@MichaelE2 Hello. :)

@MichaelE2 I note you have 112,067 rep.If I downvote your answer,you have that permission to know that is my behavior?

@yode No one can see who downvoted. Not even moderators (I think).

@Szabolcs Correct, mods can find voting patterns, but cannot determine who voted for what.

@yode Hi. :) Sometimes my computer stays logged in, even though it and I have gone to sleep. :) Breakfast time now. Even though it's not yet light.

@Szabolcs Thanks.(also J.M. and Michael E2)

I have a group of points, over which a mesh should form one or more closed surfaces. How to express the constraint on existence of edges that describes only valid closed surface solutions (where every 3-cycle over present edges is a valid triangle on the surface mesh, and all these triangles form one or more valid closed surfaces)?

Clearly there tend to be multiple valid solutions, choosing one that makes most sense is left as a minimization problem over valid solutions.

I suspect this ILP problem can't be solved in reasonable time for real-world datasets... :(

@kirma This mesh you speak of can be self-intersecting?

I'm trying to make a IHM with mathematica to play music on a raspberry. p=StartProcess[{"omxplayer","thefile.mp3"}] launches the player (named "omxplayer"). I have two problems : 1) KillProcess[p] doesn't work 2) The keyboard has no effect on oxmplayer (the characters are sent to mathematica instead of omxplayer)

@J.M. Oh well, technically yes, just that it's closed in the sense that every edge is part of exactly two 3-cycles.

And every point is part of at least three edges.

@kirma Definitely sounds like a combinatorially hard problem, without further restrictions...

:36359082 Again, we do not discuss things like that in public. :) It is the user's own business as to why s/he might leave.

@J.M. Mostly I see it exploding on amount of variables, and being integer linear programming, probably nasty in complexity.

@kirma "exploding on amount of variables" - that's what I meant by "combinatorially hard". :|

Sorry,I have not realized this also is a public occasion

:I

Just imagining how to group the triangles is making me wince.

I wonder if thinking this as a cube of boolean values listing all possible triangles might help any. Problem there is that one would very quickly have billion boolean variables...

I think this can actually be reasoned with billion-bit variables with certain ease, the problem is if Mma is capable of handling that gracefully.

@Jacccy For example:

Thread[{#[[All, 1, 1]], #[[All, All, 2]]}] &@Transpose[data]

I guess something like greedy (polygon or edge) spanning tree traversal is more or less the only way to approach my problem sensibly. Even then gigantic fan-out is probably preventive factor for actually finding optimal solutions...

But that would put certain amount of geometric intuition into the game, at least.

@C.E. thank you!!!

Anybody success to use NetModel in 11.1?I'm always in this status..

@yode it works for me

Sorry to hear that..I don't know what's problem in my this version..

VGG-16 is about 500M, maybe your connection is slow?

How long do you cost to download it?

It takes me only a second to load, but that's probability because it is cached on my system.

Does the LeNet trained on MNIST net model work for you?

@yode That's a much smaller network

I mean the first time when you success to use it. :)

@yode I don't remember, but I just tried "Inception V1 Trained on ImageNet Competition Data" which I haven't used before, and it takes about 10 seconds to load.

@yode Does NetModel["LeNet Trained on MNIST Data"] work for you?

@xslittlegrass I also don't know it should take how much time to load..

Does someone with 11.1 wants to test this dynamic Collatz visualizer? Be careful to save everything, because while in 10 it works fine, in 11.1 somehow one processor runs at 100% the whole time making the front end virtually freeze:

Import["http://halirutan.github.io/Mathematica-SE-Tools/decode.m"]["http://i.stack.imgur.com/EUGpP.png"]

@yode 如果这个小的模型都不能下载，估计是你Mathematica的问题吧。你的网速怎么样?

@halirutan That's work well here

成功了。。。

@xslittlegrass 这个多大来着？

@halirutan works for me as well

@halirutan That look beautiful!

我的网速是8M的

@yode 这个是1.5M

醉，1.5M花了十多分钟，难怪500M一天下不来

@yode 那可能是网速的问题吧

@halirutan I am interesting how do you encode you expression into Import[..][..]

@yode 如果你想用VGG-16,我可以把文件通过网盘发给你

@xslittlegrass 那太好了。。要不然我不知道要下多少天

@yode 好的我上传完了告诉你

多谢 :)

Ok,I have found your original code.Thanks.(ping halirutan)

@halirutan As you instruction,the result of Import["http://goo.gl/NaH6rM"] should same with Import["http://halirutan.github.io/Mathematica-SE-Tools/decode.m"]?But actually not

@yode pan.baidu.com/s/1jI5c6WY

1fj5

And how do I get your SEEncodeExpression from your "http://halirutan.github.io/Mathematica-SE-Tools/..."?If I use the same code,I will fail to get it.

@yode 链接：pan.baidu.com/s/1hsGBWMS 密码：wp6h

能看到吗？

@xslittlegrass 对了，请问得解压到哪个文件夹？

我没查到NetModel["LeNet Trained on MNIST Data"]的缓存跑哪去了

直接Import[".wlnet"]就可以

@xslittlegrass 奇怪，按你说的，我的NetModel["LeNet Trained on MNIST Data"]应该会在本地生成一个.wlnet，我用everytion居然没查到。。

像你正常缓存下来是在哪个文件夹来着？

@yode 不知道呢，我是直接Export导出的，我分享给你的你能用吗？

@halirutan it works fine on osx. What os are you running on?

爽,可以正常运行，激动得我要跳楼了，多谢多谢@xslittlegrass

不客气:)

I don't know why, but when I don't understand a language (Chinese), it looks like more fascinating. I hope that's not why I have chosen Mathematica.

@xslittlegrass @yode @rcollyer Thanks for testing guys. So again a Linux issue.. :(

@yode I'm not entirely sure what you want. If you have the SETools installed, you automatically have all these functions.

<< SETools`

?? SEEncodeExpression