Matches in SemOpenAlex for { <https://semopenalex.org/work/W4322760368> ?p ?o ?g. }
Showing items 1 to 69 of
69
with 100 items per page.
- W4322760368 abstract "The past decade has witnessed a drastic increase in modern deep neural networks (DNNs) size, especially for generative adversarial networks (GANs). Since GANs usually suffer from high computational complexity, researchers have shown an increased interest in applying pruning methods to reduce the training and inference costs of GANs. Among different pruning methods invented for supervised learning, dynamic sparse training (DST) has gained increasing attention recently as it enjoys excellent training efficiency with comparable performance to post-hoc pruning. Hence, applying DST on GANs, where we train a sparse GAN with a fixed parameter count throughout training, seems to be a good candidate for reducing GAN training costs. However, a few challenges, including the degrading training instability, emerge due to the adversarial nature of GANs. Hence, we introduce a quantity called balance ratio (BR) to quantify the balance of the generator and the discriminator. We conduct a series of experiments to show the importance of BR in understanding sparse GAN training. Building upon single dynamic sparse training (SDST), where only the generator is adjusted during training, we propose double dynamic sparse training (DDST) to control the BR during GAN training. Empirically, DDST automatically determines the density of the discriminator and greatly boosts the performance of sparse GANs on multiple datasets." @default.
- W4322760368 created "2023-03-03" @default.
- W4322760368 creator A5000101445 @default.
- W4322760368 creator A5029998748 @default.
- W4322760368 creator A5074954379 @default.
- W4322760368 creator A5089713542 @default.
- W4322760368 date "2023-02-28" @default.
- W4322760368 modified "2023-09-27" @default.
- W4322760368 title "Double Dynamic Sparse Training for GANs" @default.
- W4322760368 doi "https://doi.org/10.48550/arxiv.2302.14670" @default.
- W4322760368 hasPublicationYear "2023" @default.
- W4322760368 type Work @default.
- W4322760368 citedByCount "0" @default.
- W4322760368 crossrefType "posted-content" @default.
- W4322760368 hasAuthorship W4322760368A5000101445 @default.
- W4322760368 hasAuthorship W4322760368A5029998748 @default.
- W4322760368 hasAuthorship W4322760368A5074954379 @default.
- W4322760368 hasAuthorship W4322760368A5089713542 @default.
- W4322760368 hasBestOaLocation W43227603681 @default.
- W4322760368 hasConcept C108010975 @default.
- W4322760368 hasConcept C119857082 @default.
- W4322760368 hasConcept C121332964 @default.
- W4322760368 hasConcept C153180895 @default.
- W4322760368 hasConcept C153294291 @default.
- W4322760368 hasConcept C154945302 @default.
- W4322760368 hasConcept C163258240 @default.
- W4322760368 hasConcept C2776214188 @default.
- W4322760368 hasConcept C2777211547 @default.
- W4322760368 hasConcept C2779803651 @default.
- W4322760368 hasConcept C2780992000 @default.
- W4322760368 hasConcept C41008148 @default.
- W4322760368 hasConcept C62520636 @default.
- W4322760368 hasConcept C6557445 @default.
- W4322760368 hasConcept C76155785 @default.
- W4322760368 hasConcept C86803240 @default.
- W4322760368 hasConcept C94915269 @default.
- W4322760368 hasConceptScore W4322760368C108010975 @default.
- W4322760368 hasConceptScore W4322760368C119857082 @default.
- W4322760368 hasConceptScore W4322760368C121332964 @default.
- W4322760368 hasConceptScore W4322760368C153180895 @default.
- W4322760368 hasConceptScore W4322760368C153294291 @default.
- W4322760368 hasConceptScore W4322760368C154945302 @default.
- W4322760368 hasConceptScore W4322760368C163258240 @default.
- W4322760368 hasConceptScore W4322760368C2776214188 @default.
- W4322760368 hasConceptScore W4322760368C2777211547 @default.
- W4322760368 hasConceptScore W4322760368C2779803651 @default.
- W4322760368 hasConceptScore W4322760368C2780992000 @default.
- W4322760368 hasConceptScore W4322760368C41008148 @default.
- W4322760368 hasConceptScore W4322760368C62520636 @default.
- W4322760368 hasConceptScore W4322760368C6557445 @default.
- W4322760368 hasConceptScore W4322760368C76155785 @default.
- W4322760368 hasConceptScore W4322760368C86803240 @default.
- W4322760368 hasConceptScore W4322760368C94915269 @default.
- W4322760368 hasLocation W43227603681 @default.
- W4322760368 hasOpenAccess W4322760368 @default.
- W4322760368 hasPrimaryLocation W43227603681 @default.
- W4322760368 hasRelatedWork W2511279186 @default.
- W4322760368 hasRelatedWork W2884581909 @default.
- W4322760368 hasRelatedWork W2950378732 @default.
- W4322760368 hasRelatedWork W2998257508 @default.
- W4322760368 hasRelatedWork W3190733223 @default.
- W4322760368 hasRelatedWork W3195426195 @default.
- W4322760368 hasRelatedWork W3199608561 @default.
- W4322760368 hasRelatedWork W4287020189 @default.
- W4322760368 hasRelatedWork W4287728236 @default.
- W4322760368 hasRelatedWork W4301908144 @default.
- W4322760368 isParatext "false" @default.
- W4322760368 isRetracted "false" @default.
- W4322760368 workType "article" @default.