Skip to content

Commit bf8bf3e

Browse files
committed
update
1 parent c979fd0 commit bf8bf3e

5 files changed

Lines changed: 5 additions & 0 deletions

File tree

README.md

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -32,6 +32,11 @@ It improves throughput and reduces memory usage with minimal accuracy drop.
3232
Shrinks KV cache size during decoding, enabling longer generations under memory constraints.
3333

3434
### Key Results
35+
36+
<div align=center>
37+
<img width=90% src="./images/redundancy.png"/>
38+
</div>
39+
3540
<div align=center>
3641
<img width=90% src="./images/accuracy.png"/>
3742
</div>

images/accuracy.png

42.2 KB
Loading

images/efficiency.png

-706 Bytes
Loading

images/overview.png

-104 KB
Binary file not shown.

images/redundancy_rate.png

32.8 KB
Loading

0 commit comments

Comments
 (0)