A universal vision transformer for fast calorimeter simulations -- Supplementary material

This repository contains supplementary material to "A universal vision transformer for fast calorimeter simulations". We include the full set of high-level features for the CaloChallenge and LEMURS datasets in zip files. For the LEMURS dataset, we also include the binning XML files used in the evaluation script.We also include the generated samples used for the evaluation of all the datasets studied in the paper: CaloChallenge: ds1-$\gamma$, ds1-$\pi^{+}$, ds2, ds3; LEMURS: full evaluation for the five detectors; CaloHadronic: sample of 20k showers for the down-sampled geometry. Feel free to contact us for questions and requests.Changelog:v1.1.0: Include generated samples and high-level features for the CaloChallenge ds2 from a network pretrained on the LEMURS dataset. This samples correspond to a network trained until convergence on the full ds2 training dataset.本仓库为《面向快速量热计（calorimeter）模拟的通用视觉Transformer（universal vision transformer）》的补充材料。 我们以压缩包形式提供了CaloChallenge与LEMURS数据集的全部高阶特征。针对LEMURS数据集，我们还提供了评估脚本中使用的分箱XML文件。此外，我们还提供了用于评估本文研究的所有数据集的生成样本： CaloChallenge：ds1-γ、ds1-π⁺、ds2、ds3； LEMURS：五种探测器的完整评估集； CaloHadronic：用于下采样几何结构的20k次簇射（shower）样本。 如有疑问与需求，欢迎联系我们。 更新日志：v1.1.0：新增基于LEMURS数据集预训练的网络生成的CaloChallenge ds2数据集的生成样本与高阶特征，该样本对应于在完整ds2训练数据集上训练至收敛的网络。