Shrinks 1 GB to 105 MB – Takes a Week to Unpack
3 Articles
3 Articles
Tufazip claims to have safely compressed a gigabyte of Wikipedia text at 105.39 MB with a Transformer as the only predictor. Space saving, however, requires two Nvidia B200 GPUs and about a week to compress or decompress.
One gigabyte of Wikipedia losslessly compressed to 105.39MB with self-learning AI transformer model — 'tufazip' needs two Nvidia B200 GPUs and about a week to compress or decompress
Tufa Labs’ tufazip compresses 1GB of Wikipedia to 105.39MB with a 163M-parameter Transformer as its only predictor, on two Nvidia B200s.
Tufa Labs, a small group of AI researchers, has used an AI model to make 1 GB of Wikipedia text almost 90 percent smaller. The AI learns as it compresses Tufa Labs researcher Tommy He developed the compression program tufazip and tested it on enwik9 – what is that, you ask, because we did: it’s a billion-byte file from the English edition of Wikipedia. The result of 105.39 MB includes both the compressed text and what it takes to unpack the text…
Coverage Details
Bias Distribution
- There is no tracked Bias information for the sources covering this story.
Factuality
To view factuality data please Upgrade to Premium






