Skip to content

Instantly share code, notes, and snippets.

@TimSC

TimSC/tensorflow-gpu-environment.md

Last active April 15, 2020 21:47
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save TimSC/2fee3bf32bac622b1160b361b774f2e2 to your computer and use it in GitHub Desktop.
Save TimSC/2fee3bf32bac622b1160b361b774f2e2 to your computer and use it in GitHub Desktop.
Download ZIP
Working tensorflow GPU environments (Ubuntu 18.04 with Geforce RTX 2060 super)
Raw
tensorflow-gpu-environment.md

Working environments (Ubuntu 18.04 with Geforce RTX 2060 super)

tensorflow-gpu-1.13 and after require TF_FORCE_GPU_ALLOW_GROWTH work around tensorflow/tensorflow#24496

Keras-2.2.5
tensorflow-gpu-1.12.0 
libcudnn7 7.6.5.32-1+cuda10.0
cuda-libraries-10-0
nvidia driver 435

import os
os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'
keras-2.3.1
tensorflow-gpu-1.13.2
libcudnn7 7.6.5.32-1+cuda10.0
cuda-libraries-10-0
nvidia driver 435

import os
os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'
keras-2.3.1
tensorflow-gpu-2.0.1
libcudnn7 7.6.5.32-1+cuda10.0
cuda-libraries-10-0
nvidia driver 435

import os
os.environ['TF_FORCE_GPU_ALLOW_GROWTH'] = 'true'
keras-2.3.1
tensorflow-gpu-2.1.0
libcudnn7 7.6.5.32-1+cuda10.2
cuda-libraries-10-2
nvidia driver 440
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment