Zack Ziegler

Everything’s in nats by default, so you’ll have to convert from base e to base 2 to get bits.

Hi, This is a preprocessing choice we inherited from previous summarization work with OpenNMT, which found that the first 400 tokens is often plenty to compose a good summary. That...

Weird model! If you’re doing the usual VQ-VAE thing with the straight through estimator, could there be some kind of accumulation of error? I can’t picture the model exactly, but...