Comments (6)
HydrogenSulfate
commented on July 28, 2024
F.interpolate存在diff好像是符合预期的，可以试下在CPU上运行上述程序，印象里CPU的结果能跟torch的GPU对齐。
from paddlescience.
Yang-Changhui
commented on July 28, 2024
好的，谢谢。但是，这段测试代码，经过判断条件后，并没有进入generalize_padding函数，没有用到F.interpolate，这个能麻烦您再看一下吗
from paddlescience.
HydrogenSulfate
commented on July 28, 2024
好的，谢谢。但是，这段测试代码，经过判断条件后，并没有进入generalize_padding函数，没有用到F.interpolate，这个能麻烦您再看一下吗
可以使用这个工具具体定位一下是具体哪一层导致的diff，https://github.com/PaddlePaddle/PaDiff
from paddlescience.
Yang-Changhui
commented on July 28, 2024
@HydrogenSulfate 你好，设置"rtol":1e-6,"atol":1e-6，显示linear层有误差，但是linear的权重是一样的；
设置"rtol":1e-4,"atol":1e-7，可以 通过；但是使用reprod_log误差在相同的机器上为0.001，之前torch使用3060，paddle使用A100，误差为0.0006
from paddlescience.
HydrogenSulfate
commented on July 28, 2024
@HydrogenSulfate 你好，设置"rtol":1e-6,"atol":1e-6，显示linear层有误差，但是linear的权重是一样的； 设置"rtol":1e-4,"atol":1e-7，可以 通过；但是使用reprod_log误差在相同的机器上为0.001，之前torch使用3060，paddle使用A100，误差为0.0006
- 确认对于同样的输入，F.pad在上述代码中是否能产生完全一致的结果
- 含有XXXNorm层时，输出的diff达到1e-4是有可能的，可以尝试把双方的norm层全部注释掉再对齐
from paddlescience.
HydrogenSulfate
commented on July 28, 2024
@HydrogenSulfate 你好，设置"rtol":1e-6,"atol":1e-6，显示linear层有误差，但是linear的权重是一样的； 设置"rtol":1e-4,"atol":1e-7，可以 通过；但是使用reprod_log误差在相同的机器上为0.001，之前torch使用3060，paddle使用A100，误差为0.0006
- 确认对于同样的输入，F.pad在上述代码中是否能产生完全一致的结果
- 含有XXXNorm层时，输出的diff达到1e-4是有可能的，可以尝试把双方的norm层全部注释掉再对齐
另外对齐的时候尽量使用同样的卡，避免GPU硬件不同带来的的影响
from paddlescience.
Related Issues (20)
- 【快乐开源】迁移 AMGNet 案例至 PaddleScience HOT 2
- 【快乐开源】复现 Phy-LSTM 论文 HOT 2
- 【快乐开源】为 PaddleScience 添加 HamiltonianMonteCarlo API HOT 3
- where is volterra HOT 1
- 优化文档结构，增强非专业领域人员可读性，增加应用场景描述，实验图表注明各变量物理意义 HOT 3
- 安装paddlescience报错 HOT 7
- 【快乐开源】基于hydra的案例改造计划 HOT 8
- bracket评估命令错误 HOT 1
- 关于定常问题初始流场设置 HOT 2
- 【快乐开源】补充汽车CFD数据集(Shape-Net Car)至 PaddleScience
- 【快乐开源】补充汽车CFD数据集(Ahmed)至 PaddleScience
- 【快乐开源】科学计算 API 文档补全 HOT 13
- 如何实现求三阶导数 HOT 3
- AI Studio快速体验ldc2d_steady_Re10完整代码出现error HOT 1
- 【快乐开源】为 PaddleScience 案例添加 export 和 inference 功能 HOT 21
- paddlescience训练模型，验证时出错 HOT 1
- 【快乐开源】为PaddleScience添加chamfer distance loss HOT 1
- 训练的时候出现反向传播错误 HOT 7
- 带有paddle.fft.fftshift网络模型的导出 HOT 9
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from paddlescience.