lmcinnes/umap

Document reproducibility guarantees

Open

#298 创建于 2019年9月25日

在 GitHub 查看
 (16 评论) (0 反应) (0 负责人)Python (6,478 star) (751 fork)batch import
0.4documentationgood first issue

描述

Following up on the discussion here, it would be good to document how to get reproducible results with UMAP.

I think we should consider changing random_state in the UMAP constructor to a seed (e.g. 42, like the new transform_seed default) so that UMAP is reproducible by default.

We should document that users can set random_state to None to get faster results at the expense of reproducibility. In this mode there is no seed that would produce the same output due to the multithreading. (This was introduced in #294.)

贡献者指南

技术栈
python
领域
documentation
议题类型
documentation
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
3
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
half day
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
stale
清晰度议题是否清楚说明期望改动、验收标准和下一步。
clear
前置要求
PythonUMAP libraryrandom state conceptdocumentation writing
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
40
研究方向
Review the discussion in PR #231 and issue #294 to understand the current behavior of random state and transform seed. The main task is to update the UMAP documentation to explain how to achieve reproducible results, including possibly changing the default random state to a seed. Files likely include the UMAP class docstring and the main documentation page. Also consider the trade offs with performance when setting random state to None.