TensorFlow 1.4.1 (GPU, CUDA 9.1)
amberfly
released this
14 Dec 22:28
·
11 commits
to master
since this release
GPU wheels for TensorFlow 1.4.1, built by TinyMind, the cloud machine learning platform.
To use the wheels on your own machine, Intel Broadwell or later CPU, and Nvidia computing capability 3.7 or later GPU with CUDA 9.1 and CuDNN 7 are required.
This version is optimized for compute capabilities 3.7 (K80, AWS P2/GCP), 6.0 (P100, GCP) and 7.0 (V100, AWS P3).
NOTE: These wheels contain MKL support. If you don't have it, install MKL by following the instructions here. You may want to take a look at the performance guide for MKL.