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

分享训练时候的 memory, cpu, gpu 信息 #4

Open
Dreamgoing opened this issue May 20, 2023 · 9 comments
Open

分享训练时候的 memory, cpu, gpu 信息 #4

Dreamgoing opened this issue May 20, 2023 · 9 comments

Comments

@Dreamgoing
Copy link

Dreamgoing commented May 20, 2023

非常赞的project!我可以在本地test,但是train的时候,memory和cpu oom了。作者可否分享一下cpu和memory的信息

@lang07123
Copy link

图片

瞎点..不知道对不对..你可以参考一下...

@Dreamgoing
Copy link
Author

谢谢

@lang07123
Copy link

CPU: I5-10400
内存: 32G内存
显卡: RTX2060 6G
sata SSD

图片
跑了15个小时跟作者5个小时差不多......

@pangyouzhen
Copy link

pangyouzhen commented May 23, 2023

@lang07123 我用linux 服务器训练的,设置SDL_VIDEODRIVER='dummy',16C128G,GPU利用率周期性0-28%,env=64,batch_size=1024, 到50M时候也15个小时。。。

@pangyouzhen
Copy link

@lang07123
Copy link

@lang07123 我用linux 服务器训练的,设置SDL_VIDEODRIVER='dummy',16C128G,GPU利用率周期性0-28%,env=64,batch_size=1024, 到50M时候也15个小时。。。

我刚才用M2 的 macbook pro 测试了一下.... 大概是RTX 2060 的两倍...但是热的烫手..舍不得了....

@lang07123
Copy link

@lang07123 我用linux 服务器训练的,设置SDL_VIDEODRIVER='dummy',16C128G,GPU利用率周期性0-28%,env=64,batch_size=1024, 到50M时候也15个小时。。。

那你的估计还可以加操作... 我的2060 基本上100%了..也就是这个速度了...

@pangyouzhen
Copy link

@lang07123 发现linux服务器不行,提高env和batch_size效果也一般,如果看进程的话,CPU应该只占用了1核,GPU周期性波动,linux服务器根本发挥不出优势,也不知道咋改了

@sjgfa
Copy link

sjgfa commented May 28, 2023

CPU: I5-10400 内存: 32G内存 显卡: RTX2060 6G sata SSD

图片 跑了15个小时跟作者5个小时差不多......

作者用的是4090,我用4090和作者的速率差不多

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

4 participants