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
4.18k
stars
222
forks
source link
Feature(LLMLingua-2): update the meetingbank datasets
#160
Closed
iofu728
closed
1 month ago
iofu728
commented
1 month ago
Update the MeetingBank datasets, docs, and scripts;
Fix the typos in #157.