Yejin Korean 8B v3 cpt5000-5750 + Qwen3.6-27B distill lr2e-4 soup4 nothink-v2

Korean instruction-following language model.

Chat Template

<s><|user|>\n{instruction}<|end|>\n<|assistant|>\n<think> </think>{response}<|end|>

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("cooler8/yejin-korean-8b-v3-cpt5000-5750-distill-qwen36-lr2e4-soup4-nothink-v2", torch_dtype="bfloat16", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("cooler8/yejin-korean-8b-v3-cpt5000-5750-distill-qwen36-lr2e4-soup4-nothink-v2")

messages = [{"role": "user", "content": "한국의 수도는 어디인가요?"}]
inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt", return_dict=True).to(model.device)
inputs.pop("token_type_ids", None)
outputs = model.generate(**inputs)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))

Training

Uniform weight average (model soup) of four Yejin fine-tunes: continued-pretraining checkpoints 5,000, 5,250, 5,500 and 5,750 of Yejin 8B v3 (pretrained from scratch by Yejin: cooler8/yejin-korean-8b-v3-base), each with the Yejin chat vector from our own SFT model and a LoRA (rank 64, lr 2e-4). The LoRA data mixes our phase-1 format data (45,013 samples) with 6,052 short Korean solutions written by Qwen/Qwen3.6-27B (Apache-2.0) for questions from our own puzzle generator, KMMLU train and AI Hub school exams; only solutions that reached the known answer were kept (sequence-level distillation; no teacher weights or logits). No benchmark test items were used, and the teacher solutions share no word 13-gram with KMMLU-Pro, KMMLU test, CLIcK, MuSR or Korean MuSR items.

License

Apache 2.0