piskvorky/gensim

Improving computational time in LDASeqModel

Open

#1,545 建立於 2017年8月22日

在 GitHub 查看
 (9 留言) (1 反應) (0 負責人)Python (15,144 star) (4,349 fork)batch import
Hacktoberfestdifficulty mediumfeaturegood first issueperformance

描述

Hi,

I was training a LDASeqModel on ~40k documents through 20 years (20 time slices in my case) and after 2 days of computation, still didn't finish.

I realized that only LDAModel is used in this class. Shouldn't be better to use LDAMulticore since it can also handle the case of single core ?

Thanks

貢獻者指南

技術棧
python
領域
datamachine learningperformance
議題類型
performance
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
Pythongensim
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
50
研究方向
Investigate the LDASeqModel class in gensim's source code to identify where LDAModel is used and assess the feasibility of replacing it with LDAMulticore. Check the LDAMulticore implementation to ensure it supports the same interface. Consider any potential side effects, such as changes in behavior or dependencies. Reference the issue comments for any maintainer feedback.