search

dselivanov / rsparse

Fast and accurate machine learning on sparse matrices - matrix factorizations, regression, classification, top-N recommendations.
https://www.slideshare.net/DmitriySelivanov/matrix-factorizations-for-recommender-systems
170 stars 31 forks source link

Deployment as RESTful service #4

Open dselivanov opened 7 years ago

dselivanov commented 7 years ago

Several choices:

talegari commented 7 years ago

opencpu ?

pshashk commented 7 years ago

jug? There is also a node.js based plugin for parallel processing, but I haven't tried it.

Things like request memoization and approximate nearest neighbor search (e.g., Annoy if L2 norms are approximately equal) might also be useful for serving trained models.

dselivanov commented 6 years ago

Sadly jug is orphaned now.

filipwastberg commented 5 years ago

Is there a plumber vignette coming for this?

dselivanov commented 5 years ago

https://github.com/dselivanov/RestRserve

пт, 1 февр. 2019 г., 13:32 Filip Wästberg notifications@github.com:

Is there a plumber vignette coming for this?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/dselivanov/rsparse/issues/4#issuecomment-459661512, or mute the thread https://github.com/notifications/unsubscribe-auth/AE4u3WELg6TKzgYghJYjR39pWOIQbOgKks5vJAmQgaJpZM4PVp0S .

filipwastberg commented 5 years ago

Thanks, seems like a great alternative. I'll try and figure out how to user rsparse with it. Nevertheless, it would be helpful with a vignette for how you could deploy rsparse as a microservice.