Συγκριτική μελέτη εναλλακτικής υλοποίησης βαθιάς μάθησης σε παράλληλα περιβάλλοντα
Comparative study of alternative implementation of depth learning in parallel environments
View/
Keywords
Βαθιά μηχανική μάθηση ; Βαθιά μάθηση ; Βαθιά νευρωνικά δίκτυα ; Αρχιτεκτονικές μάθησης ; Deep learning algorithms ; Deep learning ; Deep neural networksAbstract
This paper explores comparatively the differences in alternative implementation of Deep Learning algorithms in parallel environments. We are in the third wave with Convolution Neural Networks and Big Data presenting and developing hardware technology with GPU graphics processors. Because of the latter, we can investigate the Neural Networks of Deep Learning as the basic functions on these cards are very fast in relation to common CPUs. We review the most active frameworks and choose the most appropriate ones based on speed, reliability, portability, and software implementation time.
Postgraduate Studies Programme
ΠληροφορικήDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα ΠληροφορικήςNumber of pages
65Language
GreekCollections
Except where otherwise noted, this item's license is described as
Attribution-NonCommercial-NoDerivatives 4.0 Διεθνές
Attribution-NonCommercial-NoDerivatives 4.0 Διεθνές