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

飞桨黑客马拉松第四期飞桨科学计算套件No.204设计文档RFC #490

Closed
wants to merge 57 commits into from

Conversation

v0919
Copy link

@v0919 v0919 commented Mar 29, 2023

薄板弯曲模块设计文档
一、概述
1、相关背景
PaddlePaddle Hackathon飞桨黑客马拉松第四期,赛道二:模型套件开源贡献飞桨科学计算套件,No.204:开放赛题:车辆标准部件受力、变形分析【PaddlePaddle Hackathon 4】模型套件开源贡献任务合集 · Issue #50631 · PaddlePaddle/Paddle (github.com)
2、功能目标
“本任务属于开放性赛题”,围绕汽车、飞机等装备的零部件,进行结构变形、受力分析。本课题选择汽车某标准部件,如下图所示。具体解决两类问题:
1)能够评估在给定任意的截面外形下,可得到不变的约束、负载条件下的结构受力结果,如应力、形变
2)能够基于目标的应力、变形,可以逆向优化得到部件截面的理想外形特征
3、意义
通过本次升级开发,使paddlescience具有直接面向工程结构的建模能力。
二、飞桨现状
飞桨目前具有求解PDE的功能,可以基于PDE建模,计算结构的应力和变形,但是这样不方便工程应用。
本次开发面向工程结构,应用工程师只需要输入结构参数、载荷及约束,就可以进行建模计算,得到需要的结果。
三、业内方案调研
目前求解PDE多用PINN或hPINN,PHYSICS-INFORMED NEURAL NETWORKS WITH HARD CONSTRAINTS FOR INVERSE DESIGN,arXiv:2102.04626v1。
对于多任务神经网络,采用GradNorm方法进行损失函数正则化处理,GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks,arXiv:1711.02257v4。
四、对比分析
本次开发利用第三部分所述的求解PDE的理论方法,用于工程结构的建模和计算。
五、设计思路与实现方案
工程结构往往是多个构件组成的组合结构,为了使软件具有灵活的建模能力,就有先对单个构件建模再组装成整个结构的需要。
例如由板组成的箱型结构,本次开发就是开发具有组装功能的板的建模和计算软件模块。在结构上板分为薄板和厚板,它们的PDE方程不同,本次开发主要做薄板。
为了进行结构组装,组装关系用硬约束实现,结构件之间能够进行坐标系转换,最终计算结果可以在整体坐标系和局部坐标系显示。
本次开发基于PadlleScience/DeepXDE,网络模型训练可以用PDE方程和数据进行。
六、测试和验收的考量
参考:新增API 测试及验收规范
七、可行性分析和排期规划
按赛程要求进行
八、影响面
对其他模块没有影响

heavyrain-lzy and others added 17 commits March 17, 2023 10:50
* 【Hackathon No.32】

为 Paddle 优化 expand_as 前向&反向 op 在 GPU 上的计算性能

* Update 20230305_expand_as_optimization.md

* Update 20230305_expand_as_optimization.md

* Update 20230305_expand_as_optimization.md

开头表格中的文件名写错了

* Update 20230305_expand_as_optimization.md

补充了fp16输入数据类型在paddle baseline以及tensorflow上的性能测试结果,并重新对两个框架的表现进行了对比分析。

---------

Co-authored-by: Hanchiao <46010859+Timber-Ye@users.noreply.github.com>
* update api doc

* update api doc

* update api doc format

* update api doc format

* update api doc format

* update test case

* update edge case

* update datatypes

* update api name

* update api doc
* upload the agenda for the 20230307 meeting

* upload meeting minutes for 20230307 meeting and correct a typo for 20230222 meeting minutes

* upload meeting minutes for 20230314 PPWG-TVM meeting
* 【Hackathon 4th No.13】为 Paddle 新增 Bernoulli API

* [Change]修改初始化参数及相关测试项,去掉logits;修改继承父类为ExponentialFamily
* Create 20220320_api_design_for_polygamma.md

* Create 20230320_api_design_for_polygamma.md

* apidoctest

* review0322

* review0323

* 'review0323(2)'

* 'review0323(3)'

* 'review0324'
* add sparse sum rfc

* fix sparse_backward_ops.yaml

* modify title

* modify api

* modify 7

* modified

* modify sec 7

* modify sec 7
@paddle-bot
Copy link

paddle-bot bot commented Mar 29, 2023

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.

zrr1999 and others added 4 commits April 6, 2023 11:54
* add meeting agenda

* 格式化 Contributor 列表

* update

* 添加新人

* 添加会议议程

* modify

* modify

* 添加会议链接

* add  meeting 19 minutes
@CLAassistant
Copy link

CLAassistant commented Apr 11, 2023

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
35 out of 36 committers have signed the CLA.

✅ gouzil
✅ heavyrain-lzy
✅ BrianQian1999
✅ Ligoml
✅ LyndonKong
✅ onecatcn
✅ cos43
✅ megemini
✅ luotao1
✅ ZwhElliott
✅ Sblue-G
✅ zrr1999
✅ KateJing1212
✅ PommesPeter
✅ TanWei
✅ zeroRains
✅ Patrick-Star125
✅ From00
✅ jeff41404
✅ yangguohao
✅ leiqing1
✅ MayYouBeProsperous
✅ unseenme
✅ thunder95
✅ lijingkai2023
✅ Tomoko-hjf
✅ WintersMontagne10335
✅ enkilee
✅ sanbuphy
✅ Zheng-Bicheng
✅ Asthestarsfalll
✅ qlogcn
✅ dasenCoding
✅ zhangyuqin1998
✅ TheFormerWalker
❌ XiaBing992
You have signed the CLA already but the status is still pending? Let us recheck it.

thunder95 and others added 23 commits April 11, 2023 13:59
* erfinv

* false commit

* erfinv

* false commit

* rfc for layoutxlm
* No.179:将PointPillars集成到FastDeploy,并在Jetson Orin硬件上部署验证精度和速度

* No.179:将PointPillars集成到FastDeploy,并在Jetson Orin硬件上部署验证精度和速度

* 标题修正
* Add PPOCRV3 Deploy PFC

* Add PPOCRV3 Deploy PFC

* update 20230221_add_solov2_for_fastdeploy

* update 20230228_pplcnet_deploy_for_arm

* update 20230228_pplcnet_deploy_for_arm

* update
* 使用java完成PaddleDetection部署提案

* deploy paddledetection by java

* modify RFC

* update RFC

* update RFC
…正确 (#497)

* Create 20230402_add_CANN_backend_for_FastDeploy.md

* Update 20230402_add_CANN_backend_for_FastDeploy.md

* Update 20230402_add_CANN_backend_for_FastDeploy.md
* Create 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md

* Update 20230331_add_MagicMind_backend_for_FastDeploy.md
* erfinv

* false commit

* erfinv

* false commit

* add rfc for slanet
* add rfc_nextafter

* fix rfc

* fix

* fix

* fix
* add 0302 meeting

* add 0302 meeting

* fix link

* fix

* fix

* fix

* fix

* Update 2023-03-02-meeting-agenda.md

* fix

* add pfcc-minutes

* add pfcc-minutes

* fix

* add CI

* fix

* fix

* fix
* [ reports ]add dev_guides

* [reports] Supplementary content

* [reports] Supplementary docs content

* [reports] Supplementary i18n content

* [reports] Fix errors

* [reports] Supplementary mindspore i18n content
* add 20230328_histogram_op_optimization.md

* Correct typos

* Correct typos

* modify the optimization way
* Create 2023-04-20-meeting-agenda.md

* Update 2023-04-20-meeting-agenda.md

---------

Co-authored-by: Tao Luo <luotao02@baidu.com>
为TVM PaddlePaddle前端完善 logical_or_cn/batch_norm/gaussian_random/one_hot_v2 算子支持程度,并完善算子测试。
* add_codegen_task2

* fix some comment
* 【Hackathon No.93】增加 linux 下 cpu tensor file_descriptor 传输方案

* Update 20230222_design_for_cpu_tensor_file_descriptor.md

* Update 20230222_design_for_cpu_tensor_file_descriptor.md
…式计算逻辑 (#514)

* erfinv

* false commit

* erfinv

* false commit

* ass softmax coo

* add cuda kernel design
@luotao1 luotao1 deleted the branch luotao1-patch-1 April 27, 2023 08:11
@luotao1 luotao1 closed this Apr 27, 2023
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.

None yet