undefined | Better HN

0 pointstrott1y ago0 comments

> This is no different than a transformer, which, after all, is bound by a finite state, just organized in a different manner.

It's not just a matter of organizing things differently. Suppose your network dimension and sequence length are both X.

Then your memory usage (per layer) will be O(X^2), while your training update cost will be O(X^3). That's for both Transformers and RNNs.

However, at the end of the sequence, a Transformer layer can look back see O(X^2) numbers, while an RNN can only see O(X) numbers.

0 comments

f_devd1y ago

Transformers actually have an quantifiable state size (see https://hazyresearch.stanford.edu/static/posts/2024-06-22-ac...) although it's anywhere between 200k and 2M floats (for 360M and 1.33B respectively iinm). So a sufficiently sized RNN could have the same state capacity as a transformer.

(this is from the Based paper: https://arxiv.org/pdf/2402.18668)

trottOP1y ago

> Transformers actually have an quantifiable state size

Are you griping about my writing O(X^2) above instead of precisely 2X^2, like this paper? The latter implies the former.

> So a sufficiently sized RNN could have the same state capacity as a transformer.

Does this contradict anything I've said? If you increase the size of the RNN, while keeping the Transformer fixed, you can match their recurrent state sizes (if you don't run out of RAM or funding)

f_devd1y ago

I was responding to

> a Transformer layer can look back see O(X^2) numbers, while an RNN can only see O(X) numbers

The thing is RNN can look back infinitely if you don't exceed the state capacity. For transformers the state it is defined semi-implicitly (you can change the hidden dims but you cannot extend the look back; ignoring transformer-xl et al.) defined by the amount of tokens, for an RNN it's defined explicitly by the state size.

The big-O here is irrelevant for the architectures since it's all in the configuration & implementation of the model; i.e. there is no relevant asymptote to compare.

As an aside this was what was shown in the based paper, the fact that you can have a continuity of state (as with RNN) while have the same associative recall capability as a transformer (the main downfall of recurrent methods at that point).

1 more reply

slashdave1y ago

Simplistic thinking. An RNN hidden parameter space of high dimension provides plenty of room for linear projections of token histories. I think people just do not realize just how huge R^N can be.

trottOP1y ago

> Simplistic thinking. An RNN hidden parameter space of high dimension provides plenty of room for linear projections of token histories. I think people just do not realize just how huge R^N can be.

16N bits as hard limit, but more realistically, about 2N bits or less of useful information probably.

You'd need to grow the network dimension in proportion to the maximum sequence length just to avoid the information theoretical limit.

j / k navigate · click thread line to collapse

0 comments

f_devd1y ago

(this is from the Based paper: https://arxiv.org/pdf/2402.18668)

trottOP1y ago

> Transformers actually have an quantifiable state size

Are you griping about my writing O(X^2) above instead of precisely 2X^2, like this paper? The latter implies the former.

> So a sufficiently sized RNN could have the same state capacity as a transformer.

Does this contradict anything I've said? If you increase the size of the RNN, while keeping the Transformer fixed, you can match their recurrent state sizes (if you don't run out of RAM or funding)

f_devd1y ago

I was responding to

> a Transformer layer can look back see O(X^2) numbers, while an RNN can only see O(X) numbers

The big-O here is irrelevant for the architectures since it's all in the configuration & implementation of the model; i.e. there is no relevant asymptote to compare.

1 more reply

slashdave1y ago

Simplistic thinking. An RNN hidden parameter space of high dimension provides plenty of room for linear projections of token histories. I think people just do not realize just how huge R^N can be.

trottOP1y ago

> Simplistic thinking. An RNN hidden parameter space of high dimension provides plenty of room for linear projections of token histories. I think people just do not realize just how huge R^N can be.

16N bits as hard limit, but more realistically, about 2N bits or less of useful information probably.

You'd need to grow the network dimension in proportion to the maximum sequence length just to avoid the information theoretical limit.

j / k navigate · click thread line to collapse