Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Shile, Karmann, Markus, Urfalioglu, Onay
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2602.18861
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866910029082263552
author Li, Shile
Karmann, Markus
Urfalioglu, Onay
author_facet Li, Shile
Karmann, Markus
Urfalioglu, Onay
contents We present a framework for end-to-end joint quantization of Vision Transformers trained on ImageNet for the purpose of image classification. Unlike prior post-training or block-wise reconstruction methods, we jointly optimize over the entire set of all layers and inter-block dependencies without any labeled data, scaling effectively with the number of samples and completing in just one hour on a single GPU for ViT-small. We achieve state-of-the-art W4A4 and W3A3 accuracies on ImageNet and, to the best of our knowledge, the first PTQ results that maintain strong accuracy on ViT, DeiT, and Swin-T models under extremely low-bit settings (W1.58A8), demonstrating the potential for efficient edge deployment. Furthermore, we introduce a data-free calibration strategy that synthesizes diverse, label-free samples using Stable Diffusion Turbo guided by learned multi-mode prompts. By encouraging diversity in both the learned prompt embeddings and the generated image features, our data-free approach achieves performance on par with real-data ImageNet calibration and surpasses simple text-prompt baselines such as "a <adjective> photo of <adjective> <cls>".
format Preprint
id arxiv_https___arxiv_org_abs_2602_18861
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Joint Post-Training Quantization of Vision Transformers with Learned Prompt-Guided Data Generation
Li, Shile
Karmann, Markus
Urfalioglu, Onay
Computer Vision and Pattern Recognition
We present a framework for end-to-end joint quantization of Vision Transformers trained on ImageNet for the purpose of image classification. Unlike prior post-training or block-wise reconstruction methods, we jointly optimize over the entire set of all layers and inter-block dependencies without any labeled data, scaling effectively with the number of samples and completing in just one hour on a single GPU for ViT-small. We achieve state-of-the-art W4A4 and W3A3 accuracies on ImageNet and, to the best of our knowledge, the first PTQ results that maintain strong accuracy on ViT, DeiT, and Swin-T models under extremely low-bit settings (W1.58A8), demonstrating the potential for efficient edge deployment. Furthermore, we introduce a data-free calibration strategy that synthesizes diverse, label-free samples using Stable Diffusion Turbo guided by learned multi-mode prompts. By encouraging diversity in both the learned prompt embeddings and the generated image features, our data-free approach achieves performance on par with real-data ImageNet calibration and surpasses simple text-prompt baselines such as "a <adjective> photo of <adjective> <cls>".
title Joint Post-Training Quantization of Vision Transformers with Learned Prompt-Guided Data Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2602.18861