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/
vansika__ has quit [Read error: Connection reset by peer]
vansika__ has joined #mlpack
< jeffin143[m]>
walragatver: I will try to
favre49 has joined #mlpack
k3nz0__ has quit [Remote host closed the connection]
k3nz0_ has quit [Remote host closed the connection]
k3nz0 has joined #mlpack
k3nz0_ has joined #mlpack
ImQ009 has joined #mlpack
favre49 has quit [Quit: Lost terminal]
favre49 has joined #mlpack
favre49 has quit [Remote host closed the connection]
ImQ009 has quit [Quit: Leaving]
< shrit[m]>
@rcurtin if you agree, would rename all CLI to IO instead of CMD, all related classes, file names, comments, etc. this will make it the difference clear between IO bindings and CLI11
< rcurtin>
shrit[m]: sure, "IO" is just fine with me unless you have a better idea
< rcurtin>
we should still call the command-line bindings 'cli' bindings (i.e. they should live in the namespace/directory mlpack::bindings::cli)
< shrit[m]>
Exactly, so the file names cli.*** should be kept like that?
< shrit[m]>
I am usually bad at naming things, so I do not have a better idea
< shrit[m]>
I understand, So file name, Cmake binding, and directory stays CLI, otherwise everything else related to namespace goes to CMD
< shrit[m]>
sorry goes to IO
< rcurtin>
right, in any case, it's pretty easy to change either way :)
< rcurtin>
I'll try to make some time to do another pass on the PR today but please forgive me if I don't---my bigger priority is finishing some parts replacements in my car :)
< jeffin143[m]>
kartikdutt18 (@kartikdutt18:matrix.org): does this work as checkpoint
< kartikdutt18[m]>
Periodic save ?
< kartikdutt18[m]>
Periodic save weights only for now, but it can extended to load the last saved weights as well.
< jeffin143[m]>
> Periodic save ?
< jeffin143[m]>
Yes
< jeffin143[m]>
Last saved weights as well mean?
< jeffin143[m]>
Could you elaborate
< kartikdutt18[m]>
Right now, After every k epochs weights are saved but not used. If needed, we could clear the model and start the next epoch with the saved weights. So that later on it can be extended to load best weights based on some metric.
< jeffin143[m]>
> Right now, After every k epochs weights are saved but not used. If needed, we could clear the model and start the next epoch with the saved weights. So that later on it can be extended to load best weights based on some metric.