Skip to content

Commit

Permalink
update doc
Browse files Browse the repository at this point in the history
  • Loading branch information
Albert Tseng committed Jan 11, 2024
1 parent 1673d81 commit 8ec2f1a
Show file tree
Hide file tree
Showing 3 changed files with 887 additions and 889 deletions.
1 change: 1 addition & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -13,6 +13,7 @@ We also provide a full codebase that allows users to quantize and deploy their o
| **QuIP#** | **2 bit** | **4.159** | **6.529** | **0.595** | **0.786** |

Quantization results on Llama 2 70B. QuIP# achieves near-native performance at 2 bits, outperforming all other presented baselines.
Results for other models available [here](https://docs.google.com/spreadsheets/d/18woLrIBdVGUr9CuFDbK9pl_6QzEBl09hfnoe4Qkg7Hg/edit?usp=sharing).

## ☞ Read more about QuIP# and how it works [here](https://cornell-relaxml.github.io/quip-sharp/)!

Expand Down
Loading

0 comments on commit 8ec2f1a

Please sign in to comment.