rcurtin_irc changed the topic of #mlpack to: mlpack: a scalable machine learning library (https://www.mlpack.org/) -- channel logs: https://libera.irclog.whitequark.org/mlpack -- NOTE: messages sent here might not be seen by bridged users on matrix, gitter, or slack
<Liu[m]> Hey, guys, I ran into a little problem when trying to remove boost parts in mlpack like issue <https://github.com/mlpack/mlpack/issues/3095|#3095> suggests. I chose to replace `boost::heap::priority_queue` in `mlpack/src/mlpack/core/tree/cosine_tree/cosine_tree.hpp`.... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/ccf7d844d1bf64844383ea26743a9f68756ad6b7)
halfy has quit [Quit: Bridge terminating on SIGTERM]
Cadair has quit [Quit: Bridge terminating on SIGTERM]
rcurtin[m] has quit [Quit: Bridge terminating on SIGTERM]
NabanitaDash[m] has quit [Quit: Bridge terminating on SIGTERM]
AnwaarKhalid[m] has quit [Quit: Bridge terminating on SIGTERM]
Sarrah[m] has quit [Quit: Bridge terminating on SIGTERM]
SlackIntegration has quit [Quit: Bridge terminating on SIGTERM]
heisenbuugGopiMT has quit [Quit: Bridge terminating on SIGTERM]
BenJoiceBenJoice has quit [Quit: Bridge terminating on SIGTERM]
zoq[m]1 has quit [Quit: Bridge terminating on SIGTERM]
RajatMishra[m] has quit [Quit: Bridge terminating on SIGTERM]
EshaanAgarwal[m4 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_U0 has quit [Quit: Bridge terminating on SIGTERM]
PKhurana09Pranav has quit [Quit: Bridge terminating on SIGTERM]
zoq[m] has quit [Quit: Bridge terminating on SIGTERM]
HimanshuPathak[m has quit [Quit: Bridge terminating on SIGTERM]
AnoushkaJha[m] has quit [Quit: Bridge terminating on SIGTERM]
say4n[m] has quit [Quit: Bridge terminating on SIGTERM]
jonpsy[m] has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_31 has quit [Quit: Bridge terminating on SIGTERM]
GopiMTatiraju[m] has quit [Quit: Bridge terminating on SIGTERM]
manav71ManavSang has quit [Quit: Bridge terminating on SIGTERM]
AshvithShetty[m] has quit [Quit: Bridge terminating on SIGTERM]
AkifEmreapolu[m] has quit [Quit: Bridge terminating on SIGTERM]
DarshKaushik[m] has quit [Quit: Bridge terminating on SIGTERM]
mohamedAdhamcmoh has quit [Quit: Bridge terminating on SIGTERM]
Aakash-kaushikAa has quit [Quit: Bridge terminating on SIGTERM]
TarekNasser[m] has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_U4 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_34 has quit [Quit: Bridge terminating on SIGTERM]
EshaanAgarwal[m] has quit [Quit: Bridge terminating on SIGTERM]
Liu[m] has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_13 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_U7 has quit [Quit: Bridge terminating on SIGTERM]
rcurtin_matrixor has quit [Quit: Bridge terminating on SIGTERM]
psydroid has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_28 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_25 has quit [Quit: Bridge terminating on SIGTERM]
kuries has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_10 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_37 has quit [Quit: Bridge terminating on SIGTERM]
shrit[m] has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_16 has quit [Quit: Bridge terminating on SIGTERM]
ShubhamAgrawal[4 has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_19 has quit [Quit: Bridge terminating on SIGTERM]
hitesh-anandhite has quit [Quit: Bridge terminating on SIGTERM]
Niketkumardheery has quit [Quit: Bridge terminating on SIGTERM]
VageeshaDattaGVa has quit [Quit: Bridge terminating on SIGTERM]
_slack_mlpack_22 has quit [Quit: Bridge terminating on SIGTERM]
hitesh-anandhite has joined #mlpack
Cadair has joined #mlpack
rcurtin[m] has joined #mlpack
halfy has joined #mlpack
psydroid has joined #mlpack
SlackIntegration has joined #mlpack
Sarrah[m] has joined #mlpack
TarekNasser[m] has joined #mlpack
jonpsy[m] has joined #mlpack
Niketkumardheery has joined #mlpack
zoq[m]1 has joined #mlpack
Liu[m] has joined #mlpack
AnwaarKhalid[m] has joined #mlpack
ShaikhMohammedAm has joined #mlpack
BenJoiceBenJoice has joined #mlpack
heisenbuugGopiMT has joined #mlpack
mohamedAdhamcmoh has joined #mlpack
VageeshaDattaGVa has joined #mlpack
GopiMTatiraju[m] has joined #mlpack
shrit[m] has joined #mlpack
Aakash-kaushikAa has joined #mlpack
PKhurana09Pranav has joined #mlpack
ShubhamAgrawal[m has joined #mlpack
EshaanAgarwal[m] has joined #mlpack
RajatMishra[m] has joined #mlpack
zoq[m] has joined #mlpack
_slack_mlpack_U0 has joined #mlpack
NabanitaDash[m] has joined #mlpack
manav71ManavSang has joined #mlpack
say4n[m] has joined #mlpack
DarshKaushik[m] has joined #mlpack
AshvithShetty[m] has joined #mlpack
HimanshuPathak[m has joined #mlpack
EshaanAgarwal[m4 has joined #mlpack
kuries has joined #mlpack
_slack_mlpack_U4 has joined #mlpack
TarekNasser[m]1 has joined #mlpack
AkifEmreapolu[m] has joined #mlpack
AnoushkaJha[m] has joined #mlpack
_slack_mlpack_U7 has joined #mlpack
082AAGIGE has joined #mlpack
082AAGIGE has quit [Client Quit]
078AAHU16 has joined #mlpack
074AALN9Y has joined #mlpack
076AAIJUF has joined #mlpack
_slack_mlpack_10 has joined #mlpack
_slack_mlpack_13 has joined #mlpack
_slack_mlpack_16 has joined #mlpack
_slack_mlpack_19 has joined #mlpack
<ShubhamAgrawal[m> > <@_slack_mlpack_U038MB7TMMG:matrix.org> Hey, guys, I ran into a little problem when trying to remove boost parts in mlpack like issue <https://github.com/mlpack/mlpack/issues/3095|#3095> suggests. I chose to replace `boost::heap::priority_queue` in `mlpack/src/mlpack/core/tree/cosine_tree/cosine_tree.hpp`.... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/3e6f97b53acea4f773da6c071aaecae479cb78bb)
<ShubhamAgrawal[m> > <@shubhamag:matrix.org> I think it will not hurt to implement this as a separate internal method in mlpack. But first try to run test with replaced implentation.
<ShubhamAgrawal[m> And can you link the reproducible examples you created
<ShubhamAgrawal[m> > Anyone with different thoughts?
halfy has quit [Quit: User was banned]
psydroid has quit [Quit: User was banned]
rcurtin[m] has quit [Quit: User was banned]
SlackIntegration has quit [Quit: User was banned]
hitesh-anandhite has quit [Quit: User was banned]
Cadair has quit [Quit: User was banned]
TarekNasser[m] has quit [Quit: User was banned]
Sarrah[m] has quit [Quit: User was banned]
zoq[m] has quit [Quit: User was banned]
Niketkumardheery has quit [Quit: User was banned]
jonpsy[m] has quit [Quit: User was banned]
shrit[m] has quit [Quit: User was banned]
_slack_mlpack_13 has quit [Quit: User was banned]
074AALN9Y has quit [Quit: User was banned]
ShubhamAgrawal[m has quit [Quit: User was banned]
zoq[m]1 has quit [Quit: User was banned]
mohamedAdhamcmoh has quit [Quit: User was banned]
AnwaarKhalid[m] has quit [Quit: User was banned]
NabanitaDash[m] has quit [Quit: User was banned]
RajatMishra[m] has quit [Quit: User was banned]
VageeshaDattaGVa has quit [Quit: User was banned]
_slack_mlpack_U4 has quit [Quit: User was banned]
Liu[m] has quit [Quit: User was banned]
PKhurana09Pranav has quit [Quit: User was banned]
GopiMTatiraju[m] has quit [Quit: User was banned]
EshaanAgarwal[m] has quit [Quit: User was banned]
heisenbuugGopiMT has quit [Quit: User was banned]
076AAIJUF has quit [Quit: User was banned]
say4n[m] has quit [Quit: User was banned]
DarshKaushik[m] has quit [Quit: User was banned]
_slack_mlpack_U7 has quit [Quit: User was banned]
078AAHU16 has quit [Quit: User was banned]
Aakash-kaushikAa has quit [Quit: User was banned]
HimanshuPathak[m has quit [Quit: User was banned]
_slack_mlpack_19 has quit [Quit: User was banned]
_slack_mlpack_10 has quit [Quit: User was banned]
kuries has quit [Quit: User was banned]
_slack_mlpack_U0 has quit [Quit: User was banned]
BenJoiceBenJoice has quit [Quit: User was banned]
EshaanAgarwal[m4 has quit [Quit: User was banned]
ShaikhMohammedAm has quit [Quit: User was banned]
_slack_mlpack_16 has quit [Quit: User was banned]
AnoushkaJha[m] has quit [Quit: User was banned]
manav71ManavSang has quit [Quit: User was banned]
AkifEmreapolu[m] has quit [Quit: User was banned]
AshvithShetty[m] has quit [Quit: User was banned]
TarekNasser[m]1 has quit [Quit: User was banned]
GopiMTatiraju[m] has joined #mlpack
<GopiMTatiraju[m]> zoq: rcurtin when we do `mat.get_dev_mem()` the memory which we get is that contiguous?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/ee5645d71887cc18d35989ab50ed8a06c6a8e5f5)
<GopiMTatiraju[m]> * zoq: rcurtin when we do `mat.get_dev_mem()` the memory which we get is that contiguous?
<GopiMTatiraju[m]> when I am doing
<GopiMTatiraju[m]> `cudaMemcpy(dest, source1, size1, cudaMemcpyDeviceToDevice); cudaMemcpy(dest+size1, source2, size2, cudaMemcpyDeviceToDevice);`
<GopiMTatiraju[m]> Although I am not getting any error but values seem random.
<GopiMTatiraju[m]> When I tried it with a basic c style 1D array the above worked.
<GopiMTatiraju[m]> * zoq: rcurtin when we do `mat.get_dev_mem()` the memory which we get is that contiguous?
<GopiMTatiraju[m]> Although I am not getting any error but values seem random.
<GopiMTatiraju[m]> when I am doing
<GopiMTatiraju[m]> ```cudaMemcpy(dest, source1, size1, cudaMemcpyDeviceToDevice); cudaMemcpy(dest+size1, source2, size2, cudaMemcpyDeviceToDevice);`
<GopiMTatiraju[m]> * zoq: rcurtin Is memory returned by `mat.get_dev_mem()` contiguous?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/9de7249fab8c40087e25ba777c4e9748dbacd971)
<GopiMTatiraju[m]> * zoq: rcurtin Is memory returned by `mat.get_dev_mem()` contiguous?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/20ebe67c6a5a35fee11b1d4d83d1d6dcfee60689)
EshaanAgarwal[m] has joined #mlpack
<EshaanAgarwal[m]> > <@shubhamag:matrix.org> I think it will not hurt to implement this as a separate internal method in mlpack. But first try to run test with replaced implentation.... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/5af888357bb1cbd8460a9dab827e9ffbd487e84b)
<EshaanAgarwal[m]> s/://, s///, s///
GopiMTatiraju[m] has quit [Quit: Bridge terminating on SIGTERM]
EshaanAgarwal[m] has quit [Quit: Bridge terminating on SIGTERM]
hitesh-anandhite has joined #mlpack
Cadair has joined #mlpack
rcurtin[m] has joined #mlpack
halfy1 has joined #mlpack
psydroid has joined #mlpack
SlackIntegration has joined #mlpack
TarekNasser[m] has joined #mlpack
Sarrah[m] has joined #mlpack
shrit[m] has joined #mlpack
jonpsy[m] has joined #mlpack
Niketkumardheery has joined #mlpack
_slack_mlpack_U0 has joined #mlpack
GopiMTatiraju[m] has joined #mlpack
ShaikhMohammedAm has joined #mlpack
Liu[m] has joined #mlpack
PKhurana09Pranav has joined #mlpack
EshaanAgarwal[m] has joined #mlpack
heisenbuugGopiMT has joined #mlpack
AnwaarKhalid[m] has joined #mlpack
mohamedAdhamcmoh has joined #mlpack
ShubhamAgrawal[m has joined #mlpack
zoq[m]1 has joined #mlpack
VageeshaDattaGVa has joined #mlpack
RajatMishra[m] has joined #mlpack
say4n[m] has joined #mlpack
NabanitaDash[m] has joined #mlpack
Aakash-kaushikAa has joined #mlpack
BenJoiceBenJoice has joined #mlpack
AshvithShetty[m] has joined #mlpack
zoq[m] has joined #mlpack
HimanshuPathak[m has joined #mlpack
AkifEmreapolu[m] has joined #mlpack
DarshKaushik[m] has joined #mlpack
TarekNasser[m]1 has joined #mlpack
manav71ManavSang has joined #mlpack
AnoushkaJha[m] has joined #mlpack
EshaanAgarwal[m4 has joined #mlpack
kuries has joined #mlpack
_slack_mlpack_U4 has joined #mlpack
_slack_mlpack_U7 has joined #mlpack
074AALPVB has joined #mlpack
030AAFSSJ has joined #mlpack
078AAHW1E has joined #mlpack
_slack_mlpack_10 has joined #mlpack
_slack_mlpack_13 has joined #mlpack
_slack_mlpack_16 has joined #mlpack
078AAHW1J has joined #mlpack
077AAIAM5 has joined #mlpack
<zoq[m]> > <@gopitatiraju:matrix.org> zoq: rcurtin Is memory returned by `mat.get_dev_mem()` contiguous?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/053ba6d825f877a73a26adacdfa800c1cbb26250)
<rcurtin[m]> yeah, all memory for `mat` objects is contiguous (whether it's OpenCL or CUDA memory)
<jonpsy[m]> hey guys, I have a generic question.
<jonpsy[m]> So, I have a complex logic where I take every combination of every length of a list and run it through an optimizer.
<jonpsy[m]> For ex: if my list is ['man', 'ban', 'can'], I'll run ['man'], ['can'], ['ban'], ['man', 'ban'], ['man','can'] ... ['man','ban','can']. aka this would grow exponentially with `len(list)`.
<jonpsy[m]> The catch is, I know the max len is 8, so at max it could grow to 2^8 combinations. I was wondering can I parallelize this since so that each thread takes 1 of the 256 tasks (in the worst case).
<rcurtin[m]> certainly you could parallelize it, but is it worth it? that will depend on how much time each individual optimization takes; if it's a trivial operation, the overhead of parallelization may not be worthwhile
<rcurtin[m]> you can use things like OpenMP `task`s; that may make the implementation easier
<jonpsy[m]> We're using python
<rcurtin[m]> (if you're going to do it in C++ with OpenMP, that is :))
<jonpsy[m]> rcurtin[m]: hehe
<rcurtin[m]> I don't know anything about Python parallelism
<jonpsy[m]> rcurtin[m]: I'd say yeah. The optimization process is **very** costly like about 50 seconds or so, if we run 2^8 * 50 it would take forever
<jonpsy[m]> rcurtin[m]: nw you can just tell me if its a valid thing to do. Since my knowledge about parallelism is very limited
<Liu[m]> hey Curtin, would u mind take a look at my question above? Thanks
<rcurtin[m]> Liu: I saw it, I'll do that when I have a chance. 👍️ like Eshaan Agarwal pointed out, I think this has been discussed before
<Liu[m]> Is the solution mentioned in StackOverflow acceptable in mlpack? If so, I’d love to have a try 🙂
<rcurtin[m]> I have no idea, I haven't had a chance to look at the issue and formulate a response---I can't do that for you right now, I will get to it when I get to it 👍️
<Liu[m]> Sure, I’ll look at some other simple issues and have a try. Thanks~
<GopiMTatiraju[m]> <zoq[m]> "> <@gopitatiraju:matrix.org> zoq..." <- dest and source1 are matrix.
<GopiMTatiraju[m]> And I used sizeof(eT)*size_t(n_elem)
<GopiMTatiraju[m]> When I am filling the `dest` all at once it's working, but when I am trying to fill it in parts(like I mentioned above) the values are incorrect....
<zoq[m]> <GopiMTatiraju[m]> "When I am filling the `dest` all..." <- https://gist.github.com/zoq/11101e2022d11527854a263061a3939b
<zoq[m]> zoq[m]: Returns the expected results.
<zoq[m]> zoq[m]: As I mentioned keep in mind the size you like to copy is in bytes.
<zoq[m]> zoq[m]: `cudaMemcpy(dest+size1, source2, size2, cudaMemcpyDeviceToDevice);` isn't correct because you used `size1` for two different things.
<zoq[m]> zoq[m]: To iterate the pointer and to specify the number of bytes to copy.
<GopiMTatiraju[m]> Oh okay...
<GopiMTatiraju[m]> I will make the changes accordingly and update...
<GopiMTatiraju[m]> <zoq[m]> "https://gist.github.com/zoq/1110..."; <- Hey zoq , I commented on the gist...
<TarekNasser[m]> Hey
<TarekNasser[m]> I am trying to contribute to example zoo project in GSoC this year
<TarekNasser[m]> What can I do next to enhance my chances?
<TarekNasser[m]> I did an example using Python and made a pull request in the example repository
<zoq[m]> zoq[m]: Also, if you are interested, you can take a look at some of the open issues or PR, having more eyes on code is always a good thing.
<zoq[m]> s/PR/PRs/
<zoq[m]> <GopiMTatiraju[m]> "Hey zoq , I commented on the..." <- Looks like, you figured it out, column major :)
<GopiMTatiraju[m]> zoq: rcurtin done with the implementation, also added tests.
<GopiMTatiraju[m]> Pushed the code...
<rcurtin[m]> zoq: do you think we should do a new ensmallen release now that the Adam fix is merged? or do you want to merge #211 first too?
<rcurtin[m]> (and #231)
<zoq[m]> Ideally we can include #211 as well, it's ready from my side.
<zoq[m]> I think #231 needs some extra work, so we can just include it in the next one.
<rcurtin[m]> ahh, ok, I'll try to review that today and then we can finally get it merged. hopefully before the two year mark 😄