March 15, 2024Applied and Computational Engineering

Hyperparameter search and performance assessment of convolutional neural networks for image classification

View Full Paper

Ask AI

Bookmark

Model accuracy and convergence behavior depend significantly on input complexity when training convolutional neural networks across distinct parameter settings.
Experiments with the ResNet-18 residual network evaluate varied combinations of learning rates and optimizers to isolate determinants of classification performance.
Highlights how systematic hyperparameter search and careful optimizer selection remain essential to optimize convergence and mitigate overfitting during training.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Experimental analysis reveals how input complexity and optimizer selection alter classification accuracy in ResNet-18, highlighting the importance of hyperparameter search.

Cite This Study

Jianwen Lyu (2024) studied this question.

synapsesocial.com/papers/68e73fdcb6db6435876b93d2 https://doi.org/10.54254/2755-2721/46/20241177

View Full Paper

Ask AI

Bookmark

Implication

Experimental analysis reveals how input complexity and optimizer selection alter classification accuracy in ResNet-18, highlighting the importance of hyperparameter search.

Hyperparameter search and performance assessment of convolutional neural networks for image classification

Key Points

Discussion

Implication

Cite This Study

Hyperparameter search and performance assessment of convolutional neural networks for image classification

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider