Skip to first unread message
Eren Gölge
Nov 30, 2014, 3:56:37 PM11/30/14
to caffe...@googlegroups.com
As I understand from http://caffe.berkeleyvision.org/tutorial/layers.html it is possible to provide raw images to caffe with a file that includes image_names and labels. Since I try to use caffe as a AE, what type of label information should I give? And could you explain this file format a bit more to include raw images in the caffe training?
Evan Shelhamer
Nov 30, 2014, 4:14:58 PM11/30/14
to Eren Gölge, caffe...@googlegroups.com
You could start from the MNIST autoencoder example's model definition and solver. For an IMAGE_DATA layer you can just pass dummy "labels" of 0 for an autoencoder that only uses the image data.
Evan Shelhamer
On Sun, Nov 30, 2014 at 3:56 PM, Eren Gölge <eren...@gmail.com> wrote:
As I understand from http://caffe.berkeleyvision.org/tutorial/layers.html it is possible to provide raw images to caffe with a file that includes image_names and labels. Since I try to use caffe as a AE, what type of label information should I give? And could you explain this file format a bit more to include raw images in the caffe training?
--
You received this message because you are subscribed to the Google Groups "Caffe Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to caffe-users...@googlegroups.com.
To post to this group, send email to caffe...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/caffe-users/32e15a93-9873-4c57-a5de-04a1ef4636e8%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages