[abstract] This approach significantly reduces the KV cache size relative to traditional multi-head attention
[3.3] For saving the KV cache, only the intermediate
latent representations need to be stored: [latex] where r is much smaller than nh · dh [n-sub-h, d-sub-h]
[background] In traditional multi-head attention you must cache full key and value matrices of size T x (nh · dh) where T is the token length, nh is the number of attention heads, dh is the dimensionality of each individual head
sounds like a big win for memory constrained environments like local inference
killerstorm · 24m ago
Another paper related to attention distillation, although doing something far more radical: transformer attention is distilled onto RWKV-like model: https://huggingface.co/papers/2505.03005
octocop · 4h ago
These titles need to stop, we've seen that in fact it is not all you need.
ghc · 45m ago
It's become the equivalent of the stupid faces on YouTube thumbnails.
insin · 2h ago
Why we're moving away from all you need considered harmful
jsheard · 1h ago
Those words are all you need to get to the top of HN though. Think of the karma!
seeknotfind · 4h ago
All you need titles stopping is all you need.
Etheryte · 3h ago
All you need is love, and for these titles to stop. (But they won't do that.)
EGreg · 4h ago
We need more than that, and all you need to stop saying that!!
tankenmate · 4h ago
The title of this paper is a reference to a previous paper titled "Attention Is All You Need"[0][1]. This seminal work described the transformer model that is the basis for almost all LLMs, and is almost certainly the most cited paper on AI even though it was only published in 2017.
Right, it's an 8 year old reference that's been made hundreds of times.
People seem to love going to the references graveyard, digging up tired and dead ones and drag them around town hoping everyone thinks they're clever.
Also this was from 3 months ago.
nihzm · 3h ago
It has definitely been overused by too many authors.
This reminds me a passage of Orwell's essay "Politics and the English Language":
> A newly−invented metaphor assists thought by evoking a visual image, while on the other hand a metaphor which is technically "dead" (e.g., iron resolution) has in effect reverted to being an ordinary word and can generally be used without loss of vividness. But in between these two classes there is a huge dump of worn−out metaphors which have lost all evocative power and are merely used because they save people the trouble of inventing phrases for themselves
tankenmate · 3h ago
By that argument you must also hate anything that mentions the term "considered harmful", or makes any form of derivative cultural reference (like just about every episode of the Simpsons). Why do you let it get to you?
netdevphoenix · 3h ago
Why is this the most cited paper in AI and not the original 1943 paper who started it all?
zaptrem · 3h ago
Transformers are what made ML infinitely scalable and caused a huge amount of progress in very few years since everyone could just go scale things. However, idk how many of those papers actually even cite the transformer paper?
netdevphoenix · 45m ago
As I understand, the transformer architecture is built on deep learning.
Would you say that transformers made a bigger progress RELATIVE to the progress made by deep learning?
AFAIK, before the first wave of AI powered apps that were visible to users appeared thanks to deep learning in the early 10s. Users went from nothing to fancy AI features, the question is likely subjective but is the jump from nothing to fancy AI features the same as the jump from fancy AI features to GenAI in relative terms?
We can't forget that new tech builds upon older tech hence merits need to be relative
tankenmate · 2h ago
I just checked Google Scholar, not perfect but good for an indicative; "A logical calculus of the ideas immanent in nervous activity" [WS McCulloch, W Pitts - The bulletin of mathematical biophysics, 1943] has ~33,000 citations, and "Attention is all you need" [A Vaswani, N Shazeer, et al, Advances in Neural Information Processing Systems, 2017] has ~180,000 citations.
tankenmate · 3h ago
Probably because of the modern "publish or perish" mantra led to an exponential growth in publications, and "newer is better" means that newer impactful papers get cited more than older impactful publications. But that thesis is probably a paper in itself (of the meta analysis navel gazing variety).
magicalhippo · 1h ago
I'm just following the field from the sidelines, but this looks interesting to me. Especially the increase in expressiveness that the new model allows for over GQA, at the cost of just ~10% more memory, and the fact that you can convert existing GQA models like LLaMA, Qwen etc with just a bit of fine-tuning.
Perhaps a trivial insight but I feel a lot of progress often comes in the form of generalizations, where existing approaches can be seen as special cases. Here the authors show that Group Query Attention (GQA) and Multi-Query Attention (MQA) falls out as special cases of their new model.
edit:
Adding my own summary, as I understand it.
The key to what they're doing, no pun intended, is to rely on the fact that large, high-dimensional, matrices may contain a lot of redundant information. Thus one may be able to find an good approximation which has less redundant information, by going through an intermediary stage which has fewer dimensions.
A n-by-m matrix M takes n-dimensional vectors and transforms them to m-dimensional vectors. The trick here is to replace matrix A by two matrices, L and R, which are n-by-r and r-by-m respectively, where r is smaller than n and m. This is called a low-rank approximation.
In a sense you're "straining the matrix", by forcing the information to pass through an intermediary, low-dimensional vector.
The memory savings come from the fact that matrix A has n*m entries, while L and R have n*r and r*m entries respectively. Say n = m = 100 and r = 20, that means A has 100*100 = 10k entries, while L and R have just 100*20 + 20*100 = 4k entries in total.
The trick itself is not new, for example it is also used in LoRA where an additional low-rank approximation matrix is used to tweak the output of an existing model. The low rank means there's far fewer the matrix entries, aka parameters, to train than if one had used a regular fully dense matrix.
The extra expressiveness of MLA comes from the fact that in GQA, in order to save memory, some of the matrices are actually built by gluing copies of a narrower matrix together. This means the information in the glued-up matrices are very redundant and fixed in a certain way, and thus are restricted in how they can transform the inputs.
By using the low-rank approximation instead, the information in the full, reconstructed matrices are not fixed in the same way compared to the glued-up result. Thus the inputs can be transformed in a less restrictive way, leading to the increase in expressiveness.
The GQA method saves a bit more memory compared to MLA as the narrower matrices are even smaller than the low-rank matrices in MLA, but at the cost of expressiveness.
olq_plo · 7h ago
Very cool idea. Can't wait for converted models on HF.
wiz21c · 4h ago
Not quite related, but do the mamba models gain ground ?
Still wrapping my head around this architecture, but the idea of reducing headcount while maintaining performance is compelling. Would love to see a benchmark against something like FlashAttention.
kavalg · 6h ago
My (possibly wrong) TLDR: TransMLA is a method to "compress" an already trained GQA model, with the additional option to further fine tune it. Shall make inference faster.
yorwba · 6h ago
It is not a method to compress a Grouped-Query Attention model, but to expand it into an equivalent Multi-head Latent Attention model with the same key-value cache size but larger effective key/value vectors and a correspondingly larger number of trainable parameters. With additional training, you can then obtain a better model that only uses a little bit more memory.
[3.3] For saving the KV cache, only the intermediate latent representations need to be stored: [latex] where r is much smaller than nh · dh [n-sub-h, d-sub-h]
[background] In traditional multi-head attention you must cache full key and value matrices of size T x (nh · dh) where T is the token length, nh is the number of attention heads, dh is the dimensionality of each individual head
sounds like a big win for memory constrained environments like local inference
[0] https://arxiv.org/abs/1706.03762 [1] https://en.wikipedia.org/wiki/Attention_Is_All_You_Need
People seem to love going to the references graveyard, digging up tired and dead ones and drag them around town hoping everyone thinks they're clever.
Also this was from 3 months ago.
> A newly−invented metaphor assists thought by evoking a visual image, while on the other hand a metaphor which is technically "dead" (e.g., iron resolution) has in effect reverted to being an ordinary word and can generally be used without loss of vividness. But in between these two classes there is a huge dump of worn−out metaphors which have lost all evocative power and are merely used because they save people the trouble of inventing phrases for themselves
Would you say that transformers made a bigger progress RELATIVE to the progress made by deep learning? AFAIK, before the first wave of AI powered apps that were visible to users appeared thanks to deep learning in the early 10s. Users went from nothing to fancy AI features, the question is likely subjective but is the jump from nothing to fancy AI features the same as the jump from fancy AI features to GenAI in relative terms?
We can't forget that new tech builds upon older tech hence merits need to be relative
Perhaps a trivial insight but I feel a lot of progress often comes in the form of generalizations, where existing approaches can be seen as special cases. Here the authors show that Group Query Attention (GQA) and Multi-Query Attention (MQA) falls out as special cases of their new model.
edit:
Adding my own summary, as I understand it.
The key to what they're doing, no pun intended, is to rely on the fact that large, high-dimensional, matrices may contain a lot of redundant information. Thus one may be able to find an good approximation which has less redundant information, by going through an intermediary stage which has fewer dimensions.
A n-by-m matrix M takes n-dimensional vectors and transforms them to m-dimensional vectors. The trick here is to replace matrix A by two matrices, L and R, which are n-by-r and r-by-m respectively, where r is smaller than n and m. This is called a low-rank approximation.
In a sense you're "straining the matrix", by forcing the information to pass through an intermediary, low-dimensional vector.
The memory savings come from the fact that matrix A has n*m entries, while L and R have n*r and r*m entries respectively. Say n = m = 100 and r = 20, that means A has 100*100 = 10k entries, while L and R have just 100*20 + 20*100 = 4k entries in total.
The trick itself is not new, for example it is also used in LoRA where an additional low-rank approximation matrix is used to tweak the output of an existing model. The low rank means there's far fewer the matrix entries, aka parameters, to train than if one had used a regular fully dense matrix.
The extra expressiveness of MLA comes from the fact that in GQA, in order to save memory, some of the matrices are actually built by gluing copies of a narrower matrix together. This means the information in the glued-up matrices are very redundant and fixed in a certain way, and thus are restricted in how they can transform the inputs.
By using the low-rank approximation instead, the information in the full, reconstructed matrices are not fixed in the same way compared to the glued-up result. Thus the inputs can be transformed in a less restrictive way, leading to the increase in expressiveness.
The GQA method saves a bit more memory compared to MLA as the narrower matrices are even smaller than the low-rank matrices in MLA, but at the cost of expressiveness.
Answering my own question: https://www.reddit.com/r/MachineLearning/comments/1hpg91o/d_...