8 сент 2016 г. - Ultimate tensorization: Compressing convolutional and FC layers alike (CoRR 2016)

Описание:

tensor-train decomposition (TT decomposition)
直接应用 tensor-train decomposition 会导致实验结果不好，我们提出一种 reshape 的方法，把 4 维的卷积核重构为一个多维张量

结果：
We combine the proposed approach with the previous work to compress both convolutional and fully-connected layers of a network and achieve 80× network compression rate with 1.1% accuracy drop on the CIFAR-10 dataset.

Добавлено на ленту времени:

Дата: