# DyCE: Dynamically Configurable Exiting for Deep Learning Compression and Real-time Scaling

arXiv: https://arxiv.org/abs/2403.01695 · HTML: https://mlsanity.com/paper/2403.01695/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 35 references that are in the map.
- First posted on arXiv 4 Mar 2024.

## Builds on (most cited first, 21 of 21)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://mlsanity.com/paper/1905.11946/) (2019)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [Scaling Vision with Sparse Mixture of Experts](https://mlsanity.com/paper/2106.05974/) (2021)
- [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://mlsanity.com/paper/2301.00808/) (2023)
- [BranchyNet: Fast Inference via Early Exiting from Deep Neural Networks](https://mlsanity.com/paper/1709.01686/) (2017)
- [What is the State of Neural Network Pruning?](https://mlsanity.com/paper/2003.03033/) (2020)
- [Dynamic Neural Networks: A Survey](https://mlsanity.com/paper/2102.04906/) (2021)
- [FastBERT: a Self-distilling BERT with Adaptive Inference Time](https://mlsanity.com/paper/2004.02178/) (2020)
- [DaViT: Dual Attention Vision Transformers](https://mlsanity.com/paper/2204.03645/) (2022)
- [Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognition](https://mlsanity.com/paper/2105.15075/) (2021)
- [Fixed Point Quantization of Deep Convolutional Networks](https://mlsanity.com/paper/1511.06393/) (2015)
- [CF-ViT: A General Coarse-to-Fine Method for Vision Transformer](https://mlsanity.com/paper/2203.03821/) (2022)
- [A Panda? No, It's a Sloth: Slowdown Attacks on Adaptive Multi-Exit Neural Network Inference](https://mlsanity.com/paper/2010.02432/) (2020)
- [Learning to Stop While Learning to Predict](https://mlsanity.com/paper/2006.05082/) (2020)
- [Filter Sketch for Network Pruning](https://mlsanity.com/paper/2001.08514/) (2020)
- [It's always personal: Using Early Exits for Efficient On-Device CNN Personalisation](https://mlsanity.com/paper/2102.01393/) (2021)
- [Mobile V-MoEs: Scaling Down Vision Transformers via Sparse Mixture-of-Experts](https://mlsanity.com/paper/2309.04354/) (2023)

## Built on by (most cited first, 3 of 3)

- [Tiny Models are the Computational Saver for Large Models](https://mlsanity.com/paper/2403.17726/) (2024)
- [ORXE: Orchestrating Experts for Dynamically Configurable Efficiency](https://mlsanity.com/paper/2505.04850/) (2025)
- [Adaptive AI: Energy Efficient Multi-exit TinyML on Intelligent Vision Systems at the Edge](https://mlsanity.com/paper/2609.11939/) (2026)

## Abstract (arXiv)

Conventional deep learning (DL) model compression and scaling methods focus on altering the model's components, impacting the results across all samples uniformly. However, since samples vary in difficulty, a dynamic model that adapts computation based on sample complexity offers a novel perspective for compression and scaling. Despite this potential, existing dynamic models are typically monolithic and model-specific, limiting their generalizability as broad compression and scaling methods. Additionally, most deployed DL systems are fixed, unable to adjust their scale once deployed and, therefore, cannot adapt to the varying real-time demands. This paper introduces DyCE, a dynamically configurable system that can adjust the performance-complexity trade-off of a DL model at runtime without requiring re-initialization or redeployment on inference hardware. DyCE achieves this by adding small exit networks to intermediate layers of the original model, allowing computation to terminate early if acceptable results are obtained. DyCE also decouples the design of an efficient dynamic model, facilitating easy adaptation to new base models and potential general use in compression and scaling. We also propose methods for generating optimized configurations and determining the types and positions of exit networks to achieve desired performance and complexity trade-offs. By enabling simple configuration switching, DyCE provides fine-grained performance tuning in real-time. We demonstrate the effectiveness of DyCE through image classification tasks using deep convolutional neural networks (CNNs). DyCE significantly reduces computational complexity by 23.5% for ResNet152 and 25.9% for ConvNextv2-tiny on ImageNet, with accuracy reductions of less than 0.5%.