search

mind / wheels

Performance-optimized wheels for TensorFlow (SSE, AVX, FMA, XLA, MPI)
886 stars 108 forks source link

Loaded runtime CuDNN library: 7101 (compatibility version 7100) #33

Closed secretbatcave closed 6 years ago

secretbatcave commented 6 years ago

Hello there,

firstly, thank you for producing these wheels, they are spectacularly useful. However I have run into a small compatibility issue. Nvidia have shipped cuDNN 7.1 in their cuda docker images, which is incompatible with your current 1.6 release. Running it yields:

E tensorflow/stream_executor/cuda/cuda_dnn.cc:378] Loaded runtime CuDNN library: 7101 (compatibility version 7100) but source was compiled with 7005 (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.

If you would be so kind as to do a release built with cuDNN 7.1, that would be most wonderful.

danqing commented 6 years ago

Will provide with TF 1.7. Thanks for reporting!

yangkang779 commented 6 years ago

thanks,How long does it take to update? (tf1.7)

secretbatcave commented 6 years ago

@danqing Wonderful news!

danqing commented 6 years ago

https://github.com/mind/wheels/releases/tag/tf1.7-gpu-cuda91-nomkl

(python 3.6 coming soon)