issues
search
microsoft
/
LLMLingua
To speed up LLMs' inference and enhance LLM's perceive of key information, compress the prompt and KV-Cache, which achieves up to 20x compression with minimal performance loss.
https://llmlingua.com/
MIT License
4.27k
stars
228
forks
source link
Hotfix (LLMLingua): fix the out of range
#87
Closed
iofu728
closed
5 months ago
iofu728
commented
5 months ago
fix the out of range due to the very short sentence.
iofu728
commented
5 months ago
84