代码可用性

[TimGor1997]

尊敬的作者，您好！
请问现在开源的代码是否已经可以用于训练呢？

[TimGor1997]

Dear author, hello!
Can the open source code be used for training and testing？

[jiangxb98]

Dear author, hello! Can the open source code be used for training and testing？

The current code supports training and testing, please feel free to contact me if you have any questions

[TimGor1997]

Dear author,   Thank you for your reply!   However, in the README, 'replace the CCL' and 'Reorganize the current code(pts)' have not been completed yet. Will it have any impact on training or testing? TimGor ***@***.***  

…

------------------ 原始邮件 ------------------ 发件人: "jiangxb98/mwsis-plugin" ***@***.***>; 发送时间: 2024年2月12日(星期一) 晚上8:17 ***@***.***>; ***@***.******@***.***>; 主题: Re: [jiangxb98/mwsis-plugin] 代码可用性 (Issue #1) The current code supports training and testing, please feel free to contact me if you have any questions — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[jiangxb98]

1. In order to accelerate the clustering speed, the paper uses the voxelized CCL (not open source), the normal use of scipy library CCL results in little difference, but the speed is slow. 2. The code is consistent with the code used in the paper, and there may be some differences in loading data from the data processing pipeline (in order to speed up the training, the SPG module is used to generate pseudo labels in advance and save them to the local disk).