Note: This is already a very old caffe version. Please check the caffe-multigpu repo for a new version based on caffe-rc (https://github.com/BVLC/caffe/releases/tag/rc). Multilabel caffe, with multiple label outputs on dataLayer. No need to modify Loss layer, each loss is dealed separately.
I found that it writes "layers" in the "imagenet_finetune_overfeat_train.prototxt" while in other caffe net prototxt file it writes "layer". I don't know what's the difference about "layer" and "layers". For example,
in the caffe official documents, it shows: layer { name: "relu1" type: "ReLU" bottom: "ip1" top: "ip1" }
but in the imagenet_finetune_overfeat_train.prototxt, it shows layers { name: "relu3" type: RELU bottom: "conv3" top: "conv3" }
I really want to know what's the difference between them. Any thoughts would be greatly appreciated!
Thanks.