a toy example of triplet loss in keras (in 2d for illustration only) https://gist.github.com/kylemcdonald/e91b903332d6a94a5403bca378ac8c2e …pic.twitter.com/yoIIN7HZZF
Show this thread
-
@fchollet it feels like a hack to pass unused zeros from the generator, but i don't see any better way to do thispic.twitter.com/z3YkpiUZwI
Show this thread
model.add_loss(K.mean(...)) model.compile(loss=None, ...)
-
-
much cleaner! it gives me a warning, but still works. it sounds like the output of the shared model is expected to be used in the loss, or it's just reminding me that i should be passing `None` for y.pic.twitter.com/CNZaaJ9Bcb
-
It's a warning that's safe to ignore, it's just to make sure you did it on purpose
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.