Convolutional Neural Network with Optimized Learning Parameters
The error of training and testing of convolutional neural network when using SGD, Adam, AdaGrad, Nadam, RMSprop training optimizers, which was implemented using the Keras library in the Python programming language, is analyzed. An analysis of the learning error function was carried out, and Fourier spectra and branching diagrams were obtained for it.
Authors
- I. Katerynchuk (ORCID: https://orcid.org/0000-0001-8877-8324)
- I. Kunyo
- Volodymyr Brygilevych
- Volodymyr Franiv (ORCID: https://orcid.org/0000-0001-9856-1962)
- S. Sveleba
- Oleksandr Krychevets
Institutions
- Lviv University (UA)
- State University of Applied Sciences in Jaroslaw (PL)
Publication Details
- Published
- 2024-09-19
- DOI
- https://doi.org/10.1109/acit62333.2024.10712630
- Primary Topic
- Neural Networks and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00