« first day (1901 days earlier)      last day (2580 days later) » 

1:16 AM
@Szabolcs You are right,I have realized Wolfram Community is more suitable for this post,but another frustrated thing is I cannot access my account to that site.I will post to it to there by some method.
 
1:40 AM
@xslittlegrass The style transfers in Mathematica cost my one day and one night,and get a bad result.Do you have another method can implement this?
I mean the example in ref/NetTrain|applications|computer vision|style transfer cost my time.And I just use my CPU
@xslittlegrass Like this
 
 
2 hours later…
4:03 AM
@yode I haven't played with the style net yet. Do you mean you get different results running exactly the same code in the documentation?
 
4:25 AM
@xslittlegrass I have tried this example,then I get such result.:-)
 
 
1 hour later…
5:54 AM
@yode I think the bad results you get is expected since that paper is specifically written to address some of the problems in the original style-net.
 
@xslittlegrass Yes,I think so,just bad for my that request,and I don't know how to adjust that code for my demend.
And that code cost so much time that I have no impetus to research.
Of course,I'd like to see you master-hand in neural-networks can find more beautiful solution. :)
 
 
3 hours later…
9:10 AM
What does this mean in practice?
 
@P.Fonseca Why wouldn't it mean what it says? Blockchain related functions in a release due this year.
 
 
3 hours later…
11:54 AM
What is the most reliable way to measure memory of consumption of a function? I tried mem = MemoryInUse[]; maxMem = MaxMemoryUsed[f]; maxMem - mem but it doesn't seem to be correct. It can return negative values.
 
I can even generalize it further: it means what it means (this could come directly out of Matrix)
 
@P.Fonseca Maybe I misunderstood the question. I don't really know exactly what their product does either, I just read it as "there will be functions related to their product in a coming release of Mathematica". That is all I know.
 
 
3 hours later…
3:31 PM
I keep getting prompted that my question contains code that isn't properly formatted and that I should indent with 4 spaces... But IT IS properly formatted!! Is anyone familiar with this error?
I believe I found the issue... I was defining variables using "Where" and the system though it was a loop function.
 
4:35 PM
@C.E. sounds like a good question for main.
 
 
1 hour later…
5:45 PM
Does anybody see a way to create a function getAxes3D that takes a 3D graphics and returns a new 3D graphics that contains lines and text primitives that exactly reconstruct the box, axes and ticks of the input graphics?
This comes down to the problem that I don't know any way to extract ticks with AbsoluteOptions.
 
Yikes. DateValue[{2017, 1, 1}, "Week"] gives 52.
 
@halirutan The name of the function FullGraphics but it doesn't work.
 
@MichaelHale And that is certainly correct.. in some reality.
 
@halirutan Hopefully it didn't mess up the post I just made about Trump's weekly average likes per Tweet too much.
 
@MichaelHale hehe... it corrects Trump's likes down :)
@C.E. PrintDefinitions suggests that this only works for Graphics not for Graphics3D although it doesn't even work there :)
 
6:01 PM
@halirutan I had plans for writing my own Graphics to SVG exporter once and I couldn't figure out how to convert axes and ticks to graphics primitives, it's a shame. Mr. Wizard has done some spelunking which is a step towards recreating ticks but that's the best I know of.
 
@C.E. I'm wrapping my head around a decent 3D export to any suitable format for a long time..
I'm off for some time. See you later.
 
@halirutan ok, is that exported with your exporter? To what format? Looks pretty good.
 
6:40 PM
I guess this is what I wanted for the date stuff anyway. Normal allows GroupBy to handle the week spans properly.
Normal@DateObject[{2017, 1, 1}, "Week"]
Pretty happy with the code now. data // Select[#fav > 0 &] //
GroupBy[#, Normal@DateObject[#@"created_at", "Week"] & -> (#fav &),
Mean] & // DateListPlot
Although SocialMediaData doesn't let you get tweets. Just a bunch of stuff about their Twitter follower network and stuff.
 

« first day (1901 days earlier)      last day (2580 days later) »