Thank you for this video, however, please don't call your variable ŝ 😆 (or at least don't say it out loud)
@jonathanyang2359
3 жыл бұрын
Thanks! I don't attend this institution, but this was an extremely clear lecture :)
@sdfrtyhfds
3 жыл бұрын
do you train the pixel cnn on the same data and just not update the Vae weights while training?
@davidmcallester4973
3 жыл бұрын
yes, the vector quantization is held constant as the pixel CNN is trained.
@LyoshaTheZebra
3 жыл бұрын
Thanks for explaining that! Great job. Subscribed!
@K3pukk4
11 ай бұрын
what a legend!
@sdfrtyhfds
3 жыл бұрын
also, what if you skip the quantization during inference? would you still get images that make sense?
@davidmcallester4973
3 жыл бұрын
Do you mean "during generation"? During generation you can't skip the quantization because the pixel-CNN is defined to generate the quantized vectors (the symbols).
@sdfrtyhfds
3 жыл бұрын
@@davidmcallester4973 I guess that during generation it wouldn't make much sense, i was thinking more in the direction of interpolating smoothly between two different symbols.
Пікірлер: 9