|
return0 posted:why is clustering some data so hard, it’s 2019 ugh have you tried adding more bias to your training set
|
# ? Dec 3, 2019 01:04 |
|
|
# ? May 25, 2024 03:15 |
|
i found out about tensorflow arduonios and now I want to join the club
|
# ? Dec 3, 2019 03:36 |
|
why the hell has no one ported numpy/scipy cobol libraries i mean i know the practical answer to this question, but no one got a wild one up their rear end to at least make a go of it?
|
# ? Dec 4, 2019 01:24 |
|
cobol programmers get paid too well doing cobol to bother doing python for free. i am also not quite sure why you want it.
|
# ? Dec 4, 2019 10:12 |
|
if you want to do math in a dumb dead language use fortran
|
# ? Dec 4, 2019 12:09 |
|
power botton posted:if you want to do math in a dumb dead language use fortran which is exactly why there are python-FORTRAN bindings
|
# ? Dec 4, 2019 14:47 |
|
Some of our computational physicists still use FORTRAN And these are people like my age Living FORTRAN physics libraries
|
# ? Dec 4, 2019 14:48 |
|
uh, and, numpy/scipy are a bunch of fortran with some python coating, i am not sure where you were expecting fortran to go?
|
# ? Dec 4, 2019 14:58 |
|
i swear to god alice...
|
# ? Dec 4, 2019 15:41 |
|
is fortran still faster than c for some types of numeric calculations for deeply technical yet tedious reasons?
|
# ? Dec 5, 2019 02:00 |
|
redleader posted:is fortran still faster than c for some types of numeric calculations for deeply technical yet tedious reasons? That's not changing any time soon. tl;dr is that Fortran doesn't allow programmers to do arbitrary things with memory and that means the compiler can make more assumptions about how code behaves and that makes optimization easier.
|
# ? Dec 5, 2019 02:30 |
|
if you're serious about high-performance numerics you should be writing CUDA anyway oh also, check this out: https://tvm.apache.org/2018/10/09/ml-in-tees a custom compiler for running ML models inside SGX enclaves. now the system that implacably denies your loan application could be running inside your very own computer! And you *still* won't be able to see how it works! animist fucked around with this message at 03:33 on Dec 5, 2019 |
# ? Dec 5, 2019 03:30 |
|
CUDA's great if what you're trying to do exhibits SIMD parallelism. If not, it's a non-starter.
|
# ? Dec 5, 2019 03:49 |
|
redleader posted:is fortran still faster than c for some types of numeric calculations for deeply technical yet tedious reasons? this is a weird comparison since fortran is in many ways a much more alive language than c, having been modernized to add e.g. proper modules, operator overloading (pretty big deal for numerics), forall loops, and if you dip a far in as fortran'03, object orientation. also, yeah, cuda is another super-important leg to all this stuff, but tons of kernels are extremely hard to parallelize and require a lot of very careful work to split even across a small number of well-interconnected general cpus. e.g. the vast zoo of bespoke matrix decompositions for extremely specific subcases.
|
# ? Dec 5, 2019 10:55 |
|
is fortran the one u have to dress as a cyber gladiator for?
|
# ? Dec 5, 2019 14:12 |
|
Fanged Lawn Wormy posted:is fortran the one u have to dress as a cyber gladiator for? dont forget the huge boner
|
# ? Dec 5, 2019 15:36 |
|
that’s forth
|
# ? Dec 5, 2019 16:11 |
|
you can always start your own traditions though all these cowards using woodcuts of animals or portraits instead of commissioning extremely horny fantasy artists smh
|
# ? Dec 5, 2019 17:07 |
|
Share Bear posted:you can always start your own traditions though
|
# ? Dec 5, 2019 19:51 |
|
ultrafilter posted:That's not changing any time soon. the main thing here is restrictions on pointer aliasing and you can do that in c with "restrict" and get the same performance. i think it's more that much fortran code is straight textbook math, "formula translation" as they say, and it's battle tested and works fast so why rewrite it in c and risk loving it up
|
# ? Dec 5, 2019 20:49 |
|
can confirm that debugging high-performance scientific code is hell (because i'm doing that right now)
|
# ? Dec 5, 2019 21:58 |
|
it is easy to get in a mindset where one imagines that the code would be easy to work with if it was just structured just so (or written in ones preferred language), but some code just expresses sufficiently complicated concepts that it will always be difficult. godspeed etc. vv
|
# ? Dec 5, 2019 22:04 |
|
e: wrong thread
animist fucked around with this message at 23:09 on Dec 5, 2019 |
# ? Dec 5, 2019 23:06 |
|
https://twitter.com/ryxcommar/status/1202577620644352001
|
# ? Dec 6, 2019 00:22 |
|
also if you take the initials of each step in the circle it spells IDIOM which is one letter off from what the creator of that slide is
|
# ? Dec 6, 2019 01:28 |
|
|
# ? Dec 6, 2019 05:17 |
|
akadajet posted:that’s forth this explains why i couldn't find it lol the more you look the better it gets
|
# ? Dec 6, 2019 05:41 |
|
where are his nipples
|
# ? Dec 6, 2019 19:54 |
|
https://twitter.com/JanelleCShane/status/1202968242286784512
|
# ? Dec 7, 2019 01:51 |
|
60 years of AI for a slightly more advanced Eliza
|
# ? Dec 7, 2019 02:05 |
|
Vomik posted:60 years of AI for a slightly more advanced Eliza yeah for once computers are being used for something nontoxic. its nice
|
# ? Dec 7, 2019 04:33 |
|
https://twitter.com/nxthompson/status/1203840778608529408
|
# ? Dec 10, 2019 01:09 |
|
Fanged Lawn Wormy posted:this explains why i couldn't find it the most amazing thing about this to mr is the sheer quantity of underboob
|
# ? Dec 13, 2019 22:49 |
|
looking at that cover again, i wonder what kind of facial hair e. floegel had
|
# ? Dec 13, 2019 22:54 |
|
i guess in the space future there is no gravity so the bra is actually just there to keep boobs *down* so they dont float away or whatever so that leads to a reverse cleavage paradigm
|
# ? Dec 13, 2019 22:57 |
|
I can’t tell if things are less horny now or different horny
|
# ? Dec 13, 2019 22:58 |
|
that's surely the horniest stack programming in particular has ever been
|
# ? Dec 13, 2019 23:01 |
|
Cybernetic Vermin posted:that's surely the horniest stack programming in particular has ever been horny & stacked programmers???
|
# ? Dec 13, 2019 23:15 |
|
lancemantis posted:I can’t tell if things are less horny now or different horny different horny, 100%
|
# ? Dec 13, 2019 23:28 |
|
|
# ? May 25, 2024 03:15 |
|
dumb serious question. why do we use nonlinear functions instead of linear ones if they're all dumb poo poo like ReLU which seems like it barely works. I guess second question: is there any theoretical basis for why ReLU works at all?
|
# ? Dec 15, 2019 07:55 |