Ubuntu TensorFlow-GPU 1.8.0 with CUDA 9.1, cuDNN 7.0, Python 3.6.5, no AVX

yaroslavvb / tensorflow-community-wheels

Place to upload links to TensorFlow wheels

452 stars 35 forks source link

Ubuntu TensorFlow-GPU 1.8.0 with CUDA 9.1, cuDNN 7.0, Python 3.6.5, no AVX #72

Open laurentS opened 6 years ago

laurentS commented 6 years ago

This was built on a paperspace P5000 instance, with AVX optimisations disabled as they don't support them for now.

laurentS commented 6 years ago

https://github.com/laurentS/tensorflow-wheels/raw/master/tensorflow-1.8.0-cp36-cp36m-linux_x86_64.whl

and the repo: https://github.com/laurentS/tensorflow-wheels

zhanwenchen commented 6 years ago

@laurentS at this point I'm sure that tensorflow-gpu doesn't support anything other than CUDA 9.0. The reason you are not getting errors is probably that you had had CUDA 9.0 before 9.1. Under the hood, even though tf is compiled with 9.1, it is actually using 9.0 because "9.0" is hard-coded.