gibberish master 10000

  • 17 Replies
  • 3789 Views
*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
gibberish master 10000
« on: September 06, 2019, 09:43:34 am »
Lockster made me do it,    and im just showing off.

https://www.youtube.com/watch?v=fW2VjLWWmQU&feature=youtu.be

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #1 on: September 06, 2019, 10:23:54 am »
I am supreme goddess of all, master of none
Emergent          https://openai.com/blog/

*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #2 on: September 06, 2019, 11:10:57 am »
here something more edumacational -  this actually highlights the forks - other than that its a bit of a cheating device.   O0

https://www.youtube.com/watch?v=8bv6nxm35rc&feature=youtu.be

Thanks Lock for giving me the big 1 gig text store.


« Last Edit: September 06, 2019, 11:44:49 am by goaty »

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #3 on: September 06, 2019, 12:50:09 pm »
Don't forget to tell them it runs on GPU power!
Emergent          https://openai.com/blog/

*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #4 on: September 06, 2019, 02:24:14 pm »
heres my code to dispatch the video card.


<code>
 INVOKE(dc, derezz0, 0, views, nullptr, nullptr, 0, output_uav, 1, 1, 1 );  //derezz the closest match score
 
 views[0]=query_srv[0];
 for(j=0;j<PAGES;j++){views[j+1]=biblepage_srv[j];} 
 INVOKE(dc, runmatch, PAGES+1, views, nullptr, nullptr, 0, output_uav,  (BIBLEPAGE_SIZE*4*SQRTPAGES)/32, (BIBLEPAGE_SIZE*SQRTPAGES)/32, 1 );  //HERES THE BIG GPU BASTARD CALL!    //THIS THING CALLS 2048x2048x16 THREADS AND STILL GOES RELATIVELY QUICKISH.  its the limit of my little card.

 views[0]=query_srv[0];views[1]=output_srv;
 INVOKE(dc, stepquery, 2, views, nullptr, nullptr, 0, query_uav[1], 1, 1, 1 );  //step the query along.
 </code>



*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #5 on: September 06, 2019, 03:37:13 pm »
Looks like ive gone embarrassed myself again,  banned off all seven seas off the internet.

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #6 on: September 07, 2019, 06:10:33 am »
The algorithm searches up to 250MB of text for an exact match on GPU. It comes out just as fast as CPUs in the end :P
Emergent          https://openai.com/blog/

*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #7 on: September 07, 2019, 03:01:35 pm »
But Lock,  you don't realize the beauty of the situation.

Its working acceleration structureless!   And that's going to mean something,  so back to work I go.

Next thing im adding is order invariance (which would work whether its linear search or not.)
Then it should yoda similar to sound,  itll start saying verbs like they are nouns.

and a synonym generator (its done simple, and words which have identical sorroundings become the same symbol,  and u do it as long as u want but it guarantees less sense each compounding.)
« Last Edit: September 07, 2019, 07:49:47 pm by goaty »

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #8 on: September 08, 2019, 01:15:03 am »
My synonyms are way better ;)

GPU has it all spread on many cores like random access but it's as slow as CPU.
Emergent          https://openai.com/blog/

*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #9 on: September 08, 2019, 05:24:44 pm »
The more you say words are substitutable with each other, the more jumps your system gets, and then the more attention span you can give it for so much data in the system.

But - how do you get it to communicate instead of just spitting out a random story?

*

Freddy

  • Administrator
  • **********************
  • Colossus
  • *
  • 6855
  • Mostly Harmless
Re: gibberish master 10000
« Reply #10 on: September 08, 2019, 06:21:20 pm »
But - how do you get it to communicate instead of just spitting out a random story?

I think that's why we're all here  :)

*

WriterOfMinds

  • Trusty Member
  • ********
  • Replicant
  • *
  • 606
    • WriterOfMinds Blog
Re: gibberish master 10000
« Reply #11 on: September 08, 2019, 09:29:32 pm »
So what's happening here exactly?  Are you just running Lock's generator on your computer?  Running Lock's generator with your own improvements?  Or is this a different generator entirely?

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #12 on: September 08, 2019, 09:37:41 pm »
Lock generators is wayyyyy better ;)

This is just the most basic simple search engine running on GPU.
Emergent          https://openai.com/blog/

*

goaty

  • Trusty Member
  • ********
  • Replicant
  • *
  • 552
Re: gibberish master 10000
« Reply #13 on: September 09, 2019, 01:01:04 pm »
So what's happening here exactly?  Are you just running Lock's generator on your computer?  Running Lock's generator with your own improvements?  Or is this a different generator entirely?

Lock gave me an idea,  just to have 60 meg or so just raw ascii text then every new letter it just brute scans it with all the gpu power of my gtx980 -  that way I don't need to store it in a binary tree, and I can actually get lossy matching as well.

Turns out its pretty wasteful,  and it needs the proper binary tree to go faster.

I have an idea of how to get it chatting,  but I need a source of proper chat text,  and I then I have an idea that if I search during the jumps,  I can make the robot try and direct the other speaker to say a word,  or change their mood,  if I have that in the jump information.

I think that if I had videos of people chatting to each other,  I could use computer vision with computer audio,  to get a useful set of data,  but I think brute searching it all is a bit of a waste of time, (it was a failure, when I said it was running quick, I only meant because it was actually linear searching/brute forcing.)  so ill get back to it later.

*

LOCKSUIT

  • Emerged from nothing
  • Trusty Member
  • *******************
  • Prometheus
  • *
  • 4659
  • First it wiggles, then it is rewarded.
    • Main Project Thread
Re: gibberish master 10000
« Reply #14 on: September 09, 2019, 01:13:35 pm »
I proved tree storage is worse, you waste RAM like no tomorrow. You can see my new little invention has extra low mem and extra fast search. Btw, for Next Word prediction, searching 1GB as 1Search is not long at all. That is not issue...

Actually let me think about binary trees while I sleep, I know of a way i think that stores them without the huge overhead they make.
Emergent          https://openai.com/blog/

 


OpenAI Speech-to-Speech Reasoning Demo
by ivan.moony (AI News )
Today at 01:31:53 pm
Say good-bye to GPUs...
by MikeB (AI News )
March 23, 2024, 09:23:52 am
Google Bard report
by ivan.moony (AI News )
February 14, 2024, 04:42:23 pm
Elon Musk's xAI Grok Chatbot
by MikeB (AI News )
December 11, 2023, 06:26:33 am
Nvidia Hype
by 8pla.net (AI News )
December 06, 2023, 10:04:52 pm
How will the OpenAI CEO being Fired affect ChatGPT?
by 8pla.net (AI News )
December 06, 2023, 09:54:25 pm
Independent AI sovereignties
by WriterOfMinds (AI News )
November 08, 2023, 04:51:21 am
LLaMA2 Meta's chatbot released
by 8pla.net (AI News )
October 18, 2023, 11:41:21 pm

Users Online

279 Guests, 0 Users

Most Online Today: 346. Most Online Ever: 2369 (November 21, 2020, 04:08:13 pm)

Articles