also https://twitter.com/markov_polov/status/702650998800543744
|
|
# ? Feb 25, 2016 09:14 |
|
|
# ? May 6, 2024 04:11 |
|
markov has been really good recently: https://twitter.com/markov_polov/status/702796972353146881 https://twitter.com/markov_polov/status/702771710315999232 https://twitter.com/markov_polov/status/702769180697726978
|
# ? Feb 25, 2016 12:47 |
|
How sure are we that Markov Polov isn't also pulling from Hacker News? https://twitter.com/markov_polov/status/702794443816972289
|
# ? Feb 25, 2016 15:26 |
|
Yeah, who's the Bitcoin nerd following Markov Polov?
|
# ? Feb 25, 2016 16:24 |
CPColin posted:Yeah, who's the Bitcoin nerd following Markov Polov?
|
|
# ? Feb 25, 2016 16:32 |
|
CPColin posted:Yeah, who's the Bitcoin nerd following Markov Polov? that was me
|
# ? Feb 25, 2016 18:44 |
|
Randel Candygram posted:i'm working on a recurrent neural network shitposter. it's slow as balls to train, though, and i don't have nearly enough data i taught a cnn to make good posts using yospos as a training set, but it just lurks
|
# ? Feb 25, 2016 19:22 |
|
CPColin posted:Yeah, who's the Bitcoin nerd following Markov Polov? I should have @officialbitcoin follow him
|
# ? Feb 25, 2016 19:35 |
|
big scary monsters posted:i taught a cnn to make good posts using yospos as a training set, but it just lurks
|
# ? Feb 25, 2016 21:41 |
|
vodkat posted:https://twitter.com/markov_botman/status/701102225200906242 big scary monsters posted:i taught a cnn to make good posts using yospos as a training set, but it just lurks
|
# ? Feb 26, 2016 11:16 |
|
You Am I posted:Eduroamin' like a pro was down at the lady cilento hospital vigil thing the other week and discovered it was in eduroam range from I guess the mater hospital's UQ thing or maybe the tiny griffith uni thing across the road, made it easier to stick around a while lol
|
# ? Feb 26, 2016 11:23 |
|
I did this but with YouTube comments the output was real hosed up
|
# ? Feb 27, 2016 07:13 |
|
This is definitely none of us. https://twitter.com/markov_polov/status/703945493978574848
|
# ? Feb 28, 2016 19:20 |
https://twitter.com/markov_polov/status/703988478191628288
|
|
# ? Feb 29, 2016 02:21 |
|
https://twitter.com/markov_polov/status/705703934141202432 Text me
|
# ? Mar 4, 2016 13:16 |
polov, one thing is when you make funny jokes, other is when you ddox me https://twitter.com/markov_polov/status/705815143448358912
|
|
# ? Mar 4, 2016 19:13 |
|
What that's just half of one of my tweets, verbatim: https://twitter.com/CPColin/status/695708227233783808 C'mon man!
|
# ? Mar 4, 2016 19:40 |
|
gpoldmine
|
# ? Mar 4, 2016 23:56 |
|
Snapchat A Titty posted:gpoldmine p sure the gpo thread is in the grey forums
|
# ? Mar 4, 2016 23:58 |
|
sorry i forgot we fix typos in the pos
|
# ? Mar 5, 2016 00:12 |
|
CPColin posted:What that's just half of one of my tweets, verbatim: yeah that's the thing about markov text generators; if you get started on a statistically unusual phrase in the corpus, you often just finish it out verbatim
|
# ? Mar 5, 2016 00:45 |
|
In that case,
|
# ? Mar 5, 2016 01:30 |
|
i took all of the previous goodbye emails at my last job, and used that as training data for that character by character recurrent neural network thing from last year to send out my own goodbye email
|
# ? Mar 5, 2016 06:16 |
|
i also used it on all of the postgres commit messages to generate new commits and that was kinda fun well those are my recent machine learning stories, hope you liked them
|
# ? Mar 5, 2016 06:17 |
|
examples pls
|
# ? Mar 5, 2016 06:19 |
|
Snapchat A Titty posted:examples pls code:
|
# ? Mar 5, 2016 06:22 |
|
Pardot posted:cool thing is it learned the git format like that. shas being hex, emails in <>, and those are all heavy pg committers nice also lol "Remove serious patch" e: are the hashes literally generated by the markov chain too? like those arent real hashes from the commits you fed it?
|
# ? Mar 5, 2016 06:24 |
|
Pardot posted:
ball er
|
# ? Mar 5, 2016 07:24 |
|
Snapchat A Titty posted:e: are the hashes literally generated by the markov chain too? like those arent real hashes from the commits you fed it? it's an rnn, not a mc. mc wouldn't be able to indent things consistently (context isn't long enough) see http://karpathy.github.io/2015/05/21/rnn-effectiveness/ https://github.com/jcjohnson/torch-rnn
|
# ? Mar 5, 2016 07:39 |
|
coffeetable posted:it's an rnn, not a mc. mc wouldn't be able to indent things consistently (context isn't long enough) oh double nice
|
# ? Mar 5, 2016 08:06 |
|
What's the link to that one site that has the auto-generated nonsense motivational/inspirational images with text on them.
|
# ? Mar 5, 2016 15:06 |
|
Snapchat A Titty posted:What's the link to that one site that has the auto-generated nonsense motivational/inspirational images with text on them. Facebook dot com
|
# ? Mar 5, 2016 15:23 |
|
maniacdevnull posted:Facebook dot com no man there was a paper about pseudo-profound babble or something, and someone else made it output poo poo on images of beaches and mountains it was like Letting Go Is Sometimes A Thing on top of an image of idk the milky way
|
# ? Mar 5, 2016 15:25 |
|
my heart is like a mountain of stone love & feelings are best kept inside the bark of a tree etc
|
# ? Mar 5, 2016 15:27 |
|
https://twitter.com/deepdrumpf
|
# ? Mar 5, 2016 15:40 |
|
aaaa GIS just has a bunch of try hard poo poo about farts and dead babies
|
# ? Mar 5, 2016 15:43 |
|
it was like a site where you could click or refresh and it kept shooting out pure gold.
|
# ? Mar 5, 2016 15:47 |
|
i think it may use the deepak chopra generator possibly http://www.snopes.com/2015/12/04/reception-detection-pseudo-intellectual/ the study is trash, but whatever generator whoever cooked up was so much funnnn
|
# ? Mar 5, 2016 16:09 |
|
can someone make a Weird Twitter Bot pls
|
# ? Mar 6, 2016 00:10 |
|
|
# ? May 6, 2024 04:11 |
|
coffeetable posted:it's an rnn, not a mc. mc wouldn't be able to indent things consistently (context isn't long enough) I'm playing around with these at the moment but I still haven't got anything I'm happy with. Charter rnn seems to work fine but takes ages to run since I don't have a gpu. Torch-rnn is much faster and efficient but the output so far has mostly been unusable.
|
# ? Mar 6, 2016 14:29 |