跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2606.16119 · v2

EdgeZSAD: Value-Range Quantization for Deploying Zero-Shot Anomaly Detection on Edge Neural Processing Units

首次提交 2026-06-15 10:17
版本更新 2026-11-13 18:43
雷达观测 2026-11-16 11:42
arXiv ↗PDF ↗

摘要

Quantizing zero-shot industrial anomaly detectors for edge neural processing units can degrade defect detection and localization. We present EdgeZSAD (Edge Zero-Shot Anomaly Detection), a deployment framework whose methodological contribution is Value-Range Quantization (VRQ). We diagnose how attention Value tensors inherit overly wide ranges from fused query-key-value projections and assign independent ranges calibrated on source data, without retraining or changing quantized weights. The engineering application is native deployment of a foundation-model detector for industrial inspection, evaluated on three public benchmarks and Qualcomm edge devices. Under minimum-maximum and mean-squared-error calibration, VRQ improves image-level discrimination and per-region overlap. In the primary calibration set, it recovers 90-98% of the loss in combined image and pixel discrimination relative to half precision, with a small increase in latency over baseline quantization. Matched tensor overrides reproduce the recovery without adding graph nodes, identifying independent Value-range assignment as the effective intervention in the evaluated runtime family.

作者备注:22 pages. Substantially revised version with a focus on Value-Range Quantization and native edge deployment evaluation

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-13 18:43 · 最新观测版本

EdgeZSAD: Value-Range Quantization for Deploying Zero-Shot Anomaly Detection on Edge Neural Processing Units

Quantizing zero-shot industrial anomaly detectors for edge neural processing units can degrade defect detection and localization. We present EdgeZSAD (Edge Zero-Shot Anomaly Detection), a deployment framework whose methodological contribution is Value-Range Quantization (VRQ). We diagnose how attention Value tensors inherit overly wide ranges from fused query-key-value projections and assign independent ranges calibrated on source data, without retraining or changing quantized weights. The engineering application is native deployment of a foundation-model detector for industrial inspection, evaluated on three public benchmarks and Qualcomm edge devices. Under minimum-maximum and mean-squared-error calibration, VRQ improves image-level discrimination and per-region overlap. In the primary calibration set, it recovers 90-98% of the loss in combined image and pixel discrimination relative to half precision, with a small increase in latency over baseline quantization. Matched tensor overrides reproduce the recovery without adding graph nodes, identifying independent Value-range assignment as the effective intervention in the evaluated runtime family.

查看此版本解读 →查看此版本原文 ↗