Unstability in training in RNN about adabelief-optimizer HOT 7 CLOSED

Just an update @juntang-zhuang . Your modified Ranger got the best accuracy with half epochs compared to regular Ranger. It is already my favorite optimizer. Thank you!

from adabelief-optimizer.

Thanks for the feedback. I think you are right, and this might be caused by the fact that the update is roughly m_t / sqrt(. (gt-mt)^2 ), the denominator is sometimes too small. Even if the denominator is small for one element, that element will explode. This is a issue I'm trying to fix in the next release, for example hard thres 1/sqrt( (gt-mt)^2 ) to a rather large value.
Please keep this issue open as a reminder of problems to fix for improvement.

from adabelief-optimizer.

I noticed that the learning rate is quite large, after reduction it's still 3e-3. Perhaps a large lr also causes the instability.

from adabelief-optimizer.

Wow, excited to hear that, thanks so much for trying it out.

from adabelief-optimizer.

Could be entirely unrelated to this issue, but at first step in AdaBelief we havem_t = grad, causing v_t = 0 and step_size = step_size / epsilon_t which seems like unintended behaviour.

Edit2:
Deleted my earlier further comments as they were not exactly correct.

from adabelief-optimizer.

Thanks for comment @henriyl . The detailed implementation is definitely not perfect now and might suffer from numerical issues, and you refer to a very good point. For this paper, it's more like a "proof-of-idea" considering the key modification of Adam(W) is only 2 lines of code, therefore many details are not well solved (these details might not be a big problem for CV task but might be more serious in RNN with exploding or vanishing gradient). We are working on the improvement, both in implementation and on theory (personally I guess the convergence bound is too loose in the paper). Thanks again for pointing this out.

from adabelief-optimizer.

@bratao Just an update, I might confuse "gradient clip" with "gradient threshold" before, please see the discussion in readme.md. Perhaps "gradient clip" still helps, which shrinks vector amplitude but keeps the direction, but it might require different clip ranges from Adam; while "gradient threshold" is element-wise operation, and for each element outputs a value in a fixed range, and each dimension of the parameter is independently thresholded, this might cause 0 denominator.

from adabelief-optimizer.