# Cyclical Log Annealing as a Learning Rate Scheduler

arXiv: https://arxiv.org/abs/2403.14685 · HTML: https://mlsanity.com/paper/2403.14685/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 9 of its 16 references that are in the map.
- First posted on arXiv 13 Mar 2024.

## Builds on (most cited first, 9 of 9)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [Going Deeper with Convolutions](https://mlsanity.com/paper/1409.4842/) (2014)
- [Xception: Deep Learning with Depthwise Separable Convolutions](https://mlsanity.com/paper/1610.02357/) (2016)
- [Introduction to Online Convex Optimization](https://mlsanity.com/paper/1909.05207/) (2019)
- [Empirical Evaluation of Rectified Activations in Convolutional Network](https://mlsanity.com/paper/1505.00853/) (2015)
- [Automated Learning Rate Scheduler for Large-batch Training](https://mlsanity.com/paper/2107.05855/) (2021)

## Built on by (most cited first, 1 of 1)

- [Efficient Deep Learning Approaches for Processing Ultra-Widefield Retinal Imaging](https://mlsanity.com/paper/2503.18151/) (2025)

## Abstract (arXiv)

A learning rate scheduler is a predefined set of instructions for varying search stepsizes during model training processes. This paper introduces a new logarithmic method using harsh restarting of step sizes through stochastic gradient descent. Cyclical log annealing implements the restart pattern more aggressively to maybe allow the usage of more greedy algorithms on the online convex optimization framework. The algorithm was tested on the CIFAR-10 image datasets, and seemed to perform analogously with cosine annealing on large transformer-enhanced residual neural networks. Future experiments would involve testing the scheduler in generative adversarial networks and finding the best parameters for the scheduler with more experiments.