{"ID":2839526,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2511.15369","arxiv_id":"2511.15369","title":"IPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformers","abstract":"Previous Quantization-Aware Training (QAT) methods for vision transformers rely on expensive retraining to recover accuracy loss in non-linear layer quantization, limiting their use in resource-constrained environments. In contrast, existing Post-Training Quantization (PTQ) methods either partially quantize non-linear functions or adjust activation distributions to maintain accuracy but fail to achieve fully integer-only inference. In this paper, we introduce IPTQ-ViT, a novel PTQ framework for fully integer-only vision transformers without retraining. We present approximation functions: a polynomial-based GELU optimized for vision data and a bit-shifting-based Softmax designed to improve approximation accuracy in PTQ. In addition, we propose a unified metric integrating quantization sensitivity, perturbation, and computational cost to select the optimal approximation function per activation layer. IPTQ-ViT outperforms previous PTQ methods, achieving up to 6.44\\%p (avg. 1.78\\%p) top-1 accuracy improvement for image classification, 1.0 mAP for object detection. IPTQ-ViT outperforms partial floating-point PTQ methods under W8A8 and W4A8, and achieves accuracy and latency comparable to integer-only QAT methods. We plan to release our code https://github.com/gihwan-kim/IPTQ-ViT.git.","short_abstract":"Previous Quantization-Aware Training (QAT) methods for vision transformers rely on expensive retraining to recover accuracy loss in non-linear layer quantization, limiting their use in resource-constrained environments. In contrast, existing Post-Training Quantization (PTQ) methods either partially quantize non-linear...","url_abs":"https://arxiv.org/abs/2511.15369","url_pdf":"https://arxiv.org/pdf/2511.15369v1","authors":"[\"Gihwan Kim\",\"Jemin Lee\",\"Hyungshin Kim\"]","published":"2025-11-19T11:56:16Z","proceeding":"cs.CV","tasks":"[\"cs.CV\",\"cs.AI\"]","methods":"[\"Vision Transformer\",\"Transformer\"]","has_code":false,"code_links":[{"ID":606880,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2839526,"paper_url":"https://arxiv.org/abs/2511.15369","paper_title":"IPTQ-ViT: Post-Training Quantization of Non-linear Functions for Integer-only Vision Transformers","repo_url":"https://github.com/gihwan-kim/IPTQ-ViT.git","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}