mu0gum/AIFT-42dot_LLM-PLM-1.3B-ao-instruct-all-v0.52

AIFT-42dot-LLM-PLM-1.3B-ao-instruct-all-v0.52

베이스 모델 : 42dot/42dot_LLM-PLM-1.3B

학습 데이터 : 자체 제작한 Open Orca 스타일 데이터셋 약 28,000건 (데이터 수량 조정)

학습 방법 : Full finetuning

ko-lm-evaluation-harness(0-shot)

kobest_boolq	kobest_copa	kobest_hellaswag	kobest_sentineg	kohatespeech	kohatespeech_apeach	kohatespeech_gen_bias	korunsmile	nsmc	pawsx_ko
0.5826210826210826	0.68	0.436	0.7758186397984886	0.2908704883227176	0.5082228116710875	0.14225053078556263	0.39027300210119553	0.65938	0.513

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.0.0
• Tokenizers 0.15.0