Search results
Stable Low-Rank Tensor Decomposition for Compression of Convolutional Neural Network. Anh-Huy Phan, Konstantin Sobolev, + 7. August 2020Computer Vision – ECCV 2020 https://doi.org/10.1007/978-3-030-58526-6_31. View all Publications.
Aug 12, 2020 · We present a novel method, which can stabilize the low-rank approximation of convolutional kernels and ensure efficient compression while preserving the high-quality performance of the neural networks.
- Anh-Huy Phan, Konstantin Sobolev, Konstantin Sozykin, Dmitry Ermilov, Julia Gusak, Petr Tichavsky, V...
- arXiv:2008.05441 [cs.CV]
- 2020
- This paper is accepted to ECCV2020
Valeriy Glukhov | Papers With Code. Search Results for author: Valeriy Glukhov. Found 1 papers, 0 papers with code. Date Published. Stable Low-rank Tensor Decomposition for Compression of Convolutional Neural Network.
glukhov.valery@huawei.com Abstract. Most state-of-the-art deep neural networks are overparameterized and exhibit a high computational cost. A straightforward approach to this problem is to replace convolutional kernels with its low-rank tensor approximations, whereas the Canonical Polyadic tensor Decomposition is one of the most suited models.
This paper is the first study on degeneracy in the tensor decomposition of convolutional kernels. We present a novel method, which can stabilize the low-rank approximation of con-volutional kernels and ensure efficient compression while preserving the high-quality performance of the neural networks.
This paper is the first study on degeneracy in the tensor decomposition of convolutional kernels. We present a novel method, which can stabilize the low-rank approximation of convolutional kernels and ensure efficient compression while preserving the high-quality performance of the neural networks.
Read Valeriy Glukhov's latest research, browse their coauthor's research, and play around with their algorithms.
