Closed KristianHolsheimer closed 7 years ago
Lack of maintenance is definitely the issue here. I'm 'ccing Jake who may have time to look into. I'm superoverloaded for the next several weeks.
-John
On Fri, Apr 21, 2017 at 3:17 AM, Kristian Holsheimer < notifications@github.com> wrote:
Hi there,
I'm trying to train word embeddings using vw's low-rank approximation feature. I noticed the nice wiki page https://github.com/JohnLangford/vowpal_wabbit/wiki/Matrix-factorization-example that describes how to extract the latent factors.. exactly what I need.. Awesome!
I tried to compile the gd_mf_weights tool,
$ cd library $ make gd_mf_weights
but I got the this stack trace https://pastebin.com/cZtmjvX5.
Note I also noticed that there was another issue with gd_mf_weights some time ago [#933 https://github.com/JohnLangford/vowpal_wabbit/issues/933]. I opened a new issue, because the only overlap I could see is that the gd_mf_weights doesn't seem to be maintained at the moment.
Thanks!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JohnLangford/vowpal_wabbit/issues/1229, or mute the thread https://github.com/notifications/unsubscribe-auth/AAE25sWpTL8TjHgiz42tdkGS9Pv01XLsks5ryFgjgaJpZM4ND8ob .
Okay, no worries, it would've been really nice. For the time being, I'll just use my own little hack parsing the --audit
output.
Thanks John!
Huh, it actually does compile, at least on my machine and travis. Just type 'make all' in the root directory.
Okay, I'll give it another try. Thanks for looking into it
Hi there,
I'm trying to train word embeddings using vw's low-rank approximation feature. I noticed the nice wiki page that describes how to extract the latent factors.. exactly what I need.. Awesome!
I tried to compile the
gd_mf_weights
tool,but I got the this stack trace.
Note I also noticed that there was another issue with
gd_mf_weights
some time ago [#933]. I opened a new issue, because the only overlap I could see is that thegd_mf_weights
doesn't seem to be maintained at the moment.Thanks!