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.

Sign up for GitHub

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

Jump to bottom

[QA] loong train 支持packed_sample_into_one=false吗 #346

Open
Lzhang-hub opened this issue Sep 27, 2024 · 1 comment
Open

[QA] loong train 支持packed_sample_into_one=false吗 #346

Lzhang-hub opened this issue Sep 27, 2024 · 1 comment
Assignees
@ZwwWayne
Labels
question Further information is requested

Comments

@Lzhang-hub
Copy link

描述问题

咨询一下，长文本训练支持样本间的相互隔离吗？

image
@Lzhang-hub Lzhang-hub added the question Further information is requested label Sep 27, 2024
@mwiacx
Copy link
Contributor

mwiacx commented Oct 29, 2024 •
edited
Loading

支持，internevo默认配置基本上都是 use_packed_data = True， pack_sample_into_one = False；不过loongtrain我们目前只支持unpack data，由于2d attn 依赖的zigzag attn那边暂时只适配了unpack的版本

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@ZwwWayne ZwwWayne

Labels
question Further information is requested
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@mwiacx @ZwwWayne @Lzhang-hub