euagendas / m3inference

A deep learning system for demographic inference (gender, age, and individual/person) that was trained on massive Twitter dataset using profile images, screen names, names, and biographies
http://www.euagendas.org
GNU Affero General Public License v3.0
146 stars 56 forks source link

Improve requests speed #27

Closed eslamahmed235 closed 2 years ago

eslamahmed235 commented 2 years ago

Hello I already used M3inferince and work well. but in large scale of data it's not fast enough. It takes an average 1.3 second for each account while running on kaggel GPU. If there are any advice or technique to speed up it's progress.

computermacgyver commented 2 years ago

I'm glad you're finding m3 useful. We welcome pull requests and optimizations. The main thing I would do is pack images into batches. There's not enough detail here on how you're using m3, but if you do all your preprocessing and downloading of the images first, m3 will batch process the inference stage.