ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
Maik has joined #mlpack
Maik has quit [Ping timeout: 272 seconds]
< jonpsy[m]>
Hello again, I missed GSoC but I'm really interested in the Reinforcement Learning idea of mlpack. I think some Nishant Kumar is working on a select part of it (Correct me if i'm wrong). I'd like to implement the rest. I'm an intermediate c++ programmer and have good experience in github. I've built mlpack from source and am reading the docs atm. Any clarity on the availability of this project/availability of the mentors
< jonpsy[m]>
would be appreciated :). Thanks!!
DHRUBAJYOTIBORAH has left #mlpack []
ImQ009 has joined #mlpack
kyrre_ has joined #mlpack
kyrre has quit [Ping timeout: 256 seconds]
kyrre_ is now known as kyrre
bisakh[m] has quit [Ping timeout: 256 seconds]
bisakh[m]1 has joined #mlpack
kunal12298Gitter has quit [Ping timeout: 256 seconds]
kunal12298Gitter has joined #mlpack
sy0814kGitter[m] has quit [Ping timeout: 256 seconds]
TaapasAgrawalGit has quit [Ping timeout: 256 seconds]
sy0814kGitter[m4 has joined #mlpack
TaapasAgrawalGit has joined #mlpack
walragatver[m] has quit [Ping timeout: 256 seconds]
walragatver[m]1 has joined #mlpack
MostafaNabiehGit has quit [Ping timeout: 256 seconds]
MostafaNabiehGit has joined #mlpack
Nakul[m] has quit [Ping timeout: 256 seconds]
Nakul[m]1 has joined #mlpack
< jonpsy[m]>
* Hello again, I missed GSoC but I'm really interested in the Reinforcement Learning idea of mlpack. I think some Nishant Kumar is working on a select part of it (Correct me if i'm wrong). I'd like to implement the rest. I'm an intermediate c++ programmer and have good experience in github. I've built mlpack from source and am reading the docs atm. Any clarity on the availability of this project/availability of the mentors
< jonpsy[m]>
would be appreciated :). Thanks!!
< jonpsy[m]>
Edit: I think the concerned persons are zoq favre49
favre49 has joined #mlpack
< favre49>
jonpsy[m]: Glad to know you're interested. Nishant will be implementing Rainbow DQN and Soft Actor Critic. You're welcome to implement other RL algos
TaapasAgrawalGi4 has joined #mlpack
EL-SHREIFGitter4 has joined #mlpack
< favre49>
jeffin143[m]: ping me if you ever feel like playing again :)
< jonpsy[m]>
Sure , I'll start by few first issues and then jump on to a concrete plan to execute this. How do I reach you?
< favre49>
Great! Just drop a message on the IRC, if I'm not around someone else could help you too
< jonpsy[m]>
Sure thing! :)
< jeffin143[m]>
favre49: that's sound like a challenge :)
< jeffin143[m]>
If you don't mind what's your rating ?? In blitz ?
< favre49>
I mostly play rapid, haven't played blitz i think
< favre49>
i'm 1140 or something rapid
< jeffin143[m]>
Let's have. Game :)
< jeffin143[m]>
10 min is ok ?? favre49
< favre49>
sure, i'm in the middle of a game, give me 5 min
< jeffin143[m]>
Sure :)
< jeffin143[m]>
<favre49 "I mostly play rapid, haven't pla"> https://play.chess.com/wq95 : the game is active for 1 hour
< favre49>
great' i'll let you know when my current game is finished
< favre49>
alright, ready when you are
< jeffin143[m]>
Good game :)favre49
< favre49>
nice game, miscalculated that exchange
< jeffin143[m]>
I blundered Queen
< jeffin143[m]>
In first game
< jeffin143[m]>
Completely
favre49 has quit [Quit: Lost terminal]
< maddogx[m]>
<sreenik[m] "maddogx[m]: This is the communit"> sreenik[m]: IC.
< jonpsy[m]>
nishantkr18 you should know I'm a fan. Keep em coming!
< nishantkr18[m]>
@jonpsy:matrix.org: sure :D
< rcurtin>
nishantkr18[m]: thanks for posting, I enjoyed reading it. for the LayerTypes issue, maybe someone else has already said it, but this is why we have the "MoreTypes" types; it's a workaround to get more than 50 types :)
< rcurtin>
also, I am back now, but it will probably take me a while to catch up on everything :)
< rcurtin>
a little overwhelming, almost 800 emails (luckily most of those are just git commit emails, etc.)
< nishantkr18[m]>
@rcurtin:matrix.org: yeah, that's what we did to add more layer types. Thanks for reading :)
< HimanshuPathakGi>
Hey @zoq @saksham189 I tried using google colab to train RBFN model with mnist dataset it took around 4hrs with 20000 sample I was able to train but results are'nt that good I will try changing some configuration and try again https://colab.research.google.com/drive/1sGDtvTirgFWyEY3rHo7nmmkrXP3zIlOt#scrollTo=Nkicu5D7k-VK If you want to see what I am doing Please take a look.
< abernauer[m]>
Has anyone in the community had any success getting labeled data from companies for personal projects?
< saksham189Gitter>
@himanshupathak21061998 I don't think we can access the page. Can you check the permission once?
< jeffin143[m]>
abernauer: I got one during an interview , but I couldn't download :)
< jeffin143[m]>
@walragatver:matrix.org: @brim
< jeffin143[m]>
Up for meet ?
< RyanBirminghamGi>
Hello!
< jeffin143[m]>
Hi Ryan Birmingham (Gitter)
< RyanBirminghamGi>
How are you doing today?
< abernauer[m]>
Yeah was just wondering cause I am starting a personal project where the goal is to use deep learning to grade baseball cards and get close to expert grades.
< jeffin143[m]>
Ryan Birmingham (Gitter): good
< jeffin143[m]>
No difference between weekends and weekdays 😂
< RyanBirminghamGi>
I try to be a little lazier on weekends to punctuate the weeks.
< jeffin143[m]>
abernauer: that's nice
< RyanBirminghamGi>
The problem is I'm also sometimes lazy on weekdays, so, I dunno
< jeffin143[m]>
Ryan Birmingham (Gitter): haha thats a big confusion
< walragatver[m]1>
Hi birm and jeffin143 :)
< jeffin143[m]>
Hi walragatver
< walragatver[m]1>
Yeah sometimes taking a break is good pysically and mentally :)
< walragatver[m]1>
physically*
< walragatver[m]1>
Ok so let's start a meet.
< walragatver[m]1>
jeffin anything you would like to discuss first?
< jeffin143[m]>
About testing : walragatver
< jeffin143[m]>
Should we use boost or something else
< jeffin143[m]>
There is an open issue about removing testing
< jeffin143[m]>
I think so @shrit:matrix.org suggested a good alternative
< jeffin143[m]>
Let's use that
< jeffin143[m]>
Or at least try it
< walragatver[m]1>
Okay can you just point me where shrit has suggested the alternative. I will read it out.
< walragatver[m]1>
I think we are using catch in ensmallen If I am not wrong
< walragatver[m]1>
<jeffin143[m] "walragatver: also I saw your com"> I have a given a review today. Have you seen it.
< jeffin143[m]>
walragatver: yes I have resolved them
< jeffin143[m]>
I made changes and pushed it
< walragatver[m]1>
Okay I will review it.
< RyanBirminghamGi>
I think walragatver covered it, but I recommend adding a basic CI check
< RyanBirminghamGi>
it can just be cmake and make for now
< walragatver[m]1>
<RyanBirminghamGi "I think walragatver covered it, "> birm: Yeah I was also going to ask about it today.
< jeffin143[m]>
Ryan Birmingham (Gitter): yes I will add it :)
< jeffin143[m]>
Azure builds
< jeffin143[m]>
Or GitHub actions as of now*
< walragatver[m]1>
Okay
< walragatver[m]1>
Let us know what you would be using for the CI check.
< jeffin143[m]>
I will first go with azure
< jeffin143[m]>
If I am able to do it we will stick to it
< jeffin143[m]>
Since we would following same thing everywhere
< jeffin143[m]>
But if I am not able to get azure done then I will try GitHub actions
< walragatver[m]1>
Great you can refer to the .ci/ folder in mlpack/models and other repositories for azure builds.
< jeffin143[m]>
Yes
< jeffin143[m]>
I did have a look today
< jeffin143[m]>
And panned out something :)
< walragatver[m]1>
Okay birm: I reviewed the pr today.
< walragatver[m]1>
It looks like style checks are not working on comments.
< RyanBirminghamGi>
I'll look through it too soonish
< walragatver[m]1>
I am not sure but I think they should work.
< walragatver[m]1>
<RyanBirminghamGi "I'll look through it too soonish"> Great
< walragatver[m]1>
Okay so that's it from my side to discuss.
< walragatver[m]1>
birm: Anything you would like to discuss?
< walragatver[m]1>
jeffin143: I will give the second review soon.
< jeffin143[m]>
walragatver ok
< birm[m]>
I'm good for now!
< walragatver[m]1>
birm: jeffin143: Okay let's conclude the meet.
< walragatver[m]1>
Thank you for your time. Have a nice week ahead :)
< jeffin143[m]>
walragatver: bye :)
petris has quit [Quit: Bye bye.]
petris has joined #mlpack
ArabellaCurly-Q has joined #mlpack
ArabellaCurly-Q has quit [Client Quit]
< saksham189Gitter>
@himanshupathak21061998 something must be wrong. Because this is performing worse than a random algorithm would.
< HimanshuPathakGi>
> @himanshupathak21061998 something must be wrong. Because this is performing worse than a random algorithm would.
< HimanshuPathakGi>
Yes I am investigating over this but right now it is not clear what the problem is ?? Also, I think I should close that k-means issue because it is working with other csv. So, there is some problem with mnist_train.csv What do you think?
< saksham189Gitter>
yes please do
< HimanshuPathakGi>
Ok :)
ImQ009 has quit [Quit: Leaving]
< saksham189Gitter>
@himanshupathak21061998 can you test out different approaches and see which one works best?
< HimanshuPathakGi>
> @himanshupathak21061998 can you test out different approaches and see which one works best?