BSE_Training_Tutorial

合集下载

MacTesseract4.1.1样本训练超详细教程

MacTesseract4.1.1样本训练超详细教程

MacTesseract4.1.1样本训练超详细教程Mac Tesseract 4.1.1 样本训练超详细教程乔布斯的橘⼦ 2021-03-17 01:40:17 483 收藏 2⽂章标签: opencv python 图像识别 ocr版权安装Mac直接安装tesseract的话⽆法附带安装training tools如果已经安装了没有training tools的tesseract,请先卸载brew uninstall tesseract先安装⼀些依赖的包# Packages which are always needed.brew install automake autoconf libtoolbrew install pkgconfigbrew install icu4cbrew install leptonica# Packages required for training tools.brew install pango# Optional packages for extra features.brew install libarchive# Optional package for builds using g++.brew install gcc从下列链接下载tesseract-4.1.1.tar.gz并解压编译并安装cd tesseract-4.1.1./autogen.shmkdir buildcd build# Optionally add CXX=g++-8 to the configure command if you really want to use a different compiler.../configure PKG_CONFIG_PATH=/usr/local/opt/icu4c/lib/pkgconfig:/usr/local/opt/libarchive/lib/pkgconfig:/usr/local/opt/libffi/lib/pkgconfig make -j# Optionally install Tesseract.sudo make install# Optionally build and install training tools.make trainingsudo make training-install下载完不会附带着⼀起下载数据集,通过下列链接⾃⾏下载需要的语⾔训练⾸先,收集数据样本(若⼲张需要训练的图⽚)图⽚格式需要转换为tif下载并打开jTessBoxEditor (注意,该软件需要java8环境,请⾃⾏配置):在jTessBoxEditor中Tools->Merge TIFF将所有tif⽂件合并将合并后的tif⽂件重命名为eng.num.exp0.tif⽣成box⽂件,⽤来纠正识别错误tesseract eng.num.exp0.tif eng.num.exp0 -l eng batch.nochop makebox此时,应该有eng.num.exp0.tif和eng.num.exp0.box两个⽂件使⽤jTessBoxEditor打开eng.num.exp0.tif(Box Editor->Open->eng.num.exp0.tif)纠正识别错误新建⼀个⽂件,取名font_properties,并填⼊下列内容font 0 0 0 0 0执⾏如下命令训练数据tesseract eng.num.exp0.tif eng.num.exp0 nobatch box.train unicharset_extractor eng.num.exp0.boxshapeclustering -F font_properties -U unicharset eng.num.exp0.tr mftraining -F font_properties -U unicharset -O unicharset eng.num.exp0.tr cntraining eng.num.exp0.trmv inttemp num.inttempmv normproto num.normprotomv pffmtable num.pffmtablemv shapetable num.shapetablemv unicharset num.unicharsetcombine_tessdata num.执⾏后,会有如下⽂件将num.traineddata移到相应路径便可使⽤我的路径是/usr/local/share/tessdata/。

Dynaform5.9BSE培训资料

Dynaform5.9BSE培训资料
DYNAFORM 5.9 Training Tutorial
坯料展开及排样
本教程通过Coat Hanger零件来描述 坯料尺寸展开计算以及排样的过程
I. 新建和保存数据库
1. 启动 Dynaform 5.9。 2. 单击文件菜单,选择另存为…子菜单(如图1所示)。 3. 输入“BSE_(user name)_(date).df ” 作为文件名。 4. 单击保存按钮保存数据库。
中(如图 26所示)。 3. 单击新建按钮在列表中新建一组轮廓(如图27所示)。 4. 单击导入线图标,并选择导入outline2.igs(如图28和29所示)。 5. 切换到Outline1组坯料轮廓线,程序自动切换当前显示的内容,把
Outline2 零件层关闭。 6. 单击 保存数据库。
20
VIII.坯料轮廓线管理
23
IX. 坯料排样
图 30
图 31
图 32
24
8. 察看排样结果
排样计算完成后,所有可能的排样结果都显示 在结果列表中。图形区中缺省显示的是在当前的 限制条件下,材料利用率最大的排样结果。用户 可以单击结果列表中的其他结果,在图形区中显 示出来。同时也可以返回到前面来详细的设置参 数,重新计算排样。如图33、图34所示。
图 13
10
V. 自动曲面网格划分
1. 选择 坯料工程 预处理 。 2. 切换到网格页面,选择曲面网格划分 (如图 14所示) 。 3. 系统默认的网格为Mstep Mesh,同时默认选中当前sheet工具中所有零
件层的曲面(如图15所示) 。 4. 在参数中输入最大尺寸, 2.00 (mm) 。 5. 单击按钮应用进行网格划分。 6. 单击是接受划分的网格。 7. 单击退出按钮退出曲面网格划分对话框。 8. 如图16所示 。

stable diffusion webui 训练方法

stable diffusion webui 训练方法

stable diffusion webui 训练方法
Stable Diffusion WebUI是一个用于训练Stable Diffusion模型
的用户界面。

稳定扩散(Stable Diffusion)是一种生成模型训
练方法,它基于扩散过程来生成符合数据分布的样本。

下面是一般的Stable Diffusion模型训练方法,可以通过Stable Diffusion WebUI进行操作:
1. 数据准备:首先,需要准备用于训练的数据集。

数据集应该包含符合所需分布的样本数据。

2. 模型架构选择:选择合适的模型架构来训练Stable Diffusion
模型。

常见的选择包括变分自动编码器(VAE)和生成对抗
网络(GAN)。

3. 模型参数设置:设置模型的超参数,如扩散步数、学习率等。

这些参数的选择可能会对模型的性能产生影响。

4. 网络训练:使用稳定扩散算法训练模型。

首先,通过生成噪声样本,利用模型生成初始样本。

然后,通过不断迭代的过程,逐步调整生成样本与真实样本之间的差异,直到生成样本的分布逼近真实样本的分布。

5. 模型评估:在训练过程中,可以通过计算生成样本与真实样本之间的差异或使用其他评估指标来评估模型的性能。

Stable Diffusion WebUI可以帮助用户以图形化界面的方式完成
上述步骤,并提供可视化工具来监控模型训练过程。

用户可以通过这个界面调整参数、查看训练进度和结果,并进行交互式地探索生成的样本。

(2024年)ABAP培训教程

(2024年)ABAP培训教程
2024/3/26
修改表
使用ALTER TABLE语句修 改已存在的数据库表结构, 如添加、删除或修改字段。
删除表
使用DROP TABLE语句删 除指定的数据库表。
15
数据库查询优化
索引优化
根据查询条件和数据量大小,合 理创建和使用索引,提高查询效
率。
2024/3/26
SQL语句优化
编写高效的SQL查询语句,避免全 表扫描和不必要的JOIN操作。
在ABAP中,使用 `INHERITING FROM`关键字 实现继承,通过方法重写实现 多态。
21
封装与接口
封装(Encapsulation)将数据 和行为捆绑在一起,隐藏内部实 现细节,只暴露必要的接口。
2024/3/26
接口(Interface)定义了一组方 法的规范,但不提供具体实现。 实现接口的类必须提供这些方法
01
设计报表的布局,包括表格、图表、文本等元素的排列和组合。
格式化
02
对报表中的数据进行格式化处理,如日期、数字、货币等格式
的显示。
样式设置
03
设置报表的样式,包括字体、颜色、背景等,以提高报表的可
读性和美观度。
26
报表输出与打印
2024/3/26
输出格式
支持多种输出格式,如PDF、Excel、Word等,以满足不同用户 的需求。
用于ABAP程序的开发、调试和执行,提 供了丰富的开发工具和界面设计功能。
SAP Web IDE
基于Web的集成开发环境,支持ABAP和 其他SAP开发语言,提供了代码编辑、版 本控制、团队协作等功能。
Eclipse ADT
基于Eclipse平台的ABAP开发工具,提供 了代码自动补全、语法检查、调试等功能。

bsr训练方法

bsr训练方法

bsr训练方法
BSR(Bilingual Sentence Representation)是一种训练方法,用
于将双语句子映射到一个高维的共享语义空间中的相似表示。

BSR的训练方法主要包括以下步骤:
1. 数据准备:首先需要准备成对的双语语料数据,包括源语言句子和目标语言句子。

2. 双语句子对齐:使用对齐工具将双语句子进行对齐,确保每个源语言句子都能找到对应的目标语言句子。

3. 双语句子编码器:建立一个双语句子编码器,它能够将源语言句子和目标语言句子分别映射到一个共享的语义空间中的表示。

4. 双语句子对损失函数:定义一个双语句子对损失函数,用来度量源语言句子和目标语言句子的相似度。

5. 迭代训练:使用双语句子对损失函数对编码器进行优化训练,通过反向传播算法更新参数,使得源语言句子和目标语言句子在共享语义空间中的表示能够更加接近。

6. 相似度计算:在训练过程中,可以使用一些相似度计算方法,比如余弦相似度,来度量不同句子在共享语义空间中的相似程度。

7. 性能评估:使用一些评估指标,比如准确率、召回率等,来
评估训练出的双语句子表示在不同任务上的性能。

通过以上步骤,BSR训练方法可以得到一个高质量的双语句子表示,这种表示能够在多个自然语言处理任务中发挥作用,比如机器翻译、文本对齐、语义匹配等。

gromacstutorial

gromacstutorial

gromacstutorialGROMACS TutorialStep One: Prepare the Protein TopologyWe must download the protein structure file we will be working with. For this tutorial, we will utilize T4 lysozyme L99A/M102Q (PDB code 3HTB). Go to the RCSB website and download the PDB text for the crystal structure.Once you have downloaded the structure, you can visualize it using a viewing program such as VMD, Chimera, PyMOL, etc. Once you've had a look at the molecule, you are going to want to strip out the crystal waters, PO4, and BME. Note that such a procedure is not universally appropriate (i.e., the case of a bound active site water molecule). For our intentions here, we do not need crystal water or other ligands. We will instead focus on the ligand called "JZ4."If you want a cleaned version of the .pdb file to check your work, you can download it here. The problem we now face is that the JZ4 ligand is not a recognized entity in any of the force fields provided with GROMACS, so pdb2gmx will give a fatal error if you were try to pass this file through it. Topologies can only be assembled automatically if an entry for a building block is present in the .rtp file for the force field. Since this is not the case, we will prepare our system topology in two steps:1.Prepare the protein topology with pdb2gmx2.Prepare the ligand topology using external toolsSince we will be preparing these two topologies separately, we must move the protein and JZ4 into separate coordinate files. Save the JZ4 coordinates like so:grep JZ4 3HTB_clean.pdb > JZ4.pdbThen simply delete the JZ4 lines from 3HTB_clean.pdb. At this point, preparing the protein topology is trivial. There are no missing atoms or residues in the 3HTB structure, so simply run pdb2gmx:pdb2gmx -f 3HTB_clean.pdb -o 3HTB_processed.gro -water spcThe structure will be processed by pdb2gmx, and you will be prompted to choose a force field:Select the Force Field:From '/usr/local/gromacs/share/gromacs/top':1: AMBER03 force field (Duan et al., J. Comp. Chem. 24, 1999-2012, 2003)2: AMBER94 force field (Cornell et al., JACS 117, 5179-5197, 1995)3: AMBER96 force field (Kollman et al., Acc. Chem. Res. 29, 461-469, 1996)4: AMBER99 force field (Wang et al., J. Comp. Chem. 21, 1049-1074, 2000)5: AMBER99SB force field (Hornak et al., Proteins 65, 712-725, 2006)6: AMBER99SB-ILDN force field (Lindorff-Larsen et al., Proteins 78, 1950-58, 2010)7: AMBERGS force field (Garcia & Sanbonmatsu, PNAS 99, 2782-2787, 2002)8: CHARMM27 all-atom force field (with CMAP) - version 2.09: GROMOS96 43a1 force field10: GROMOS96 43a2 force field (improved alkane dihedrals)11: GROMOS96 45a3 force field (Schuler JCC 2001 22 1205)12: GROMOS96 53a5 force field (JCC 2004 vol 25 pag 1656)13: GROMOS96 53a6 force field (JCC 2004 vol 25 pag 1656)14: OPLS-AA/L all-atom force field (2001 aminoacid dihedrals)15: [DEPRECATED] Encad all-atom force field, using full solvent charges16: [DEPRECATED] Encad all-atom force field, using scaled-down vacuum charges17: [DEPRECATED] Gromacs force field (see manual)18: [DEPRECATED] Gromacs force field with hydrogens for NMRFor this tutorial, we will use the united-atom GROMOS96 43A1 force field, so type 9 at the command prompt, followed by'Enter'Step Two: Prepare the Ligand TopologyFor this tutorial, we will use PRODRG to generate a starting topology for our ligand, JZ4. Go to the PRODRG site and upload your JZ4.pdb file. The server presents you with several options for how to treat your ligand:Chirality (yes/no): maintain chiral centers in the input molecule (yes), or reconstruct them (no). In our case, we have no chirality, so leave this option set to its default.Charges (full/reduced): full charges are for condensed-phase systems (43A1 force field);reduced are for "vacuum" simulations (43B1 force field). Use full charges for nearly allapplications. The accuracy of 43B1 is debatable, anyway.EM (yes/no): perform energy minimization (yes), or leave coordinates alone (no). In our case, we want to construct our protein-ligand complex from existing coordinates, so wedo not want PRODRG to minimize our molecule in vacuo. Choose "no" for this option.Force field (GROMOS96.1/GROMOS87): "GROMOS96.1" refers to the first version of the GROMOS96 force field, 43A1. "GROMOS87" refers to the (outdated!) GROMOS87 force field. Choose "GROMOS96.1" to get 43A1 parameters for our ligand.We will make use of two files that PRODRG gives us. Save the output of the field "The GROMOS87/GROMACS coordinate file (polar/aromatic hydrogens)" into a text file called"jz4.gro" and "The GROMACS topology" into a file called "drg.itp."Let's take a look at the [ atoms ] section of our JZ4 topology:[ atoms ]; nr type resnr resid atom cgnr charge mass1 CH3 1 JZ4 C4 1 0.000 15.03502 CH2 1 JZ4 C14 2 0.059 14.02703 CH2 1 JZ4 C13 2 0.060 14.02704 C 1 JZ4 C12 2 -0.041 12.01105 CR1 1 JZ4 C11 2 -0.026 12.01106 HC 1 JZ4 H11 2 0.006 1.00807 CR1 1 JZ4 C7 2 -0.026 12.01108 HC 1 JZ4 H7 2 0.006 1.00809 CR1 1 JZ4 C8 2 -0.026 12.011011 CR1 1 JZ4 C9 2 -0.026 12.011012 HC 1 JZ4 H9 2 0.007 1.008013 C 1 JZ4 C10 3 0.137 12.011014 OA 1 JZ4 OAB 3 -0.172 15.999415 H 1 JZ4 HAB 3 0.035 1.0080I immediately notice three things that are wrong with this topology:1.Charge group 2 encompasses the entire aromatic ring, nearly all of the atoms in thismolecule.2.Charges of equivalent functional groups (C-H) are not uniform. In some cases, H atomsare +0.006e, and in other cases they are +0.007e.3.Charges on these atoms bear no resemblance whatsoever to the expected chargesprescribed by the force field.So what do we do? If you read the paper linked above, you would know what I'm going to recommend already. Assign charges and charge groups from equivalent functional groups in a building block-style manner. For any unknown groups, perform the charge calculations suggested in the paper. Since JZ4 involves only functional groups found in proteins (hydrophobic chain, aromatic ring, and an alcohol), life is pretty easy. We can re-assign charges and charge groups from these moieties. I will not go into proper topology validation in this tutorial; it would be far too time-consuming. Suffice it to say that you must satisfy reviewers that the parameters applied to your molecule are reliable. They should reproduce some well-known observable, such as logP or ΔG hydr, as in the GROMOS96 literature. If you cannot satisfy these requirements for your own system, you may want to reconsider your choice of force field.I would construct a topology that has something like this for an [ atoms ] directive:[ atoms ]; nr type resnr resid atom cgnr charge mass1 CH3 1 JZ4 C4 1 0.000 15.03502 CH2 1 JZ4 C14 2 0.000 14.02703 CH2 1 JZ4 C13 2 0.000 14.02704 C 1 JZ4 C12 2 0.000 12.01105 CR1 1 JZ4 C11 3 -0.100 12.01106 HC 1 JZ4 H11 3 0.100 1.00807 CR1 1 JZ4 C7 4 -0.100 12.01108 HC 1 JZ4 H7 4 0.100 1.00809 CR1 1 JZ4 C8 5 -0.100 12.011010 HC 1 JZ4 H8 5 0.100 1.008011 CR1 1 JZ4 C9 6 -0.100 12.011012 HC 1 JZ4 H9 6 0.100 1.008013 C 1 JZ4 C10 7 0.150 12.011014 OA 1 JZ4 OAB 7 -0.548 15.9994Note that equivalent functional groups have equivalent charges, hydrophobic groups are completely uncharged, and the charge groups are of a sensible size (2-3 atoms). The other elements of the topology are sufficiently accurate. Bonded parameters assigned by PRODRG are generally reliable. We are now ready to construct the system topology and reconstruct our protein-ligand complex.Build the ComplexFrom pdb2gmx, we have a file called "conf.gro" that contains the processed, force field-compliant structure of our protein. We also have "jz4.gro" from PRODRG that has included all of the necessary H atoms. Simply copy the coordinate section ofjz4.gro and paste it into conf.gro, below the last line of the protein atoms, and before the box vectors, like so:163ASN C 1691 0.621 -0.740 -0.126163ASN O1 1692 0.624 -0.616 -0.140163ASN O2 1693 0.683 -0.703 -0.0115.99500 5.19182 9.66100 0.00000 0.00000 -2.99750 0.00000 0.00000 0.00000becomes (added text in bold green)...163ASN C 1691 0.621 -0.740 -0.126163ASN O1 1692 0.624 -0.616 -0.140163ASN O2 1693 0.683 -0.703 -0.0111JZ4 C4 1 2.429 -2.412 -0.0071JZ4 C14 2 2.392 -2.470 -0.1391JZ4 C13 3 2.246 -2.441 -0.1811JZ4 C12 4 2.229 -2.519 -0.3081JZ4 C11 5 2.169 -2.646 -0.2951JZ4 H11 6 2.135 -2.683 -0.1991JZ4 C7 7 2.155 -2.721 -0.4111JZ4 H7 8 2.104 -2.817 -0.4071JZ4 C8 9 2.207 -2.675 -0.5331JZ4 H8 10 2.199 -2.738 -0.6211JZ4 C9 11 2.267 -2.551 -0.5451JZ4 H9 12 2.306 -2.516 -0.6401JZ4 C10 13 2.277 -2.473 -0.4301JZ4 OAB 14 2.341 -2.354 -0.4341JZ4 HAB 15 2.369 -2.334 -0.5285.99500 5.19182 9.66100 0.00000 0.00000 -2.99750 0.00000 0.00000 0.00000Since we have added 15 more atoms into the .gro file, increment the second line of conf.gro to reflect this change. There should be 1708 atoms in the coordinate file now.Build the TopologyIncluding the parameters for the JZ4 ligand in the system topology is very easy. Just insert a line that says #include "drg.itp" into topol.top after the position restraint file is included. The inclusion of position restraints indicates the end of the "Protein" moleculetype section.#ifdef POSRES#include "posre.itp"#endif; Include water topology#include "gromos43a1.ff/spc.itp"becomes...; Include Position restraint file#ifdef POSRES#include "posre.itp"#endif; Include ligand topology#include "drg.itp"; Include water topology#include "gromos43a1.ff/spc.itp"The last adjustment to be made is in the [ molecules ] directive. To account for the fact that there is a new molecule in conf.gro, we have to add it here, like so:[ molecules ]; Compound #molsProtein_chain_A 1JZ4 1The topology and coordinate file are now in agreement with respect to the contents of the system. Step Three: Defining the Unit Cell & Adding SolventAt this point, the workflow is just like any other MD simulation. We will define the unit cell andfill it with water.editconf -f conf.gro -o newbox.gro -bt dodecahedron -d 1.0genbox -cp newbox.gro -cs spc216.gro -p topol.top -o solv.groStep Four: Adding IonsWe now have a solvated system that contains a charged protein. The output of pdb2gmx told us that the protein has a net charge of +6e (based on its amino acid composition). If you missed this information in the pdb2gmx output, look at the last line of your [ atoms ] directive in topol.top; it should read (in part) qtot。

沃尔沃英文缩写含义清单-20150604

沃尔沃英文缩写含义清单-20150604

第 1 页,共 76 页
AME AMEC AMK AMP AMPPE AMS AMTD ANC ANOM ANOVA ANS ANSI ANX AO AOC AOS AP APD APEAL APEL APG API APP APQP APR APS AQL AQS AR ARB ARL ARM ARO ARP ARPA ARS AS ASA ASAP ASC ASDM ASESO ASIC ASIL ASM ASN ASP ASR ASSY ASTA ASTM AT ATACQ ATC ATEO ATI ATM ATP
​Advanced Manufacturing Engineering ​After Market Executive Committee ​W ork Environment Committee ​After Market Program ​After Market Pre-Program Engineering ​Application Management Services ​Advanced Manufacturing Technology and Development ​Active Noise Cancellation ​Analysis Of Means ​Analysis Of Variation ​Alarm Notification System ​American National Standards Institute ​Automotive Network Exchange ​Application Owner ​Active On-demand Coupling A.Attribute utomotivePrototype Occupancy Sensor Process 3. Accelerator ​1 2. Annual Pedal ​Availability Performance Data ​Automotive Performance Execution & Layout Study ​Assessment for Prior Experience & Learning ​Arizona Proving Ground Application Program Interface ​Availability Performance Program ​Advanced Product Quality Planning ​Annual Percentage Rate ​Affordable Personel Structure ​Accepted Quality Level ​Air Quality Sensor ​Application Ready ​Architecture Review Board ​Analytical Road Loads ​Accelerated Revitalization Management ​Assembly Receive Order ​1. Application Recovery Plan 2. Abrasive Resistant Paint ​Advanced Research Projects Agency ​1. Active Rear Steering 2. Active Roll Stabilization ​1. Automatic Synchronizer 2. Automatic System ​After Sales Analysis ​As Soon As Possible ​1.Adaptive Steering Column 2. Active Short Cut ​Air Bag System Diagnostic Module ​Automotive Safety Engineering Standards Office ​Application Specified Integrated Circuit ​A Safety Integrity Level 1.utomotive Automative Strategy Meeting 2. Active Safety Management ​Advanced Shipping Notice ​Active Server Pages ​Attribute Status Report ​Assembly ​Active Safety Test Arena ​American Society for Testing And Material ​Automatic Transmission ​Answers To All Car Questions ​Automatic Temperature Control ​Automatic Transmission Engineering Office ​Advanced Technical Information ​1.Antenna Module 2. Antenna Tuner Module ​Affordable Target Process

python基础教程英文版

python基础教程英文版

python基础教程英文版A Python Basic Tutorial (English Version)。

Python is a widely-used high-level programming language known for its simplicity and readability. It is anexcellent choice for beginners who want to learn programming. In this tutorial, we will cover the basics of Python programming.1. Introduction to Python:Python is an interpreted language, which means that the code is executed line by line. It does not require a compilation step, making it easy to write and test code quickly.2. Installation:To get started with Python, you need to install the Python interpreter on your computer. You can download thelatest version of Python from the official website andfollow the installation instructions.3. Variables and Data Types:In Python, you can create variables to store data. Python supports various data types such as integers, floats, strings, booleans, lists, tuples, and dictionaries. Understanding these data types is crucial for writing effective code.4. Operators:Python provides a range of operators for performing arithmetic, comparison, logical, and assignment operations. These operators allow you to manipulate data and controlthe flow of your program.5. Control Flow:Control flow statements, such as if-else, for loops,and while loops, enable you to control the execution ofyour code based on certain conditions. These statements help in making decisions and iterating over data.6. Functions:Functions are reusable blocks of code that perform specific tasks. They help in organizing code and making it more modular. Python allows you to define and call functions, passing arguments and returning values.7. Modules and Packages:Python has a vast collection of built-in modules and packages that provide additional functionality. You can import these modules into your code and use their functions and classes. Additionally, you can create your own modules for code reusability.8. File Handling:Python provides various functions and methods for working with files. You can open, read, write, and closefiles using built-in file handling operations. Understanding file handling is essential for dealing with data stored in files.9. Exception Handling:Exception handling allows you to catch and handleerrors that may occur during the execution of your program. Python provides try-except blocks to handle exceptions gracefully and prevent your program from crashing.10. Object-Oriented Programming (OOP):Python supports object-oriented programming, which allows you to create classes and objects. OOP helps in organizing code and implementing complex systems byutilizing concepts such as inheritance, encapsulation, and polymorphism.11. Libraries and Frameworks:Python has a vast ecosystem of libraries and frameworksthat extend its capabilities. Libraries like NumPy, Pandas, and Matplotlib are widely used for scientific computing and data analysis. Frameworks like Django and Flask are popular for web development.12. Debugging and Testing:Python provides tools and techniques for debugging and testing your code. You can use debugging tools to find and fix errors in your program. Testing frameworks likeunittest and pytest help in writing automated tests to ensure the correctness of your code.In conclusion, this Python basic tutorial provides an overview of the fundamental concepts and features of the Python programming language. By understanding these concepts, you will be well-equipped to start writing your own Python programs and explore more advanced topics. Remember to practice writing code and experimenting with different examples to enhance your learning experience.。

训练冻结bn层参数

训练冻结bn层参数

训练冻结bn层参数
要训练冻结BN层参数,需要在网络的 BN 层后添加一个
Scale 层。

Scale 层中的参数可以进行训练,并在训练过程中对BN 层的结果进行缩放,从而达到冻结 BN 层参数的目的。

具体步骤如下:
1. 将原来的 BN 层替换为 Scale 层。

在网络结构中找到所有的BN 层,并将其替换为Scale 层。

可以使用工具库(如pytorch)提供的函数或自行实现。

2. 将 Scale 层的训练标志设置为 True。

将 Scale 层的
`requires_grad` 属性设置为True,表示需要对其参数进行训练。

3. 将 BN 层的训练标志设置为 False。

将 BN 层的
`requires_grad` 属性设置为 False,表示不需要对其参数进行训练。

4. 将整个网络的参数划分为可训练和不可训练部分。

将整个网络的参数划分为两个部分,一个是需要训练的参数(包括Scale 层和其他需要训练的层参数),一个是不需要训练的参
数(包括 BN 层参数和其他不需要训练的层参数)。

5. 设置优化器。

使用优化器来更新可训练部分的参数。

可以根据需要选择适合的优化器,如Adam、SGD等。

6. 进行训练。

使用训练数据对网络进行训练,更新可训练部分
的参数。

通过以上步骤,可以训练冻结 BN 层参数的网络模型。

需要注意的是,在训练过程中,BN 层的统计量(均值和方差)仍然是通过输入数据动态计算得出的,Scale 层只对 BN 层的结果进行缩放,不会更新 BN 层的统计量。

c++ 信奥赛 常用英语

c++ 信奥赛 常用英语

c++ 信奥赛常用英语在C++ 信奥赛中(计算机奥林匹克竞赛),常用英语词汇主要包括以下几方面:1. 基本概念:- Algorithm(算法)- Data structure(数据结构)- Programming language(编程语言)- C++(C++ 编程语言)- Object-oriented(面向对象)- Function(函数)- Variable(变量)- Constants(常量)- Loops(循环)- Conditional statements(条件语句)- Operators(运算符)- Control structures(控制结构)- Memory management(内存管理)2. 常用算法与数据结构:- Sorting algorithms(排序算法)- Searching algorithms(搜索算法)- Graph algorithms(图算法)- Tree algorithms(树算法)- Dynamic programming(动态规划)- Backtracking(回溯)- Brute force(暴力破解)- Divide and conquer(分治)- Greedy algorithms(贪心算法)- Integer array(整数数组)- Linked list(链表)- Stack(栈)- Queue(队列)- Tree(树)- Graph(图)3. 编程实践:- Code optimization(代码优化)- Debugging(调试)- Testing(测试)- Time complexity(时间复杂度)- Space complexity(空间复杂度)- Input/output(输入/输出)- File handling(文件处理)- Console output(控制台输出)4. 竞赛相关:- IOI(国际信息学奥林匹克竞赛)- NOI(全国信息学奥林匹克竞赛)- ACM-ICPC(ACM 国际大学生程序设计竞赛)- Codeforces(代码力)- LeetCode(力扣)- HackerRank(黑客排名)这些英语词汇在信奥赛领域具有广泛的应用,掌握这些词汇有助于提高选手之间的交流效率,同时对提升编程能力和竞赛成绩也有很大帮助。

Tutorial_Guide_IGG_82_1-Acrov5

Tutorial_Guide_IGG_82_1-Acrov5

TutorialsIGG™ v8.aDocumentation v8.aNUMECA International5, Avenue Franklin Roosevelt1050 BrusselsBelgiumTel: +32 2 647.83.11Fax: +32 2 647.93.98Web: ContentsTABLE OF CONTENTINTRODUCTIONTUTORIAL 1: Geometry Creation1-1 INTRODUCTION1-11-1.1 Introduction1-11-1.2 Prerequisites1-21-1.3 Preparation1-21-2 CARTESIAN POINT1-31-2.1 Create Cartesian Point1-31-2.2 Select Cartesian Point1-31-2.3 Delete Cartesian Point1-31-3 CURVES1-31-3.1 Create Curves1-31-3.2 Select Curves1-51-3.3 Visualize Curves1-51-3.4 Modify Curves1-61-3.5 Edit/Copy Curves1-71-3.6 Export Curves1-71-3 SURFACES1-81-3.1 Create Surfaces1-81-3.2 Select Surfaces1-101-3.3 Visualize Surfaces1-101-3.4 Modify Surfaces1-111-3.5 Edit/Copy Surfaces1-111-3.6 Export Surfaces1-11 TUTORIAL 2: 2D Airfoil Mesh Generation2-1 INTRODUCTION2-12-1.1 Introduction2-12-1.2 Prerequisites2-22-1.3 Presentation2-22-1.4 Preparation2-22-2 MESH GENERATION2-32-2.1 Define Project Configuration2-42-2.2 Define Geometry2-52-2.3 Create Blocks2-62-2.4 Define Clustering2-112-2.5 Generate Face Grid2-142-2.6 Control Mesh Quality2-162-2.7 Define Boundary Conditions2-172-2.8 Save Project2-18ContentsTUTORIAL 3: Pipe to Pipe Mesh Generation3-1 INTRODUCTION3-13-1.1 Introduction3-13-1.2 Prerequisites3-23-1.3 Presentation3-23-1.4 Preparation3-23-2 MESH GENERATION3-33-2.1 Define Geometry3-33-2.2 Create & Control Blocks3-53-2.3 Generate Block Grid3-133-2.4 Define Butterfly Topology3-143-2.5 Control Mesh Quality3-163-2.6 Define Boundary Conditions3-183-2.7 Define Full Non Matching Connection3-193-2.8 Save Project3-20TUTORIAL 4: Volute Mesh Generation4-1 INTRODUCTION4-14-1.1 Introduction4-14-1.2 Prerequisites4-24-1.3 Presentation4-24-1.4 Preparation4-24-2 MESH GENERATION4-34-2.1 Load Geometry4-34-2.2 Create & Control Blocks4-44-2.3 Generate Block Grid4-184-2.4 Control Mesh Quality4-194-2.5 Define Boundary Conditions4-204-2.6 Define Full Non Matching Connection4-224-2.7 Save Project4-23TUTORIAL 5: Axi Seal Leakage Mesh Generation5-1 INTRODUCTION5-15-1.1 Introduction5-15-1.2 Prerequisites5-25-1.3 Presentation5-25-1.4 Preparation5-25-2 MESH GENERATION5-35-2.1 Define Project Configuration5-35-2.2 Import Geometry5-45-2.3 Create & Control Blocks5-45-2.4 Define Clustering5-105-2.5 Control Mesh Quality5-135-2.6 Define Boundary Conditions5-155-2.7 Save Project5-18What’s in This Guide ?This Tutorial Guide contains a number of tutorials driving the user in IGG™ v8 to mesh different internal and external configurations. In each tutorial, features related to mesh generation are dem-onstrated.Tutorials 1 to 5 are detailed tutorials designed to introduce the beginner to IGG™ v8. These tutori-als provide explicit instructions for all steps of the mesh generation process. Tutorials 1 to 5 do not require any pre-requisite and can be treated separately, in any order. They address different types of applications, including both internal and external cases.Where to Find the Files Used in the Tutorials ?Each of the mesh generation starts from a geometry that is existing or is created. The appropriate files (and any other relevant files used in the tutorial) are stored on IGG™ v8 DVD-ROM, more precisely in the /DOC/_Tutorials directory.How to Use this Guide ?Depending upon your familiarity with computational fluid dynamics and your interest in some par-ticular configuration, you can use this tutorial guide in a variety of ways.For the BeginnerIf you are beginning user of IGG™, you should first read and solve tutorials 1 and 2, in order to familiarize yourself with the interface and basis of the mesh generation technique. You may then want to concentrate on a tutorial that demonstrates features that you are going to resolve. For exam-ple, if you are planning to mesh a volute, you should look at tutorial 4.For the Experienced UserIf you are an experienced user of IGG™, you can read and/or solve the tutorial(s) that demonstrate features that you are going to resolve. For example, if you plan to mesh a 2D airfoil, you should look at tutorial 2.Conventions Used in this GuideSeveral conventions are used in the tutorials to facilitate your learning process.Following a short introduction, each tutorial is divided into sections respectively related to the mesh generation steps from the geometry definition to the 3D mesh generation.Inputs required to execute the tutorials are restricted to the geometry, either in a ".dat" or CAD related format.The sequence of actions to be executed are described through a step-by-step approach, in the form of arabic numbers.Additional insight about some specific actions and/or features is frequently added to illustrate the tutorial further. This information is proposed for the purpose of clarity and completeness, and should not be executed. It appears in italicized type.Contact NUMECA support team at +32-2-647.83.11 or send an e-mail to sup-port@numeca.be for any question or information you may require. To allowNUMECA support to help you out within the shortest delays, please provide adetailed description of the observed behaviour and performed analysis.TUTORIAL 1:Geometry Creation1-1Introduction1-1.1IntroductionThe resolution of computational fluid dynamics (CFD) problems involves three main steps:•spatial discretization of the flow equations,•flow computation,•visualization of the results.To answer these questions, NUMECA has developed a F low IN tegrated E nvironment for internaland Turbomachinery assimilations. Called FINE™/Turbo, the environment integrates the followingtools:•IGG™ is an I nteractive G eometry modeler and G rid generator software, based on structured multi-block techniques,•AutoGrid™ is a three-dimensional Automated Grid generation software, dedicated to turboma-chinery applications. Similarly to IGG™, it is based on structured multi-block techniques,•Euranus is a state-of-the-art multi-block flow solver, able to simulate Euler and Navier-Stokes equations in the laminar, transitional and turbulent regimes,•CFView™ is a highly interactive flow visualization and post-treatment software,•FINE™ Graphical User Interface is a user-friendly environment that includes the different soft-wares. It integrates the concept of projects and allows the user to achieve complete simulations,going from the grid generation to the flow visualization, without the need of file manipulation.This tutorial is particularly adapted to the creation and modification of geometrical entities. Itmakes exclusive use of IGG™.In this tutorial you will learn how to:•Create Cartesian point,•Create and modify curve entities,•Create and modify surface entities,•Select and delete geometrical entities,Geometry Creation Introduction•Group/ungroup geometrical entities,•Save geometrical entities.1-1.2PrerequisitesThis tutorial does not require any particular prerequisite.1-1.3Preparation•Copy the files located in cdrom:\DOC\_Tutorials\IGG\Tutorial_1 to your working directory, where cdrom must be replaced by the name of your DVD-ROM.•Start IGG™ v8.xFor LINUX and UNIX systems, you can access IGG™ v8.x graphical user interface with thefollowing command lineigg -niversion 8x -print or igg -niversion autogrid8x -printFor WINDOWS systems, you can access IGG™ v8.x graphical user interface from the startmenu going to /Programs/NUMECA software/fine8x/IGG or /Programs/NUMECA software/autogrid8x/IGGMenu BarTool Bar3D ViewQuick Access PadControl Areakeyboard input areainformation areaYou’re now ready to start to create and modify geometrical entities!IGG™ v8 graphical user interface allows to visualize the geometry and mesh of the internal orexternal test case in a 3D view by default. The access to main menu and controls is proposedthrough a menu bar and a quick access pad, and is completed with a tool/icon bar and a control area(including the keyboard input area).Cartesian Point Geometry Creation1-2Cartesian Point1-2.1Create Cartesian Point1.Select the Geometry/Create Points/Cartesian Point menu to initiate the creation of aCartesian point2.Type the sequence <1 1 0> <Enter> in the keyboard input area. This action will create theCartesian point (black or white point is appearing in the graphics area)Cartesian points can also be defined as intersection between two selectedcurves or between a selected curve and a plane or between a selected curveand a surface (see User Manual for more details).3.Select the Geometry/Create Points/Cartesian Point menu to initiate the creation of a sec-ond Cartesian point4.Type the sequence <1 1 1> <Enter> in the keyboard input area. This action will create thesecond Cartesian point (second black or white point is appearing in the graphics area) 1-2.2Select Cartesian Point5.Select the Geometry/Select/Cartesian Points menu to select Cartesian points6.Move the mouse on the Cartesian point (1,1,1) and click-left on it when highlighted in blueto select it7.Click-right or <q> in the graphics area to end the selection1-2.3Delete Cartesian Point8.Select the Geometry/Delete/Cartesian Points menu to delete the selected Cartesian points(highlighted in blue)1-3Curves1-3.1Create CurvesThe following section describes how to:—create basic curves—use the keyboard or the mouse to input points—use the attraction featureThe below geometry, consisting of two polylines, one C-spline and one arc, will be created.Geometry Creation Curves9.Define a polyline curve:•Select the Geometry/Draw Polyline/Free menu (shortcut <p >) to initiate the creation of a polyline•Type the sequence <1 0 0> <Enter > in the keyboard input area . This action will create thefirst point of the polylineThe keystrokes are automatically echoed in the keyboard input area.•Enter a second point at position <1.2 0.5 0> and press <Enter >•Move the mouse near the Cartesian point. When close enough, the mouse will normally beattracted to this point if the attraction to points feature is enabled. If there is no attraction,press <a > in the graphics area. Then click-left to add this point to the polyline•Click-right or <q > in the graphics area to end the polyline creation•Repeat above steps to create another polyline passing through the points (0,0,0), (-0.2,0.5,0)and (0,1,0)10.Define a C-spline curve:•Select the Geometry/Draw CSpline/Free menu (shortcut <c >) to initiate the creation of aC-spline curve•Move the mouse near the point (0,0,0) of the second polyline. When close enough, themouse will normally be attracted to this point if the attraction to points feature is enabled. If there is no attraction, press <a > in the graphics area. Then click-left to add this point to the C-spline•Move the mouse somewhere between the points (0,0,0) and (1,0,0) and click-left to add apoint•Move the mouse near the point (1,0,0) of the first polyline. When close enough, the mousewill normally be attracted to this point if the attraction to points feature is enabled. If there is no attraction, press <a > in the graphics area. Then click-left to add this point to the C-spline•Click-right or <q > in the graphics area to end the C-spline creation11.Define a circular arc curve:•Select the Geometry/Circular Arc/Normal-Point-Point-Radius menu option to initiatethe creation of a circular arc. Several inputs will be requested to define the arcThe circular arc can be created using different methods (see User Manual for more details).•Enter <0 0 1> <Enter > to define the arc normalpolyline 1polyline2C-splinearcCurves Geometry Creation •Move the mouse near the Cartesian point. When close enough the point will be highlighted (if there is no attraction, press <a> in the graphics area). Click-left to define the arc startpoint•Move the mouse near the point (0,1,0) of the second polyline. When close enough the point will be highlighted (if there is no attraction, press <a> in the graphics area). Click-left todefine the arc end point•Enter <0.6> <Enter> to define the arc radius•Press <o> until the circle has the same shape as the one presented on above figure. Then click-left to create the arcClick-right or <q> in the graphics area to end the arc creation.1-3.2Select CurvesThe curve selection operation is used to activate one or more curves for subsequent operations ingeometry modelling or grid generation. When a curve is selected it appears highlighted in yellow(default). All the curves created in the previous steps are selected.12.Select the Geometry/Select/Curves option to initiate curve(s) selectionThe shortcut <s> can also be used to activate the option without accessing themenu.13.Press <a> to unselect all the curves, which become unhighlighted14.Move the mouse over the C-spline which is then highlighted. At the same time, the name,type of curve and approximate arc length of the curve appear in the information area15.Click-left to select it16.Repeat above step to select the first created polyline17.Click-right to quit the selectionSelection and deselection of all curves can be done by pressing <a> repeat-edly (toggle option).1-3.3Visualize CurvesWhen importing complex models, many curves may be created and visualized in IGG™, makingthe graphics unclear. It is possible to visualize only specific curves on the screen, hiding all others,in the following way:18.Select the Geometry/View/Curves option. A curve chooser appears with the name of allthe curves. All the names are highlighted since all the curves are visible19.Select the C-spline in the chooser (click-left on it) and press Apply. Only the C-splinecurve now appears in the view20.Select the first polyline in the chooser (click-left on it) while holding the <Ctrl> key. Thepolyline is highlighted in the chooser, together with the C-spline. Press Apply to visualizeboth curves21.Select the first and last curves in the chooser while holding the <Shift> key. All the curvesare now selected. Press Apply to visualize them all22.Close the chooserAfter selecting the curves by using the Geometry/Select/Curves menu, the selected curves can befurther investigated in the following way:Geometry Creation Curves23.Select the Geometry/View/Curve Orientation menu. The default orientation of theselected curves is shown. This orientation is important for other geometry modelling andgrid generation operations. These orientations can be hidden by selecting the menu onceagain (toggle option)24.Select the Geometry/View/Control Points menu. The control points of the selected curvesappear now. This options acts as a toggle (display on-off) on all selected curves25.Select the Geometry/Select/Control Points menu. A control point must be selected. Whenmoving the mouse near a control point, the point becomes highlighted. Click-left on a con-trol point to display the point coordinates in the information area26.Click-right to quit the option27.Select the Geometry/Distance menu (). A prompt appears to select two points betweenwhich the distance will be measured and displayed:•Press <c> to disable the attraction to curves (this can be verified by moving the cursor near the start point of the C-spline. Normally, there is no attraction to the curve. Otherwise, press<c> a second time)•Move and attract the cursor over the start point of the C-spline. If there is no attraction, press <a>. Click-left on curve to select the start point•When moving the mouse, the distance between the selected point and the cursor is indi-cated. Move the mouse over the last point of the C-spline. The cursor is attracted to thepoint and the distance is indicating d=1•Click-left to fix the distance on the screenThe above steps can be repeated to measure the distance between otherpoints.•Click-right to quit the option.1-3.4Modify CurvesExisting selected curves can be modified within IGG™ in the following way:28.Select the Geometry/Modify Curve/Add Control Point option to add control points onselected curve by click-left on itCurves Geometry Creation29.Select the Geometry/Modify Curve/Remove Control Point option to remove a controlpoint on selected curve by click-left on control point30.Select the Geometry/Modify Curve/Modify Point (On surface) option to move an exist-ing control point on selected curve (on surface) by click-left to select the point and click-left after moving the control point31.Select the Geometry/Modify Curve/Set Name... option to impose a userdefined name tothe selected entity (one curve should be selected)32.Select the Geometry/Modify Curve/Divide option to split the selected curve at a userde-fined location by click-left on it (one curve should be selected)33.Select the Geometry/Modify Curve/Reverse option to reverse the curve orientation plot-ted when selecting Geometry/View/Curve Orientation menu1-3.5Edit/Copy CurvesExisting selected curves can be moved or copied within IGG™ in the following way:34.Select the Geometry/Select/Curves menu to select all the curves (highlighted in yellow)35.Select the Geometry/Edit/Copy menu to copy all the selected entities with a translation,rotation, scaling or mirror operation36.Type <new> <Enter> to impose a userdefined prefix to the geometrical entities that will becreated37.Type <t> <Enter> to select a copy with a translation38.Type <1 0 0> <Enter> to impose the translation vectortranslation (1 0 0)The menus Geometry/Edit/Translate, Rotate, Scale or Mirror allow to moveand not to copy the selected geometry.1-3.6Export CurvesIt is possible to save during the work the curves created in the previous steps. Only the curvesselected are saved into a file:39.Select the Geometry/Select/Curves menu to select all the curves (highlighted in yellow)Geometry Creation Surfaces40.Select File/Export/Geometry Selection... menu. A file chooser is opened to specify the name of a file ".dat" (with corresponding Parasolid ™file "X_T"). This file can be readback using the File/Import/IGG Data... menu ().1-4Surfaces 1-4.1Create Surfaces In this section simple surface creation is described, starting from a set of curves. A new session will be opened to clear all previous drawings.41.Select File/New - yes to close the current project and open a new, empty, project.Opening a new project closes the current project without automatic saving.42.Define a lofted surface:•Select File/Import/IGG Data and choose the file "geometry_curves.dat" in the\DOC\_Tutorials\IGG\Tutorial_1 directory of the installation cdrom. Three curves are readand stored in the geometry repository•Select the curves using Geometry/Select/Curves (<s >) in the order indicated on the figure•Verify that the curves are well oriented by using the Geometry/View/Curve Orientationmenu otherwise you need to reverse the curves by using the Geometry/Modify Curve/Reverse menu in order to impose the same orientation to all the curves•Select the Geometry/Surface/Lofted menu in the Quick Access Pad . A NURBS surface,interpolating the curves is now created. Notice that two new curves, representing surfaceboundaries, are created. These curves automatically appear in the curve chooser (Geome-try/View/Curves ) when it is opened123Boundary curves automatically created1234Surfaces Geometry Creation43.Define a coons patch:A Coons surface is a surface interpolating 4 boundary curves using a bilinearinterpolation. To avoid overlapping with the lofted surface, the selected curveswill be copied and translated.•Select the four boundary curves (<s>) of the lofted surface, in the order indicated in the above figure•Select the Geometry/Edit/Copy menu in the Quick Access Pad. IGG™ interrogates whether the duplicated curves must be translated, rotated, scaled, mirror or not. To avoid overlappingwith the existing curves and surface, a translation will be performed•Type <new> <Enter> to impose a userdefined prefix to the geometrical entities that will be created•Type <t> <Enter> to select a copy with a translation•Type <1 1 1> <Enter> to impose the translation vectorThe selected curves are duplicated and the new curves are automaticallyselected (the other curves are now unselected)•Select the Quick Access Pad/Geometry/Surface/Coons menu. A new surface is created which interpolates the four selected curvesIt can be noticed that 4 additional curves have been created. These are curves following the parametricdirections of the surface and are used to provide a better visualization of the surface.44.Define a surface of revolution:A surface of revolution will be created by rotating a newly created curve aroundthe Y axis.•First create a C-spline (Geometry/Curve/CSpline) between the points (-0.5,-2,0.1), (-0.5,0,0.2) and (-0.5,2,0.1). These points were selected so that the surface of revolution thatwill be created intersects the lofted surface•Make this curve the only selected curve (Geometry/Select/Curves)•Select Geometry/Surface/Revolution in the Quick Access Pad to create a surface of revolu-tion by rotating this new curve around a line parallel to the Y axis. The rotation origin is takenso that the surface of revolution intersects the lofted surface•Type <0 1 0> <Enter> to select the rotation axis direction•Type <-0.5 0 -1> <Enter> to select the rotation axis origin•Type <300> <Enter> to select the angle of rotationGeometry Creation SurfacesAs it may be noticed, the curve used for the rotation constitutes the first boundary of the surface.Three other boundary curves are automatically created to delimitate the surface.1-4.2Select SurfacesThe surface selection operation is used to activate one or more surfaces for subsequent operations in geometry modelling (i.e surface-surface intersection) or grid generation (i.e. face grid mapping).When a surface is selected its boundary curves appear highlighted in red or yellow.45.Select the Geometry/Select/Surfaces menu to initiate surface(s) selectionThe <Ctrl-s> shortcut can also be used to activate the same option, withoutaccessing the menu.46.Press <a > to unselect all the surfaces (toggle option), which become unhighlighted (bound-ary curves are unhighlighted)47.Move the mouse over the lofted surface. The surface becomes highlighted in blue.48.Click-left to select the surface. The boundary curves remain now permanently in red or yel-low49.Repeat above steps to select the surface of revolution50.Click-right to quit the selectionSelection and deselection of all the visible surfaces can be done by pressing<a> repeatedly (toggle option).1-4.3Visualize SurfacesSurfaces stored in IGG ™ are by default visualized by displaying their boundaries. As soon as the boundary curves of a surface are visible, the surface is considered visible. The following step describes how to hide surfaces, hence hide their boundaries.51.Select the Geometry/View/Surfaces option. A surface chooser appears with the name ofall the surfaces in the geometry repository. All surfaces in the chooser are highlighted sincethey are all visible in the graphics area52.Select the lofted surface (click-left on it) in the chooser and press Apply . The lofted surface appears alone in the graphics area with all the previously created curvesboundary curvesSurfaces Geometry Creation53.Select the surface of revolution (click-left on it) in the chooser while holding the <Ctrl> key.The surface of revolution is highlighted in the chooser together with the lofted surface. PressApply to visualize both surfaces. Notice that the surface of revolution is now unselected inthe graphics area (highlighted in blue)54.Select the first and last surfaces (click-left on them) in the chooser while holding the <Shift>key. All surfaces are highlighted in the chooser. Press Apply to visualize them all in thegraphics area55.Close the chooser1-4.4Modify SurfacesWhen manipulating parametric surfaces, it is possible to create curves in the parametric directions ofthe surfaces. These curves can be used to better visualize the surfaces or for other geometry and gridmodelling operations.56.After selecting a surface, select the Geometry/Modify Surface/Representation menu.IGG™ requests the number of curves to be created in the u and v direction of each selectedsurface:•Type <15 15> <Enter> to plot 15 curves in both parametric directions of the selected surfaces•Repeat the previous step and specify 5 curves in each direction57.Select the Geometry/Modify Surface/Add uv Curves menu. Then a point must be selectedon the selected surfaces:•Move the mouse inside the limits of the selected surfaces. Two orthogonal curves appear at the mouse position. The attraction feature can be enabled, if needed•Click-left to add the two curves in the geometry repositoryThe curves created in the above steps are deleted when the surface is deleted,except if they are used by other entities.1-4.5Edit/Copy SurfacesExisting selected surfaces can be moved or copied within IGG™ as presented on the curves in section1-3.5.1-4.6Export SurfacesIt is possible to save during the work the curves and surfaces created in the previous steps. Only thecurves and surfaces selected are saved into a file:58.Select the Geometry/Select/Curves and Surfaces menu to select respectively the curves(highlighted in yellow) and the surfaces (highlighted in red or yellow)59.Select the File/Export/Geometry Selection... menu. A file chooser is opened to specify thename of a file ".dat" (with corresponding Parasolid™ file "X_T"). This file can be read backusing the File/Import/IGG Data... menu ()Geometry Creation SurfacesTUTORIAL 2:2D Airfoil MeshGeneration2-1Introduction2-1.1IntroductionThe resolution of computational fluid dynamics (CFD) problems involves three main steps:•spatial discretization of the flow equations,•flow computation,•visualization of the results.To answer these questions, NUMECA has developed a F low IN tegrated E nvironment for internaland Turbomachinery assimilations. Called FINE™/Turbo, the environment integrates the followingtools:•IGG™ is an I nteractive G eometry modeler and G rid generator software, based on structured multi-block techniques,•AutoGrid™ is a three-dimensional Automated Grid generation software, dedicated to turboma-chinery applications. Similarly to IGG™, it is based on structured multi-block techniques,•Euranus is a state-of-the-art multi-block flow solver, able to simulate Euler and Navier-Stokes equations in the laminar, transitional and turbulent regimes,•CFView™ is a highly interactive flow visualization and post-treatment software,•FINE™ Graphical User Interface is a user-friendly environment that includes the different soft-wares. It integrates the concept of projects and allows the user to achieve complete simulations,going from the grid generation to the flow visualization, without the need of file manipulation.A C-type block grid around an airfoil is proposed to explain the basic features of the major topol-ogy and grid generation modules within IGG™.The tutorial shows the successive steps that must be followed to generate a 2D mesh and to definethe boundary conditions required before starting a solver:•Set up a 2D project,•Import/Create geometry curves needed for meshing,•Define the topology before meshing,。

Tutorial 使用指南

Tutorial 使用指南

使用指南(Tutorial)修订版序从首次接触这个软件到现在,有一段时间了。

那时由于急着使用,因此对一些认为不太重要的地方没有进行整理。

后来才发现,其实每一部分都是很有用的。

此修订,一个是将LineSim(Tutorial)与后加的Crosstalk(Tutorial)的目录统一起来,再有就是原文基础上增加了多板仿真(Tutorial)一节。

同样,对于那一时期我整理的BoardSim 、LineSim使用手册,也有同样的一个没有对一些章节进行翻译整理问题(当初认为不太重要)。

而实际上使用时,有一些东西是非常重要的,同时也顺便进行了翻译。

此外,通过使用,对该软件有了更多一些理解,显然以前只从字面翻译的东西不太好理解,等我有时间将它们重新整理后,再提供给初学的朋友。

对在学习中给予我大量无私帮助的Aming、pandajohn、lzd 等网友表示忠心的感谢。

P o q i0552002-8-202002-8-20目录使用指南(TUTORIAL ) 1 第一章 LINESIM4 1.1 在L INE S IM 里时钟信号仿真的教学演示 4 第二章 时钟网络的EMC 分析 7 2.1 对是中网络进行EMC 分析7 第三章 LINESIM'S 的干扰、差分信号以及强制约束特性 8 3.1 “受害者”和 “入侵者” 8 3.2如何定线间耦合。

8 3.3 运行仿真观察交出干扰现象9 3.4 增加线间距离减少交叉干扰(从8 MILS 到 12 MILS ) 93.5 减少绝缘层介电常数减少交叉干扰 93.6 使用差分线的例子(关于差分阻抗) 93.7仿真差分线 10第四章 BOARDSIM114.1 快速分析整板的信号完整性和EMC 问题 11 4.2 检查报告文件 11 4.3 对于时钟网络详细的仿真 11 4.4 运行详细仿真步骤: 11 4.5 时钟网络CLK 的完整性仿真 12 第五章 关于集成电路的MODELS 145.1 模型M ODELS 以及如何利用T ERMINATOR W IZARD 自动创建终接负载的方法 14 5.2 修改U3的模型设置(在EASY.MOD 库里CMOS,5V,FAST ) 14 5.3 选择模型(管脚道管脚)C HOOSING M ODELS I NTERACTIVELY (交互), P IN -BY -P IN 14 5.4 搜寻模型(F INDING M ODELS (THE "M ODEL F INDER "S PREADSHEET ) 15 5.5 例子:一个没有终接的网络 15 第六章 BOARDSIM 的干扰仿真 186.1 B OARD S IM 干扰仿真如何工作 186.3仿真的例子:在一个时钟网络上预测干扰 18 6.3.1加载本例的例题“DEMO2.HYP” 18 6.3.2A UTOMATICALLY F INDING "A GGRESSOR"N ETS 18 6.3.3为仿真设置IC模型 19 6.3.4查看在耦合区域里干扰实在什么地方产生的 19 6.3.5驱动IC压摆率影响干扰和攻击网络 20 6.3.6电气门限对比几何门限 20 6.3.7用交互式仿真"CLK2"网络 20 6.4快速仿真:对整个PCB板作出干扰强度报告 20 6.5运行详细的批模式干扰仿真 21第七章关于多板仿真237.1多板仿真例题,检查交叉在两块板子上网络的信号质量 23 7.2浏览在多板向导中查看建立多板项目的方法 24 7.3仿真一个网络A024 7.4用EBD模型仿真24HyperLynxHyperLynx是高速仿真工具,包括信号完整性(signal-integrity)、交叉干扰(crosstalk)、电磁屏蔽仿真(EMC)。

tutorial 中文手册2.2

tutorial 中文手册2.2

感 谢本tutorial Manual 2.2翻译文档在许多网友的关心和支持下,得以翻译成功,在此对他们表示热烈地感谢:蝈蝈、fiona、kailinziv、Yan、杀毒软件、tiny0o0、timothy、prolee等关心TG手册翻译的热心朋友。

关于本文档内容说明:由于本手册由不同网友翻译,可能对某些概念有不同的理解,翻译可能不大一样,但决不影响理解,欢迎大家探讨。

再一次对各位网友的努力和汗水表示感谢!如有什么问题可以联系我:Mail:tiny0o0@注:本文档内容版权归X Y Z scientific company 所有,谢绝任何意图商用。

I、TrueGrid介绍True Grid是一套优秀的、功能强大的通用网格生成前处理软件。

它可以方便快速生成优化的、高质量的、多块结构的六面体网格模型。

作为一套简单易用,交互式、批处理前处理器,True Grid支持三十多款当今主流的分析软件。

True Grid是基于多块体结构(multiple-block-structured)的网格划分工具,尽管这个指南手册开始会提供一些介绍信息,新手还是强烈要求阅读用户手册(True Grid® User’s Manual)的前2章,用户指南和参考手册。

True Grid是几何和网格形成过程是分开进行。

曲面和曲线形成的方式有以下几种:内部产生,从CAD/CAE系统导出IGES格式导入TG,或用vpsd命令导入多边曲面。

块网格(block mesh)初始化然后通过各种变换与几何模型匹配形成最后的有限元模型。

True Grid网格划分过程:运用block命令初始化块网格;块网格部分会被删掉以使拓扑结构与划分目标对应;块网格部分通过移动,曲线定位,曲面投影等方法进行变换;网格插值、光滑和Zoning(控制边界节点分布)等技术可以用来形成更好的网格;块网格之间独立形成,然后通过块边界面(BB)和普通节点合并命令(指定容差范围内合并)将各块网格合并成完整的有限元模型。

stable diffusion训练代码

stable diffusion训练代码

stable diffusion训练代码Stable Diffusion训练代码Stable Diffusion是一种新型的深度学习模型,它可以在训练过程中自适应地调整学习率,从而提高模型的稳定性和收敛速度。

在实际应用中,Stable Diffusion已经被广泛应用于图像分类、目标检测、自然语言处理等领域。

为了方便大家使用Stable Diffusion模型,我们提供了一份训练代码,以下是具体的使用方法:1. 环境准备在使用Stable Diffusion训练代码之前,需要先安装以下依赖库:- PyTorch- torchvision- numpy- matplotlib2. 数据准备Stable Diffusion训练代码支持的数据集包括MNIST、CIFAR-10、CIFAR-100等。

在使用训练代码之前,需要先下载并准备好相应的数据集。

3. 模型训练在训练模型之前,需要先设置一些超参数,包括学习率、批大小、训练轮数等。

这些超参数可以在代码中进行设置。

训练代码的主要流程如下:- 加载数据集- 定义模型- 定义损失函数和优化器- 训练模型- 保存模型4. 模型测试在训练完成后,可以使用测试集对模型进行测试,计算模型的准确率和损失值。

测试代码的主要流程如下:- 加载测试集- 加载模型- 对测试集进行预测- 计算准确率和损失值5. 结果展示最后,可以使用matplotlib库将模型的训练曲线和测试结果进行可视化展示,以便更直观地了解模型的性能。

总结Stable Diffusion是一种非常有前途的深度学习模型,它可以在训练过程中自适应地调整学习率,从而提高模型的稳定性和收敛速度。

通过使用我们提供的训练代码,您可以轻松地训练和测试Stable Diffusion 模型,并对其性能进行评估和展示。

我们相信,Stable Diffusion将会在未来的深度学习领域中发挥越来越重要的作用。

深度学习训练数据集的预处理流程

深度学习训练数据集的预处理流程

深度学习训练数据集的预处理流程In the process of training deep learning models, the preprocessing of the training dataset plays a crucial role in ensuring the performance and accuracy of the model. 数据预处理是深度学习模型训练过程中的重要环节,它对于保证模型的性能和精度至关重要。

First and foremost, data cleaning is an essential step in the preprocessing process. 首先,数据清洗是预处理过程中的重要步骤。

This involves identifying and handling missing values, dealing with outliers, and removing any irrelevant or redundant data that may negatively impact the model's performance. 这包括识别和处理缺失值,处理异常值,以及删除可能对模型性能产生负面影响的不相关或冗余数据。

Furthermore, data normalization is another crucial aspect of preprocessing the training dataset. 此外,数据归一化是预处理训练数据集的另一个关键方面。

Normalizing the data involves scaling the features to a standard range, which helps the model to converge faster and be less sensitive to the scale of input features. 数据归一化涉及将特征值缩放到一个标准范围,这有助于模型更快地收敛,并且对输入特征的尺度不太敏感。

训练集 测试集 验证集英文

训练集 测试集 验证集英文

训练集测试集验证集英文The Importance of Training, Testing, and ValidationSets in Machine Learning.In the field of machine learning, the division of data into training, testing, and validation sets is crucial for ensuring the effective development and evaluation of models. Each set serves a distinct purpose in the machine learning workflow, and their integration is essential for achieving accurate and reliable results.Training Set.The training set is used to teach the machine learning model how to perform a specific task. It contains a subsetof labeled data, which the model uses to learn the underlying patterns and relationships between inputs and outputs. The model's parameters are adjusted based on the training data to minimize a predefined loss function, which measures the difference between the model's predictions andthe actual labels.During the training phase, the model's goal is to fit the training data as well as possible. However, it'scrucial to avoid overfitting, where the model performs poorly on new, unseen data because it has learned the noise or irrelevant details in the training set. To mitigate this issue, techniques such as regularization and dropout are often employed.Validation Set.The validation set serves as a middle ground between the training set and the testing set. It's used to evaluate the model's performance during the training process, allowing for adjustments to be made without corrupting the test set's integrity. The validation set helps in hyperparameter tuning, model selection, and early stopping to prevent overfitting.By monitoring the model's performance on the validation set, practitioners can assess how well it generalizes tounseen data. If the model's performance on the validation set stops improving, it's a signal to stop training to prevent overfitting. The validation set also allows for comparisons between different models or algorithms, enabling practitioners to choose the best-performing one.Testing Set.The testing set is used to assess the final performance of the trained model on unseen data. It's crucial to evaluate the model's performance on data it hasn't encountered during training or validation to ensure its generalization capabilities. The testing set should be completely separate from the training and validation sets and only used once at the end of the machine learning workflow.By comparing the model's predictions on the testing set to the actual labels, practitioners can calculate evaluation metrics such as accuracy, precision, recall, and F1 score. These metrics provide a quantitative measure of the model's performance and allow for comparisons withother models or benchmarks.Conclusion.The division of data into training, testing, and validation sets is fundamental to the success of machine learning projects. The training set teaches the model, the validation set helps in tuning and evaluating the model during training, and the testing set provides an unbiased assessment of the model's final performance. By leveraging these sets effectively, practitioners can develop accurate, robust, and reliable machine learning models that generalize well to new, unseen data.。

stable-diffusion 训练日志

stable-diffusion 训练日志

稳定扩散训练日志一、背景介绍稳定扩散是一种新型的训练模式,它结合了稳定性训练和扩散训练的优点,旨在提高运动员的稳定性和灵活性。

在这种训练模式下,运动员需要在保持稳定的情况下进行各种动作,以提高身体的协调性和控制力。

稳定扩散训练日志是记录运动员在训练过程中的表现和进步情况的重要工具。

二、日志的重要性1. 监控训练效果稳定扩散训练日志可以帮助教练和运动员监控训练效果。

通过记录每次训练的内容、时长和感受,可以及时发现问题并调整训练计划,确保训练效果最大化。

2. 追踪进步情况日志还可以用来追踪运动员的进步情况。

通过对比不同阶段的训练日志,可以清晰地看到运动员的成长和改进,这不仅可以增强运动员的信心,还可以帮助教练更好地调整训练计划。

3. 查找问题和改进空间如果运动员在训练过程中遇到问题,可以通过查看日志找到原因。

日志记录的详细情况可以帮助教练和运动员找到问题所在,及时采取措施加以改进。

三、日志内容和记录要点稳定扩散训练日志应包括以下内容和记录要点:1. 训练日期和时间2. 训练地点3. 训练内容和目标4. 训练时长5. 训练感受和反馈6. 进食和休息情况7. 体重和身体状况记录8. 运动员个人信息和过往训练历史四、日志记录的方法和技巧1. 确定记录方式可以选择纸质记录或电子记录,根据个人习惯和喜好来确定记录方式。

电子记录可以利用手机App或软件来进行,方便查阅和管理。

2. 规范记录格式建立规范的记录格式,包括必填和选填项目,确保记录的完整性和准确性。

3. 及时记录尽量在训练结束后的第一时间进行记录,以免遗忘训练的细节和感受。

4. 定期总结和分析定期对训练日志进行总结和分析,找出问题和改进空间,并根据需要调整训练计划。

五、日志的管理和保密1. 保管安全无论是纸质记录还是电子记录,都需要妥善保管,以防丢失或泄露。

2. 保密原则稳定扩散训练日志属于个人隐私,应遵守保密原则,不得随意泄露给他人,尤其是竞争对手。

stable diffusion 训练代码

stable diffusion 训练代码

稳定性是机器学习领域中一个非常重要的问题,而稳定性的训练代码则更是机器学习工程师需要熟练掌握的技能之一。

稳定性的训练代码能够有效地避免模型训练过程中出现的过拟合、梯度爆炸、梯度消失等问题,从而提高模型的泛化能力和训练效果。

在这篇文章中,我们将介绍稳定性训练代码的一些常见技巧和方法,希望能够给读者带来一些启发和帮助。

一、数据预处理在训练模型之前,数据预处理是非常关键的一步。

合理的数据预处理能够有效地减少噪声的影响,提高模型的稳定性。

常见的数据预处理方法包括归一化、标准化、特征缩放等。

1. 归一化归一化是将数据缩放到一个特定的范围,如(0,1)或(-1,1)。

这样做可以使不同特征的取值范围一致,避免某些特征对模型训练过程的主导作用。

2. 标准化标准化是将数据按其特征的均值和标准差进行缩放,使得每个特征的取值服从均值为0、方差为1的正态分布。

这样做可以减少特征间的差异,提高模型的稳定性。

3. 特征缩放对于某些特征的取值范围较大或较小,可以采用特征缩放的方法对其进行缩放,以减少特征值的差异性。

二、模型选择选择合适的模型对于提高稳定性非常重要。

不同的模型适用于不同的数据和问题,合适的模型能够更好地拟合数据,提高模型的泛化能力。

1. 深度学习模型深度学习模型如CNN、RNN等在图像、语音、自然语言处理等领域取得了很好的效果,但是深度学习模型的训练过程也相对复杂,容易出现过拟合、梯度消失等问题。

在训练深度学习模型时,需要采用一些稳定性训练代码的方法,如批标准化、dropout、学习率衰减等。

2. 传统机器学习模型传统的机器学习模型如SVM、决策树、随机森林等在一些问题上也具有很好的效果,而且训练过程相对简单稳定。

选择合适的传统机器学习模型对于提高稳定性也是非常重要的。

三、训练技巧在实际的模型训练过程中,一些训练技巧也能够有效地提高模型的稳定性。

1. 批标准化批标准化是一种常用的提高深度学习模型稳定性的方法,它能够通过规范化输入数据,使得每一层的输入数据分布相对稳定,从而提高模型的训练速度和泛化能力。

stable diffusion embedding 训练

stable diffusion embedding 训练

Stable Diffusion Embedding(稳定扩散嵌入)是一种用于学习数据分布的无监督机器学习方法。

以下是一个基本的训练步骤概述:1. 数据准备:收集和预处理你的数据集。

这可能包括清洗、标准化、编码等步骤,确保数据适合用于训练模型。

2. 初始化模型参数:初始化稳定扩散嵌入模型的参数。

这些参数可能包括扩散系数、嵌入维度、优化器设置等。

3. 定义损失函数:稳定扩散嵌入通常使用马尔可夫链蒙特卡洛(MCMC)方法进行训练,损失函数通常是基于数据点之间的相似性或距离的度量,如余弦相似度或欧氏距离。

4. 训练循环:对于每个训练迭代:从数据集中采样一个批次的数据点。

使用当前的模型参数对这些数据点进行嵌入。

计算嵌入之间的相似性或距离,并使用这些信息计算损失函数的值。

根据损失函数的梯度更新模型参数。

这通常涉及到通过反向传播算法计算梯度,并使用优化器(如SGD、Adam等)来更新参数。

5. 模型评估与调整:在训练过程中,定期评估模型在验证集上的性能。

这可能包括计算嵌入的质量、聚类效果或者其他相关的指标。

根据评估结果调整模型参数、学习率或其他训练超参数,以优化模型性能。

6. 模型测试与应用:训练完成后,在测试集上评估模型的性能,确保其泛化能力。

将训练好的稳定扩散嵌入模型应用于实际问题,如数据可视化、聚类分析、推荐系统等。

请注意,具体的实现细节可能会根据你选择的稳定扩散嵌入算法和编程库(如TensorFlow、PyT orch等)有所不同。

在实际操作中,你可能需要参考相关的研究论文或开源代码库来获取更详细的指导。

  1. 1、下载文档前请自行甄别文档内容的完整性,平台不提供额外的编辑、内容补充、找答案等附加服务。
  2. 2、"仅部分预览"的文档,不可在线预览部分如存在完整性等问题,可反馈申请退款(可完整预览的文档不适用该条件!)。
  3. 3、如文档侵犯您的权益,请联系客服反馈,我们会尽快为您处理(人工客服工作时间:9:00-18:30)。

Figure 1 Figure 2
3
II. Import Geometry
1. 2. 3. 4. 5. 6. 7. Click BSE (See Figure 3). Select Preparation. Click IMPORT. Select file location: …/Tutorial1_Hanger Select file: hanger.igs Click Ok to import part geometry. Click Exit to dismiss the BSE Preparation dialog box.
Figure 3 Figure 4
Figure 5
4
III. Rename the part
1. 2. 3. 4. 5. 6. Click Parts (See Figure 6). Select Edit. Double click on the input box for Name to highlight part name C002V000. Type in the new name HANGER. Click Modify. Click OK to exit Edit Part dialog box.
18
IX. Outer Smooth
Figure 36
Figure 34
Figure 35
Figure 37
DYNAFORM 5.7.1 Training Tutorial
Blank Unfolding and Nesting
Descriptions for blank unfolding calculation and nesting process by Coat Hanger part.
I. Create and save database
11
VI. Blank Size Estimate
Figure 17
Figure 19 Figure 20 Figure 18
12
VI. Blankห้องสมุดไป่ตู้Size Estimate
Figure 23
Figure 21
Figure 22
13
VI. Blank Size Estimate
Figure 24
15
VII. Rectangle Fitting
Figure 26 Figure 29
Figure 27
Figure 28
Figure 30
16
VIII. Generate Blank Mesh
1. 2. 3. 4. 5. 6. 7. 8. Toggle on Elements and node at the lower right corner. Click BSE Select Development Select Blank Generator Select the blank outline by unfolding BSE. Enter Element Size: 2.00 (mm) (See Figure 32). Click OK Click Yes to accept the generated blank mesh (See Figure 33).
Figure 25
14
VII. Rectangle Fitting
1. Click to show the top view. 2. Click BSE Development. 3. Select BLANK FITTING (See Figure 27). 4. Toggle on Manual option (See Figure 28). 5. Click Select Line. 6. Select blank outline. 7. Click OK to exit Select Line dialog box. 8. Click Apply to create the fitting rectangle for blank outline fitting (See Figure 30). 9. Click Close to exit Blank Fitting dialog box. 10.Click to save the database.
6
IV. Auto-Mesh the surfaces
Figure 8 Figure 10
Figure 9
7
IV. Auto-Mesh the surfaces
Figure 11
8
V. Check and Repair Mesh
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Toggle off Surface (See Figure 12). Select Model Check/Repair (See Figure 13). Click Boundary Display icon(line one, column two)(See Figure 14). Toggle off Elements and Node. Click (free rotation) to rotate the model. Click (clear highlight). Click to display the model in isometric view. Click Auto Normal icon (line one, column one). Select CURSOR PICK PART. Move the cursor to select an element on the model (See Figure 15). Select No to reverse the normal (See Figure 16). Click Exit to dismiss the dialog box. Click OK to exit Model Check/Repair dialog box.
9
V. Check and Repair Mesh
Figure 12
Figure 15
Figure 13
Figure 14 Figure 16
10
VI. Blank Size Estimate
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Select BSE PreparationBlank Size Estimate (See Figure 17). Click Null to define material in the displayed blank size estimate window (See Figure 18). Click Material Library in the displayed material window (See Figure 19). Select CQ (mild) as the material (See Figure 20). Click OK to exit MATERIAL TYPE 36 dialog box (See Figure 21). Click OK to exit Material dialog box (See Figure 22). Enter the blank thickness = 1.20 (mm) (See Figure 23). Click Apply to run BSE (See Figure 24). Click Exit to dismiss BSE Preparation dialog box. Click to open the Show Part dialog box. Select HANGER and click OK to toggle off the part hanger, with OUT Line displayed only. Click to display blank outline in TOP view (See Figure 25). Click to save the database.
1. 2. 3. 4. Start up Dynaform 5.7.1. Click File menu and select Save As … (See Figure 1). Enter BSE_(user name)_(date).df as the file name. Click Save to save the database.
Figure 31
Figure 32
Figure 33
17
IX. Outer Smooth
1. 2. 3. 4. 5. 6. Select BSE DevelopmentOuter Smooth (See Figure 34). Select the newly generated blank mesh layer blnk0000. Select Roller (See Figure 35). Type in the Roll Radius: 300.00 (mm) (default). Click Create Boundary to generate new boundary line (See Figure 36). Click Fill Boundary to fill the gap between the original part boundary and the new boundary line (See Figure 37). 7. Select BSE Development Expand (See Figure 38). 8. Select BLK0000 and BLN_SMH as the operating objects for new blank (See Figure 39). 9. Type in Extension = 5.0 (mm),which indicates that the blank outline offsets 5mm to the outer side. 10.Click Boundary Expand to generate the offset blank outline (See Figure 40). 11.Click Fill Boundary to fill the offset gap (See Figure 41). 12.Click No in the displayed dialog box. 13.Click Exit to dismiss the Outer Smooth dialog box. 14.Click Close to exit the BSE Development dialog box.
相关文档
最新文档