-gpu
