Creating a new classifier using Caffe?

[Mike Smith]

I've tested out Caffe using the prebuilt networks, and would like to now use Caffe to test cats vs dogs. I can't figure out though how to use the prebuilt networks for a different task than just categorizing for the 1000 categories from ilsvrc. I looked at the finetuning tutorial on Caffe's site but am still confused. Lets say I have a dataset of cats and dogs, and want to output whether an input image is a cat or a dog, using the prebuilt networks. Do I just change the last layer of the network then? Is there a hand on tutorial with code or examples for this? Thanks!

[Friendly User]

Yes, make the last layer output 2 'classes'

Then make class index last row in train_index.txt and test_index.txt

dog235235.png   0

dog135235.png   0

cat235235.png   1

cat135235.png   1

dog435235.png   0

Did you work through all the existing tutorial and examples?