gwonny's picture
Create README.md
6f2c65a verified
|
raw
history blame
352 Bytes
metadata
license: cc-by-nc-4.0
language:
  - ko
  - en

Base Model:

We made a LLM model with davidkim205/nox-solar-10.7b-v4

Model Description

We use QDoRA(256, 64) and NEFTune on LLM Fine-tuning lr = 1e-5, NEFTune = 3

Train Detail

Our changed private data used, 5 epoch train, and parameter changed

Others

We are making LLM model for Kolon !