Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

不做PSI,纵向直接从对齐好的数据训练可以吗? #5698

Open
FancyXun opened this issue Aug 12, 2024 · 1 comment
Open

不做PSI,纵向直接从对齐好的数据训练可以吗? #5698

FancyXun opened this issue Aug 12, 2024 · 1 comment

Comments

@FancyXun
Copy link

FancyXun commented Aug 12, 2024

请教个问题,如果我训练每次都做psi感觉没太大必要,有时候我希望psi做完一次后,就专注做模型的训练。但是我看所有的例子好像都需要做一次psi,然后进行纵向训练,

pipeline.add_tasks([reader_0, psi_0, lr_0, evaluation_0])
有没有可能,我实现(psi)对齐后数据,以纵向Lr为例,在dag图里面直接从 reader - > lr. 我尝试修改了下,如下

lr_0 = CoordinatedLR("lr_0",
                         epochs=10,
                         batch_size=300,
                         optimizer={"method": "SGD", "optimizer_params": {"lr": 0.1}, "penalty": "l2", "alpha": 0.001},
                         init_param={"fit_intercept": True, "method": "zeros"},
                         train_data=reader_0.outputs["output_data"],
                         learning_rate_scheduler={"method": "linear", "scheduler_params": {"start_factor": 0.7,
                                                                                           "total_iters": 100}})
pipeline.add_tasks([reader_0,  lr_0, evaluation_0])

但是host端报错: NoneType
Screen Shot 2024-08-12 at 18 45 07

@FancyXun FancyXun changed the title 纵向支持不做PSI,直接训练吗? 不做PSI,纵向直接从对齐好的数据训练可以吗? Aug 12, 2024
@mgqa34
Copy link
Contributor

mgqa34 commented Aug 15, 2024

可以单独跑一个reader->psi的任务,然后这一步的输出会有对齐后的数据表,后续可以直接作为reader的输入。

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants