Alec Radford

Great question, we looked into this and related objectives a bit during preliminary research before settling on the training objective described in the paper. We have to at least run...

Have you tried using sigmoid generator output instead? When you train a DCGAN on MNIST with sigmoid outputs it's pretty much always railed binary like you expect.

format is (batch, channel, height, width) also called `bc01` in theano docs.

@dribnet The generator/sample code currently uses the minibatch to calculate statistics for batchnorm. This is the reason why changing the scale and mean of sampled Zs has no effect -...

Apolgoies for the delay, did not know ipython notebook and dragged my feet on this. @udibr Madison did some wording editing on top of these notebooks and if you'd like...

Pretty sure I renamed the file names in the theano tutorials locally which caused some issues while these names correspond to the ones that are directly from downloading from Yann's...

Agreed, an example would be nice - setting the seq_output argument to any of the recurrent layers to true should work but I've only tested this for language modeling with...

Update on this, clean output sequence support starts to get into a rabbit hole of re-factoring and/or interface ugliness that's still being figured out. Alpha support is working on the...