Why change from L-BFGS to vanilla SGD? Didatical purposes?
I rewrote the old Keras neural style transfer code example to modern Keras. I switched it from using scipy's L-BFGS to vanilla SGD -- the whole thing is just a few lines of pure Keras/TF now. https://colab.research.google.com/drive/1H-lNMxVF9NPME6oZVphFRoi7yglJMZxP …pic.twitter.com/Rn53EaGvhz
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thank you I've been wanting to try to put this on mobile for some time. Think I need to re write it for JS though.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thank you for this!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is really cool, thank you! Any tricks to avoid OOM errors when bumping size to 1024 (with appropriately sized source images)?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
This is great! I just got this running on my rinky-dink 1050ti and it's really incredible how quickly Paris starts to transform. This is after only 400 iterations:pic.twitter.com/80jrhLjygV
-
Can you please explain how you let it run on your GPU? I have a strong background in Matlab, but I am fairly new to python and would like to learn. Any advise is appreciated!
- Show replies
New conversation -
-
-
Perhaps a better example than the Paris skyline:pic.twitter.com/Huo6IRk2zn
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
If I didn't follow you here, how would I find these? Thanks!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.