4–5× Less Memory, 1.6× Fewer Tokens with BoetByteTokenization #315
Boetros (Boetros)
started this conversation in
Ideas
Replies: 1 comment
|
Impressive, most impressive!!! |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Would this be an idea to squeeze even more out of the 1-bit model?
4–5× Less Memory, 1.6× Fewer Tokens: The BoetByteTokenization Leap for LLMs https://medium.com/@boetwesterburgen/can-we-make-chatgpt-4-faster-a-tiny-token-trick-1c7018651e9a
All reactions