3 ms·
Doesn’t compression ratio depend heavily on the file type / entropy of the file? You can probably zip 10 GB of zeroes to under 1 KB. Without taking the content
by echoangle 2y ago
Doesn’t compression ratio depend heavily on the file type / entropy of the file? You can probably zip 10 GB of zeroes to under 1 KB. Without taking the content of the data into account, you can’t just say “more than 10x is impossible because it’s uncommon”. The benchmarks you’re looking at are compressing text, I think. Maybe text is higher entropy than the Neuralink data?