# Joint Architecture-Token-Bitwidth Multi-Axis Optimization of Vision Transformers for Semiconductor IC Packaging

arXiv: https://arxiv.org/abs/2605.01742 · HTML: https://mlsanity.com/paper/2605.01742/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 32 references that are in the map.
- First posted on arXiv 3 May 2026.

## Builds on (most cited first, 24 of 24)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Token Merging: Your ViT But Faster](https://mlsanity.com/paper/2210.09461/) (2022)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [Post-Training Quantization for Vision Transformer](https://mlsanity.com/paper/2106.14156/) (2021)
- [A Survey on Efficient Inference for Large Language Models](https://mlsanity.com/paper/2404.14294/) (2024)
- [AutoFormer: Searching Transformers for Visual Recognition](https://mlsanity.com/paper/2107.00651/) (2021)
- [PTQ4ViT: Post-training quantization for vision transformers with twin uniform quantization](https://mlsanity.com/paper/2111.12293/) (2021)
- [SPViT: Enabling Faster Vision Transformers via Soft Token Pruning](https://mlsanity.com/paper/2112.13890/) (2021)
- [FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer](https://mlsanity.com/paper/2111.13824/) (2021)
- [Not All Tokens Are Equal: Human-centric Visual Analysis via Token Clustering Transformer](https://mlsanity.com/paper/2204.08680/) (2022)
- [I-ViT: Integer-only Quantization for Efficient Vision Transformer Inference](https://mlsanity.com/paper/2207.01405/) (2022)
- [VCT: A Video Compression Transformer](https://mlsanity.com/paper/2206.07307/) (2022)
- [Towards Accurate Post-Training Quantization for Vision Transformer](https://mlsanity.com/paper/2303.14341/) (2023)
- [NoisyQuant: Noisy Bias-Enhanced Post-Training Activation Quantization for Vision Transformers](https://mlsanity.com/paper/2211.16056/) (2022)
- [Integrating Human Gaze into Attention for Egocentric Activity Recognition](https://mlsanity.com/paper/2011.03920/) (2020)
- [Searching the Search Space of Vision Transformer](https://mlsanity.com/paper/2111.14725/) (2021)
- [Discovering Transferable Forensic Features for CNN-generated Images Detection](https://mlsanity.com/paper/2208.11342/) (2022)
- [Vision Transformers on the Edge: A Comprehensive Survey of Model Compression and Acceleration Strategies](https://mlsanity.com/paper/2503.02891/) (2025)
- [RLRC: Reinforcement Learning-based Recovery for Compressed Vision-Language-Action Models](https://mlsanity.com/paper/2506.17639/) (2025)
- [Revisit Multimodal Meta-Learning through the Lens of Multi-Task Learning](https://mlsanity.com/paper/2110.14202/) (2021)
- [Egocentric Activity Recognition with Multimodal Fisher Vector](https://mlsanity.com/paper/1601.06603/) (2016)
- [On-device Scalable Image-based Localization via Prioritized Cascade Search and Fast One-Many RANSAC](https://mlsanity.com/paper/1802.03510/) (2018)
- [On Accelerating Edge AI: Optimizing Resource-Constrained Environments](https://mlsanity.com/paper/2501.15014/) (2025)

## Abstract (arXiv)

Vision Transformers (ViTs) have achieved strong performance in visual recognition, yet their deployment in resource-constrained industrial environments remains limited. The main challenges are their high computational cost, memory requirements, and energy consumption. While individual efficiency techniques such as neural architecture search (NAS), token compression, and low-precision inference have been extensively studied, most prior work targets only a single optimization axis, limiting overall deployment gains while preserving accuracy. In this paper, we present one of the first holistic frameworks that jointly optimizes three complementary axes: architecture, token, and bit-width. Specifically, the framework identifies compact backbones via Neural Architecture Search (AutoFormer), reduces information processing via token merging (ToMe), and accelerates per-operation execution via fp16 mixed-precision inference. In our study, we analyze accuracy-efficiency trade-offs on ImageNet-1K under aggressive compression. We then apply the selected configaturions to a real-world in-house 3D X-ray semiconductor defect classification dataset for IC chip packaging inspection. Results show that the proposed multi-axis framework achieves more than 10× improvement in throughput along with over 10× reductions in parameter count, FLOPs, and energy consumption, while maintaining the required accuracy on the downstream industrial task. To the best of our knowledge, this is among the earliest works to jointly optimize architecture, token, and bit-width dimensions in ViTs and the first such resource-efficient, deployment-focused study tailored to semiconductor manufacturing.