(주)미디어그룹사람과숲과 (주)마커의 LLM 연구 컨소시엄에서 개발된 모델입니다
The license is cc-by-nc-sa-4.0.
Ko-Platypus2-13B
Model Details
More detail repo(Github): KO-Platypus
Model Developers Kyujin Han (kyujinpy)
Input Models input text only.
Output Models generate text only.
Model Architecture KO-Platypus2-13B is an auto-regressive language model based on the LLaMA2 transformer architecture.
Base Model hyunseoki/ko-en-llama2-13b
Training Dataset
I use KOpen-platypus.
It is high-quality korean translation dataset about open-platypus.
I use A100 GPU 40GB and COLAB, when trianing.
Model Benchmark
KO-LLM leaderboard
- Follow up as Open KO-LLM LeaderBoard.
| Model | Average | Ko-ARC | Ko-HellaSwag | Ko-MMLU | Ko-TruthfulQA | Ko-CommonGen V2 |
|---|---|---|---|---|---|---|
| KO-Platypus2-13B(ours) | 47.90 | 44.20 | 54.31 | 42.47 | 44.41 | 54.11 |
| hyunseoki/ko-en-llama2-13b | 46.68 | 42.15 | 54.23 | 38.90 | 40.74 | 57.39 |
| MarkrAI/kyujin-CoTy-platypus-ko-12.8b | 46.44 | 34.98 | 49.11 | 25.68 | 37.59 | 84.86 |
| momo/polyglot-ko-12.8b-Chat-QLoRA-Merge | 45.71 | 35.49 | 49.93 | 25.97 | 39.43 | 77.70 |
| KoT-platypus2-7B | 45.62 | 38.05 | 49.63 | 34.68 | 37.69 | 68.08 |
Compare with Top 4 SOTA models. (update: 10/06)
Implementation Code
### KO-Platypus
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
repo = "kyujinpy/KO-Platypus2-13B"
CoT-llama = AutoModelForCausalLM.from_pretrained(
repo,
return_dict=True,
torch_dtype=torch.float16,
device_map='auto'
)
CoT-llama_tokenizer = AutoTokenizer.from_pretrained(repo)
Readme format: kyujinpy/KoT-platypus2-7B