caffe

Maybe it would be worth looking into something like Caffe [1], it may even help you with managing the site in general.

It seems to me like they should also switch from g2.8xlarge to g2.2xlarge because caffe does not benefit much from multiple GPUs.

If someone wanted faster results and full access to edit the python code and caffe models, this would be the better option.

So I just tried to start from the Caffe instructions and then fix things that broke.

It should get you all the way to a working Caffe install.

Update: The author of Caffe just got back to me with word that the filters are in public domain, therefore safe to use as design elements.

I'm curious what hardware you have running caffe and the deepdream code for this?

Anyone used deep learning (caffe maybe?) for OCR?

There are roughly two possibilities to do this (since Caffe does not have a lookup table a la Torch yet, though there is a pull request for this): 1.

I love the fact this is using caffe by Berkeley.

Edit for clarification: Caffe is quite different from the other libraries --- you do not set up a network programmatically, but specify the topology as data (via a protocol buffer file).

Yes, even we found Caffe to be overwhelming.

Maybe it's protected by "caffe culture", but I seriously doubt that drinking half a liter caffe was ever part of any culture.

You can download the file "synset_words.txt" with class numbers in caffe via running the script./data/ilsvrc12/get_ilsvrc_aux.sh.

Caffe, for example, calls `exit` every time an error occurs, including recoverable ones like files which do not exist.

Caffe has been used for handwriting, seems like OCR of typefaces would work just the same with a typeface dataset.

Caffe has a lot of good ideas that are worth extending especially to the industry.