Stefan -- That makes sense, although I've noticed that once it runs out of memory with a failure message, it seems like even a simpler network (that would succeed if I ran it first thing) can fail and crash the kernel, so there may also be an issue for how the kernel recovers state & resources after an out of memory condition (in addition to the problem you mentioned where it's internal checks on memory requirements fail).

PS The 970 (and possibly 4GB 960s) do have that weird Nvidia memory issue, where 4GB isn't really available. I have no idea whether that might confuse an application, but it is something memory-related that's unique to these cards. (Nvidia is having to send $30 to every 970 owner as a result -- not the 960, since Nvidia never configured it with 4GB, those were done by board vendors aftermarket).

I also just noticed that sometimes the first time I run the code I get GPU Memory exhausted, and then if I try to simplify it and run again, the kernel often crashes. So perhaps there is something flawed with the way the kernel recovers from the OOM condition.

When I run the example on my machine (which has a GeForce 960 GTX) it returns with a Failure answer, saying that GPU memory was exhausted. Perhaps that is also happening on your machine but instead of properly bailing and returning a Failure the kernel is crashing...

I ripped the sample down to nothing, and built it back up. From what I can tell so far, the problem arises with the 512 neuron DotPlusLayer. If I take that down to 256 or less, I can train the network.

Is there a way to access your training data? I.e. the files that are in "c:\\projects\\Submissions\\Thumbnails\\Selects", referenced in the notebook.