You are not logged in!

F.A.Q
Log in

Register
  
 
  
(nobody)
...and 46 guests

Last 5 registered
Oplandisks
nothingstar
N_loop
yipe
foxtrotromeo

Browse members...
  
 
Members 8025
Messages 2613740
Today 18
Topics 127519
  
 
Messageboard index
hey dadonk
 

offline EpicMegatrax from Greatest Hits on 2021-09-28 05:48 [#02612327]
Points: 25264 Status: Regular



LAZY_TITLE

And that's the challenge with the various tactics that
have been used to make implementation smaller: They don't
reduce training costs enough. For example, one allows for
training a large network but penalizes complexity during
training. Another involves training a large network and then
"prunes" away unimportant connections. Yet another finds as
efficient an architecture as possible by optimizing across
many models—something called neural-architecture search.
While each of these techniques can offer significant
benefits for implementation, the effects on training are
muted—certainly not enough to address the concerns we see
in our data. And in many cases they make the training costs
higher.



 

offline EpicMegatrax from Greatest Hits on 2021-09-28 05:49 [#02612328]
Points: 25264 Status: Regular



One up-and-coming technique that could reduce training
costs goes by the name meta-learning. The idea is that the
system learns on a variety of data and then can be applied
in many areas. For example, rather than building separate
systems to recognize dogs in images, cats in images, and
cars in images, a single system could be trained on all of
them and used multiple times.


weasels. they need weasels


 

offline EpicMegatrax from Greatest Hits on 2021-09-28 05:50 [#02612329]
Points: 25264 Status: Regular



also crabs. but that's a bit of a tangent


 

offline ijonspeches from 109P/Swift-Tuttle on 2021-09-28 07:47 [#02612330]
Points: 7845 Status: Regular | Show recordbag



we need to create better algorithms by setting out the
pattern recognition to always compute the same task with a
known result, but have them shuffle their algorithms. who
reaches the goal fastest will be the new base to find even
better algorithms. there has to be absolutely NO cheating
though!
i guess there hasnt been found a way to do this properly and
maybe pattern recognition is still useless for that matter.


 

offline EpicMegatrax from Greatest Hits on 2021-09-28 08:00 [#02612333]
Points: 25264 Status: Regular | Followup to ijonspeches: #02612330



stimulated analhealing

but, no, the article covers this. and says why it doesn't
work. and how deep learning cheats that. and.. nevermind


 

offline ijonspeches from 109P/Swift-Tuttle on 2021-09-28 08:25 [#02612334]
Points: 7845 Status: Regular | Show recordbag



oh well, was worth a try...
guess we have to wait for an actual ai and hope it doesnt
get hooked up watching tiktok all day instead of doing our
calculations.


 

offline EpicMegatrax from Greatest Hits on 2021-09-28 19:40 [#02612339]
Points: 25264 Status: Regular | Followup to ijonspeches: #02612334



bedtime story


 

offline EpicMegatrax from Greatest Hits on 2021-09-29 09:05 [#02612342]
Points: 25264 Status: Regular



hay ddn k


 

offline EpicMegatrax from Greatest Hits on 2021-09-29 09:12 [#02612343]
Points: 25264 Status: Regular



The concept of mental imagery is also of interest. A
mental image is defined as a “mental experience which
occurs in the absence of stimulation, but which resembles
the experience that occurs when a stimulus is actually
present” [60]. In the sighted, mental images tend to be
strongly visual in nature. It is known, for example, that
“sighted subjects tend to visualize objects they examine
by touch” [60]. There is, however, “compelling evidence
that haptic mental imagery exists in congenitally blind
people” [60]. Haptic mental images, however, appear to be
different in nature than visual images. It seems, for
example, that the early blind “have a greater ability to
perceive vividly both the front and back of a palpated
object at the same time” [60].



 

offline EpicMegatrax from Greatest Hits on 2021-09-29 09:14 [#02612344]
Points: 25264 Status: Regular



because they have to do that to grab a flower vase on both
sides, duh.


 

offline Roger Wilco from Mo's Beans on 2021-09-29 17:58 [#02612345]
Points: 1999 Status: Lurker



Come on Dadonck. This insufferable prick wants to talk to
you.


 

offline DADONCK from here on 2021-09-29 19:34 [#02612346]
Points: 3532 Status: Regular



Sorry, havin a Premiere this Friday, working all day


 

offline mermaidman on 2021-09-29 20:20 [#02612347]
Points: 8308 Status: Regular



you think just cause epic made a thread for dadonck he needs
him to answer back? no! of course not! i mean if you had
something to say to someone you'd make the thread write your
post and wait right? not epic!


 

offline EpicMegatrax from Greatest Hits on 2021-09-30 06:42 [#02612351]
Points: 25264 Status: Regular | Followup to DADONCK: #02612346



7.0? CS4?

(good luck)


 

offline mermaidman on 2021-09-30 09:48 [#02612355]
Points: 8308 Status: Regular | Followup to EpicMegatrax: #02612351



very bad joke


 

offline EpicMegatrax from Greatest Hits on 2021-10-01 20:45 [#02612371]
Points: 25264 Status: Regular | Followup to mermaidman: #02612355



even worse joke


 

offline mermaidman on 2021-10-01 21:16 [#02612372]
Points: 8308 Status: Regular | Followup to EpicMegatrax: #02612371



which joke?


 

offline EpicMegatrax from Greatest Hits on 2021-10-02 00:41 [#02612378]
Points: 25264 Status: Regular | Followup to mermaidman: #02612372



i reply to your 'very bad joke' with 'even worse joke' and
to answer your question, even worse joke is not joking


 

offline mermaidman on 2021-10-02 09:32 [#02612383]
Points: 8308 Status: Regular



very bad joke wasn’t joking either


 

offline mermaidman on 2021-10-02 10:06 [#02612384]
Points: 8308 Status: Regular



also it's not only a very bad joke it's a TERRIBLE FUCKING
JOKE


 

offline EpicMegatrax from Greatest Hits on 2021-10-03 05:15 [#02612395]
Points: 25264 Status: Regular



even more terrible joke


 

offline EpicMegatrax from Greatest Hits on 2021-10-03 05:15 [#02612396]
Points: 25264 Status: Regular



hey dadonck,
friend and i are going to the isabella stewart gardener
museum tomorrow.


 

offline DADONCK from here on 2021-10-03 12:49 [#02612401]
Points: 3532 Status: Regular | Followup to EpicMegatrax: #02612342



i passed that campus many times. drunk, on my way home


 

offline DADONCK from here on 2021-10-03 12:54 [#02612402]
Points: 3532 Status: Regular



grimes must have been living there at that time


 

offline DADONCK from here on 2021-10-03 13:00 [#02612403]
Points: 3532 Status: Regular | Followup to EpicMegatrax: #02612396



nice!


 

offline DADONCK from here on 2021-10-03 16:20 [#02612407]
Points: 3532 Status: Regular



little challenge: find the rembrandt selfie without looking
online


 

offline EpicMegatrax from Greatest Hits on 2021-10-04 04:22 [#02612415]
Points: 25264 Status: Regular



it proved impossibru, unfortunately. tickets all sold out.
museum of science annoying 15m airlocking system because
kids kids kids. i would have settled for bryon cavestun's
cranz but ah well

we had to settle with a dive bar


 

offline EpicMegatrax from Greatest Hits on 2021-10-04 10:30 [#02612420]
Points: 25264 Status: Regular



at least we didn't run into this guy


 


Messageboard index