# Context-Aware Token Selection and Packing for Enhanced Vision Transformer

arXiv: https://arxiv.org/abs/2410.23608 · HTML: https://mlsanity.com/paper/2410.23608/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 32 of its 40 references that are in the map.
- First posted on arXiv 31 Oct 2024.

## Builds on (most cited first, 30 of 32)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer](https://mlsanity.com/paper/1701.06538/) (2017)
- [Categorical Reparameterization with Gumbel-Softmax](https://mlsanity.com/paper/1611.01144/) (2016)
- [BEiT: BERT Pre-Training of Image Transformers](https://mlsanity.com/paper/2106.08254/) (2021)
- [SwinIR: Image Restoration Using Swin Transformer](https://mlsanity.com/paper/2108.10257/) (2021)
- [GLU Variants Improve Transformer](https://mlsanity.com/paper/2002.05202/) (2020)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [Deformable Convolutional Networks](https://mlsanity.com/paper/1703.06211/) (2017)
- [Swin Transformer V2: Scaling Up Capacity and Resolution](https://mlsanity.com/paper/2111.09883/) (2021)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [Not All Patches are What You Need: Expediting Vision Transformers via Token Reorganizations](https://mlsanity.com/paper/2202.07800/) (2022)
- [Patch n' Pack: NaViT, a Vision Transformer for any Aspect Ratio and Resolution](https://mlsanity.com/paper/2307.06304/) (2023)
- [SparseViT: Revisiting Activation Sparsity for Efficient High-Resolution Vision Transformer](https://mlsanity.com/paper/2303.17605/) (2023)
- [Towards Understanding Mixture of Experts in Deep Learning](https://mlsanity.com/paper/2208.02813/) (2022)
- [Time Series Modeling for Heart Rate Prediction: From ARIMA to Transformers](https://mlsanity.com/paper/2406.12199/) (2024)
- [Transformer-VQ: Linear-Time Transformers via Vector Quantization](https://mlsanity.com/paper/2309.16354/) (2023)
- [Harnessing Earnings Reports for Stock Predictions: A QLoRA-Enhanced LLM Approach](https://mlsanity.com/paper/2408.06634/) (2024)
- [Towards Resilient and Efficient LLMs: A Comparative Study of Efficiency, Performance, and Adversarial Robustness](https://mlsanity.com/paper/2408.04585/) (2024)
- [Heterogeneous Multi-task Learning with Expert Diversity](https://mlsanity.com/paper/2106.10595/) (2021)
- [Evaluating Modern Approaches in 3D Scene Reconstruction: NeRF vs Gaussian-Based Methods](https://mlsanity.com/paper/2408.04268/) (2024)
- [Multi-manifold Attention for Vision Transformers](https://mlsanity.com/paper/2207.08569/) (2022)
- [GateNet: Gating-Enhanced Deep Network for Click-Through Rate Prediction](https://mlsanity.com/paper/2007.03519/) (2020)
- [Transformer-based Selective Super-Resolution for Efficient Image Refinement](https://mlsanity.com/paper/2312.05803/) (2023)

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

- [Enhancing Open-Vocabulary Object Detection through Multi-Level Fine-Grained Visual-Language Alignment](https://mlsanity.com/paper/2602.00531/) (2026)

## Abstract (arXiv)

In recent years, the long-range attention mechanism of vision transformers has driven significant performance breakthroughs across various computer vision tasks. However, the traditional self-attention mechanism, which processes both informative and non-informative tokens, suffers from inefficiency and inaccuracies. While sparse attention mechanisms have been introduced to mitigate these issues by pruning tokens involved in attention, they often lack context-awareness and intelligence. These mechanisms frequently apply a uniform token selection strategy across different inputs for batch training or optimize efficiency only for the inference stage. To overcome these challenges, we propose a novel algorithm: Select and Pack Attention (SPA). SPA dynamically selects informative tokens using a low-cost gating layer supervised by selection labels and packs these tokens into new batches, enabling a variable number of tokens to be used in parallelized GPU batch training and inference. Extensive experiments across diverse datasets and computer vision tasks demonstrate that SPA delivers superior performance and efficiency, including a 0.6 mAP improvement in object detection and a 16.4% reduction in computational costs.