We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
fixed typo
Added Inception21k model converted from MXNet
Added the official windows port
Soft Link Creation for Ubuntu 16.04 had wrong syntax
Updated Ubuntu 16.04 or 15.10 Installation Guide (markdown)
Added paper Neural Activation Constellations: Unsupervised Part Model Discovery with Convolutional Networks
Updated Model Zoo (markdown)
Destroyed Publications (markdown)
Updated CNN Salient Object Subitizing Models
Updated GeForce GTX 1080, CUDA 8.0, Ubuntu 16.04, Caffe (markdown)
I've gone through a test on a clean Mac with a Titan X and you'll probably find these up-to-date instructions the most helpful.
My intention with this edit is to add a bit of clarity to the section where the libhdf5 symbolic links are discussed. On my intitial reading of the document it wasn't clear to me that the process needed to be performed for Ubuntu 16.04, i.e. I thought that the section was exclusive to 15.10 at first. Hopefully I've made the process more clear through my edits.
This page seems outdated or at least, contrasts with the 'new' https://github.com/BVLC/caffe/issues/1245 and https://github.com/BVLC/caffe/pull/3211 i.e. https://github.com/BVLC/caffe/blob/master/models/bvlc_reference_caffenet/deploy.prototxt Which is not to say that this page doesn't work anymore, it does for the example given.
Updated GeForce GTX 1080 CUDA 8.0 Ubuntu 16.04 Caffe (markdown)
Created GeForce GTX 1080 + CUDA 8.0 + Ubuntu 16.04 + Caffe (markdown)
Added Pose-Aware CNN Models from CVPR16
Updated AWS EC2 GPU enabled Caffe AMI (markdown)
Created AWS EC2 GPU-enabled Caffe AMI (markdown)
link Farsi translation from the community