3 ms·
Fun to see neural nets pushed to such extremes, really enjoyed the post. > The smallest models had to be trained without data augmentation, as they would not c
by magicalhippo 2y ago
Fun to see neural nets pushed to such extremes, really enjoyed the post.
> The smallest models had to be trained without data augmentation, as they would not converge otherwise.
Was this also the case for the 2-bit model you ended up with?
- cpldcpu 2y agoYes, as far as i remember the limit was somewhere around 1kbyte total parameters size.