Skip to content

Latest commit

 

History

History
17 lines (9 loc) · 595 Bytes

README.md

File metadata and controls

17 lines (9 loc) · 595 Bytes

SmallLMEnhance

Using small LM with agent to get a better performance on unfamiliar task

For privacy corcerns, local LM models which run on device are a good solution for user.

The limit of RAM

Small local LM models would be about only 2~3B params. Due to their size, complex task would be challenging.

Using another Large LM as Agent actor

Instead of directly using small LM models as agent actor, I use another Large LM as agent as a provider and agent actor.

Workflow

SmallLMEnhance