Sakura-SOLRCA-Instruct-DPO

(주)미디어그룹사람과숲과 (주)마커의 LLM 연구 컨소시엄에서 개발된 모델입니다

Model Details

Model Developers Kyujin Han (kyujinpy)

Method
Using DPO method.
With Intel/orca_dpo_pairs.

I shared the information about my model. (training and code)
Please see: ⭐Sakura-SOLAR.

Model Benchmark

Open leaderboard

• Follow up as link.

Model	Average	ARC	HellaSwag	MMLU	TruthfulQA	Winogrande	GSM8K
Sakura-SOLRCA-Instruct-DPO	74.05	71.16	88.49	66.17	72.10	82.95	63.46
Sakura-SOLAR-Instruct-DPO-v2	74.14	70.90	88.41	66.48	71.86	83.43	63.76
kyujinpy/Sakura-SOLAR-Instruct	74.40	70.99	88.42	66.33	71.79	83.66	65.20

Implementation Code

### KO-Platypus
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

repo = "kyujinpy/Sakura-SOLRCA-Instruct-DPO"
OpenOrca = AutoModelForCausalLM.from_pretrained(
        repo,
        return_dict=True,
        torch_dtype=torch.float16,
        device_map='auto'
)
OpenOrca_tokenizer = AutoTokenizer.from_pretrained(repo)

---

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	74.05
AI2 Reasoning Challenge (25-Shot)	71.16
HellaSwag (10-Shot)	88.49
MMLU (5-Shot)	66.17
TruthfulQA (0-shot)	72.10
Winogrande (5-shot)	82.95
GSM8k (5-shot)	63.46