🏆 实验结果表格生成（多种子 mean ± std，自动加粗最佳）Results Table Generator (mean ± std over Seeds, Best in Bold)
把多次运行（不同随机种子）的实验结果粘贴进来，自动按方法汇总成“均值 ± 标准差（或 SEM、95% CI）”，每列按越大越好或越小越好自动加粗最佳、下划线次佳，可缩放 ×100、设置小数位和 ± 样式，一键导出 LaTeX booktabs 表格、Markdown 或 CSV；也可以直接粘贴已经写好的“85.3 ± 0.4”。Paste results from several runs (random seeds) and they are aggregated per method into mean ± std (or SEM, 95% CI). Each column is marked higher- or lower-is-better, with the best in bold and second best underlined; scale ×100, set decimals and ± style, and export LaTeX booktabs, Markdown or CSV. Cells already written as “85.3 ± 0.4” work too.
同一方法的多行会自动合并：均值与样本标准差（n−1，与 numpy.std(ddof=1)、pandas .std() 一致）。“最佳”按均值比较，并列时都加粗；箭头 ↑/↓ 会写入表头。LaTeX 需要 \usepackage{booktabs}，下划线用 \underline。Rows with the same method are merged: mean and sample SD (n−1, same as numpy.std(ddof=1) / pandas .std()). Best is by mean, ties all bold; ↑/↓ arrows go into the header. The LaTeX needs \usepackage{booktabs}; second best uses \underline.
使用说明与常见问题How-to & FAQ
「实验结果表格生成（多种子 mean ± std，自动加粗最佳）」是什么？What is the Results Table Generator (mean ± std over Seeds, Best in Bold)?
把多次运行（不同随机种子）的实验结果粘贴进来，自动按方法汇总成“均值 ± 标准差（或 SEM、95% CI）”，每列按越大越好或越小越好自动加粗最佳、下划线次佳，可缩放 ×100、设置小数位和 ± 样式，一键导出 LaTeX booktabs 表格、Markdown 或 CSV；也可以直接粘贴已经写好的“85.3 ± 0.4”。Paste results from several runs (random seeds) and they are aggregated per method into mean ± std (or SEM, 95% CI). Each column is marked higher- or lower-is-better, with the best in bold and second best underlined; scale ×100, set decimals and ± style, and export LaTeX booktabs, Markdown or CSV. Cells already written as “85.3 ± 0.4” work too.
如何使用「实验结果表格生成（多种子 mean ± std，自动加粗最佳）」？How do I use the Results Table Generator (mean ± std over Seeds, Best in Bold)?
第一列是方法名，同一方法的多行会自动合并（名为 Seed、Run、Fold 的列会被忽略）。名称含 loss、error、latency、params 等的列默认“越小越好”，点击上方标签可以切换。调整选项后复制 LaTeX 代码。The first column is the method; rows with the same method are merged (columns named Seed, Run or Fold are ignored). Columns with loss, error, latency, params etc. default to lower-is-better — click a tag to toggle. Then copy the LaTeX.
标准差用 n 还是 n−1？Is the SD computed with n or n−1?
使用样本标准差（除以 n−1），与 numpy.std(ddof=1)、pandas 的 .std() 和 R 的 sd() 一致，是论文中报告多次运行结果的常用做法。The sample SD (divide by n−1), as numpy.std(ddof=1), pandas .std() and R sd() — the usual choice for reporting runs.
最佳值并列怎么处理？How are ties handled?
按显示的小数位四舍五入后比较，并列的结果都会加粗。Values are compared after rounding to the displayed decimals; tied best values are all bold.
「实验结果表格生成（多种子 mean ± std，自动加粗最佳）」免费吗？文件会上传吗？Is the Results Table Generator (mean ± std over Seeds, Best in Bold) free? Are my files uploaded?
「实验结果表格生成（多种子 mean ± std，自动加粗最佳）」完全免费，无需注册。你输入的内容只在本机浏览器里处理（并保存在本机方便下次继续），不会上传到任何服务器，适合处理未发表的论文和数据。The Results Table Generator (mean ± std over Seeds, Best in Bold) is free with no sign-up. Everything you enter is processed in this browser only (and kept locally so you can continue later); nothing is uploaded, so it is safe for unpublished papers and data.