Have seen someone used even 1000 hidden layers in Convolutional Neural Network (CNN), and could that be reduced to 999 or even less? Is there any theoretical guidance? Does an optimal number of hidden layers exist?
Roman Liessner
Usually algorithms are used for hyper parameter optimization. These algorithms determine and evaluate different configurations.
Here is an example library: https://scikit-optimize.github.io/
0 votes
0 thanks
- Badges
- Science topic
More Daqing Chen's questions See All
Similar questions and discussions