[papers] Batch size

Pavel Gladkov
Nov 8 · 1 min read

Recently there was interesting discussion in twitter about optimal batch size for training neural networks.

The original paper contains two conclusions:

  • It is better to use batch size less than 32;
  • It affects optimal learning rate range for reliable convergence.

Materials for further reading:

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade