Under normal circumstances, I can save a ComputationGraph (a Convolutional Neural Network) to a file and load it in a later run and it works fine.
However, when I include it in a jar and I try to load it, it fails.
Is there some way to load ComputationGraph objects from inside of a Jar as a resource?
I got it working for the Word2Vec objects, but it fails on ComputationGraph objects.
Thank you
It's just a zip file (or should be!). I'm not sure if you're using our ModelSerializer (this is the only thing you should be using for computation graph or multi layer network) - you can use an InputStream or a File when you call:
ModelServer.restoreComputationGraph
on an InputStream. That InputStream could be a classpath resource input stream (using something like spring's ClassPathResource) or calling a YourClass.class.getResourceAsStream(..) - any of those should work fine for you.