It starts with the math.
A training run is capped by what fits in GPU memory. We make more fit, with math physicists built to simulate quantum systems on ordinary computers.
The constraint
Everything has to fit.
All of it resident at once. The run is capped by what fits, and teams give up training quality to fit it.
Why it's getting worse
GPUs have plenty of computing power. They run out of memory first.
Buying more hardware helps until the next model arrives. The question that lasts is how much training you can extract per byte.
Where the math comes from
Find the pattern. Store the pattern.
Physicists simulating quantum systems learned to write a very large object down in a fraction of the space, and to measure what was lost. That research is where we start.
What we measure
Three numbers before we call it a result.
Memory saved, throughput cost and quality cost, measured on the same run against dense precision. Compression comes at a cost. We say what it is.
The trade
Choose your trade-off.
Our methods are designed to keep the cost low, and to let you pick the point on the curve that makes sense for your team.
Thesis
It starts with the math
How mathematics from quantum physics applies to AI training today.
Read the long versionEveryone else sells you cheaper chips. We make each chip do more.
Where we are
We're early.
Akavion started with the research. We're running training jobs, measuring, and talking with teams who train models about what they design around.
Researchers and operators.
A small team. We've studied, published and worked at the places below.
