Neural Discrete Representation Learning

预备知识

核心思想

20250310215306

: straight-through estimator (STE):

z_q = z + (z_q - z).detach()

参考文献

  1. van den Oord A., Vinyals O. and Kavukcuoglu K. Neural Discrete Representation Learning. NeurIPS, 2017. [PDF] [Code]