metadata

license: cc-by-nc-4.0
base_model: Edentns/DataVortexS-10.7B-dpo-v1.11
tags:
  - trl
  - dpo
  - generated_from_trainer
model-index:
  - name: nhn_dpo_v3_DataVortexS-10.7B-dpo-v1.11_DPO
    results: []

ENERGY-DRINK-LOVE/DataVortexS_dpov3

Our Team

Youjin Chung
Jingyeom Kim

Model

Base Model

Edentns/DataVortexS-10.7B-dpo-v1.11

Hardware and Software

Hardware: A100 * 8 for training our model
Deepspeed library & Huggingface TRL Trainer

Dataset

DPO_dataset
- 자체 제작 dpo dataset(AI-hub dataset 활용)
- OpenOrca DPO 등 영어 데이터셋 번역(ENERGY-DRINK-LOVE/translate_share_gpt_dedup_llama_SFT_1024, 자체모델 활용)

Training Method

Benchmark

Ko LM Eval Harness

Ko-LLM-Leaderboard

(240316기준 7등)

Average	Ko-ARC	Ko-HellaSwag	Ko-MMLU	Ko-TruthfulQA	Ko-CommonGen V2
60.18	56.23	69.15	52.76	67.87	54.9