Open AccessPerformance analysis of the TensorFlow library with different optimisation algorithms
Author(s) -
Maciej Wadas,
Jakub Smółka
Publication year - 2021
Publication title -
journal of computer sciences institute
Language(s) - English
Resource type - Journals
ISSN - 2544-0764
DOI - 10.35784/jcsi.2738
Subject(s) - artificial neural network , computer science , artificial intelligence , graphics , algorithm , machine learning , component (thermodynamics) , deep learning , pattern recognition (psychology) , computer graphics (images) , physics , thermodynamics
This paper presents the results of performance analysis of the Tensorflow library used in machine learning and deep neural networks. The analysis focuses on comparing the parameters obtained when training the neural network model for optimization algorithms: Adam, Nadam, AdaMax, AdaDelta, AdaGrad. Special attention has been paid to the differences between the training efficiency on tasks using microprocessor and graphics card. For the study, neural network models were created in order to recognise Polish handwritten characters. The results obtained showed that the most efficient algorithm is AdaMax, while the computer component used during the research only affects the training time of the neural network model used.