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

请问一下论文中计算量与内存消耗的对比实验条件是怎样的? #33

Open
tetao00 opened this issue Dec 12, 2024 · 2 comments

Comments

@tetao00
Copy link

tetao00 commented Dec 12, 2024

您好,作者!

我对计算量与内存消耗的对比实验很感兴趣,希望能够对其进行复现。

我在编译完tvm后,在RTX 3080 10GB、Ubuntu 20.04、Python 3.7、pytorch 1.8.0、CUDA 11.1、TVM 0.8.0条件下运行
python single_step_main.py -data_path data/flow/ -dataset flow -use_tvm后,训练过程中占用显存为7.77GB,速度为4.27it/s,感觉与论文中图4展示的结果差距较大,因此对这部分实验产生了兴趣。

请问论文中图4的横坐标sequence length对应的是程序中的input_size吗?能够提供更加详细的实验条件吗?

@Zhazhan
Copy link

Zhazhan commented Dec 12, 2024

你好,感谢你关注我们的工作.
论文中图4展示的是不同Attention机制单层的显存占用和时间消耗随输入序列长度的变化,并非训练脚本的数据,旨在直观展示不同Attention机制的时空复杂度。数据可以通过下面的脚本来复现:
https://github.com/ant-research/Pyraformer/blob/master/pyraformer/graph_attention.py

@tetao00
Copy link
Author

tetao00 commented Dec 12, 2024

好的,十分感谢您的解答!

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