well done! Thanks for the tutorial!
New tutorial: building autoencoders in Keras. http://blog.keras.io/building-autoencoders-in-keras.html …pic.twitter.com/ehYr5cGutN
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
awesome!!!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
really good tutorial. Is there a way to do layer-wise pretraining (i.e. fixing the weights of lower layers) in keras?
-
sure, you can always set "trainable=False" on a layer then compile your model. It freezes the layer's weights
- Show replies
New conversation -
-
-
Good stuff, thanks!
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
nice post. suggestion might be to use inverse cdf of gaussian of the linspace prior to producing z_sample.https://twitter.com/dribnet/status/696797657809813504 …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
excellent tutorial. wish this had been my intro to VAEs, you explain and demo the concept very clearly. do jigsaw next ;)
-
minor comment: t-SNE usually preceded 30+ dim PCA on data, only "genuinely" high rank data would profit from something like AE
- Show replies
New conversation -
-
-
A minor correction: autoencoders are not in Ng's Coursera ML class AFAIR, just in http://ufldl.stanford.edu/tutorial/unsupervised/Autoencoders/ …
-
are they not? I have a vivid memory of checking out his class in 2012 and seeing AEs. Could it be something else?
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.