Alex Wiltschko
Alex Wiltschko
Is `input` always a torch tensor or scalar? You're returning a scalar, which might cause weirdness.
Try `torch.le(input, 0).value` instead of `torch.le(input, 0)`
Does it work on plain NN, on the CPU? On Sun, Jun 19, 2016 at 7:10 AM David Macêdo [email protected] wrote: > I am using cuddn... > > Input and...
Thanks for the heads up. We tried to stay away from the more "exotic" call-signatures in torch, but I think there's no reason we couldn't accept a PR implementing this...
That's pretty strange. I don't see any of the usual culprits. cc @luketwitter
I think you'd want something like this, outside of your loss function: ``` lua -- Build your own custom module mymodule = {} -- Make a sort function in your...
This is a great idea. We've been absorbed making it work at all. I think a screencast and some schematics would go a long way to making the internals clear....
Oh without `optimized=true`, this will never work, the assignment functionality is limited to optimized mode. Then the issue is that bug. One thing that I'm wary of is that you're...
Awesome, thanks. Can you open a PR, and I'll merge it right quick? On Fri, May 27, 2016 at 4:34 PM Bart van Merriënboer < [email protected]> wrote: > Making a...
Hrm, seems like we should be doing table lookups using brackets, and not dot, syntax.