README.md · dddsaty/KoSOLAR-10.7B_DPO_Adapter

metadata

license: apache-2.0
datasets: We-Want-GPU/Yi-Ko-DPO-Orca-DPO-Pairs
language:
  - ko
pipeline_tag: text-generation

Explanation

With the base model, applied DPO to the small amount of layers with the open dataset , saved the adapter part
Attached the base model and the tuned adapter together

Base Model

Used Corpus

Score

Average	Ko-ARC	Ko-HellaSwag	Ko-MMLU	Ko-TruthfulQA	Ko-CommonGen V2
56.24	53.33	64.36	55.63	45.42	62.46

Log

LICENSE