Using data layer instead of dense image data layer

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

Using data layer instead of dense image data layer #6

Closed MustafaArikan closed 7 years ago

MustafaArikan commented 7 years ago

Is it possible to use data layer for LMDB input as well. Is that updated for cudnn5 as well ?

thanks in advance

TimoSaemann commented 7 years ago

It supports everything that was supported in Caffe (Dec 2016). So, it should be possible to use another data layer with lmdb as backend.