Use with cudnn v6 and v7

TimoSaemann / caffe-segnet-cudnn5

This repository was a fork of BVLC/caffe and includes the upsample, bn, dense_image_data and softmax_with_loss (with class weighting) layers of caffe-segnet (https://github.com/alexgkendall/caffe-segnet) to run SegNet with cuDNN version 5.

Other

176 stars 127 forks source link

Use with cudnn v6 and v7 #17

Open gsedej opened 7 years ago

gsedej commented 7 years ago

Hello

I can use remote nvidia-gpu system that I do not have root and can't install software (so apt-get or dpkg). The system admin installed then-lastest version of drivers, cuda8 and cudnn v6. (now the v7 is already out).

The version on system is v6 and admin won't downgrade to v5 just because of me (multiple users). Is there any way to install local cudnn and use PATH or some other variable? (ubuntu 16.04 is installed)

kertansul commented 6 years ago

I replaced the include/caffe/util/cudnn.hpp with the one in BVLC/caffe just created a pull request https://github.com/TimoSaemann/caffe-segnet-cudnn5/pull/20