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

【Paddle Toolkit Development Competition No.3】 Paddle 适配 neuralopfiles via upload #985

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

Kai-Qi
Copy link

@Kai-Qi Kai-Qi commented Oct 25, 2024

Copy link

paddle-bot bot commented Oct 25, 2024

你的PR提交成功,感谢你对开源项目的贡献!
请检查PR提交格式和内容是否完备,具体请参考示例模版
Your PR has been submitted. Thanks for your contribution!
Please check its format and content. For this, you can refer to Template and Demo.

@CLAassistant
Copy link

CLAassistant commented Oct 25, 2024

CLA assistant check
All committers have signed the CLA.

Comment on lines +612 to +619
tensorly支持Paddle后端,但是在编写代码时发生了如下错误:

```python
import tensorly as tl
tl.set_backend('paddle')

ValueError:Unknown backend name 'paddle', known backends are ['numpy', 'mxnet', 'pytorch', 'tensorflow', 'cupy', 'jax']
```
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tensorly可能需要使用源码安装的方式,因为paddle支持的代码tensoly官方尚未发布到pypi上

image

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

好的 我会尝试 谢谢

Comment on lines +51 to +52
`neuraloperator`源代码实现的8个模型分别为, `FNO1d`, `FNO2d`, `FNO3d`, `SFNO`, `UNO`, `TFNO1d`, `TFNO2d`, `TFNO3d`。

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

8个模型只是一部分吧,neuralop需要适配的内容应该是除了training外的所有代码
image

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

我们针对模型的数量是8个 但其他所有文件也都适配了 而且我们也适配了training过程

Copy link
Contributor

@HydrogenSulfate HydrogenSulfate left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

辛苦提交RFC,有几点建议可以看一下




由于模型`UFNO(layers=1)`时的训练结果是对齐的,因此初步判断模型UFNO(layers=5)时出现的问题是由误差累积引起的。目前该问题尚未解决。
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

误差累积不太可能会这么大,可以尝试使用https://github.com/PaddlePaddle/PaDiff工具,定位一下具体的误差累积在哪一层,然后看下是什么API的计算导致的误差变化

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

好的 后面有时间时我会尝试

@HydrogenSulfate
Copy link
Contributor

hello @Kai-Qi 可以修改一RFC的内容,在文档中补充需要适配的API列表和单测列表

@Kai-Qi
Copy link
Author

Kai-Qi commented Nov 14, 2024

hello @Kai-Qi 可以修改一RFC的内容,在文档中补充需要适配的API列表和单测列表
请问单测列表 是指我们paddle模型的测试文件吗?比如文件命名为test_FNO1d, test_TFNO_1d ....... 等,谢谢 @HydrogenSulfate

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

Successfully merging this pull request may close these issues.

4 participants