VisionMX：解锁视觉模型的微缩放训练后量化
VisionMX: Unlocking Microscaling Post-Training Quantization for Vision Models
浏览论文内容
中文总结 AI 辅助
针对视觉模型微缩放训练后量化误差，提出VisionMX方法，通过优化有界权重舍入和可折叠仿射激活校正，在多项视觉任务上优于直接转换和现有基线。
中文摘要 AI 辅助
微缩放（MX）格式正成为一种支持硬件的训练和推理高效方法。它们将低精度元素与共享块缩放相结合，但其对视觉模型的影响仍未得到充分探索。我们系统地研究了跨视觉模型和任务的训练后MX量化。对直接转换的分析确定了三个误差来源：块缩放表示、某些小型卷积权重张量与非均匀元素网格的对齐不佳，以及非负激活对符号编码的利用不足。这些发现推动了VisionMX的提出，这是一种训练后MX量化方法，可优化有界权重舍入并对激活应用可折叠仿射校正。我们使用多种MX风格格式在图像分类、目标检测、语义分割和低光图像增强中评估了VisionMX。它优于直接转换和所评估的训练后量化基线，在对MX转换最敏感的架构中获得了最大的性能恢复。
英文摘要
Microscaling (MX) formats are emerging as a hardware-supported approach to efficient training and inference. They combine low-precision elements with shared block scales, but their impact on vision models remains underexplored. We systematically investigate post-training MX quantization across vision models and tasks. An analysis of direct conversion identifies three sources of error: block-scale representation, the poor alignment of some small convolutional weight tensors with nonuniform element grids, and the underuse of signed codes by nonnegative activations. These findings motivate VisionMX, a post-training MX quantization method that optimizes bounded weight rounding and applies a foldable affine correction to activations. We evaluate VisionMX across image classification, object detection, semantic segmentation, and low-light image enhancement using several MX-style formats. It improves on direct conversion and the evaluated post-training quantization baselines, with the largest performance recoveries in architectures most sensitive to MX conversion
发表机构
- Arm AI Research (AAIR), Arm Ltd（Arm AI Research (AAIR)，Arm 公司）
机构由 AI 辅助整理，请以论文原文为准。
