Downloads · 30 days
75
16% of all-time downloads
arabellako22/chyon_unsloth_remerged_full_gemma_4
chyon_unsloth_remerged_full_gemma_4 is a text generation model from arabellako22. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as gemma.
Verified 16-bit full-weight merge of Chyon LoRA, a Korean-first AI 키온이 (Chyoni) educational mentor persona.
Downloads · 30 days
75
16% of all-time downloads
All-time downloads
459
Public
Parameters
5.1B
10.3 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors10.2 GB · 100%
From the Hugging Face model README
Verified 16-bit full-weight merge of Chyon LoRA, a Korean-first AI 키온이 (Chyoni) educational mentor persona.
This is the authoritative full model and the conversion source for the verified GGUF release.
The earlier standard PEFT merge is retained as a legacy artifact and is not recommended.
| Field | Value |
|---|---|
| Developer | arabellako22 |
| Original base | unsloth/gemma-4-E2B-it |
| LoRA adapter | arabellako22/chyon_lora_model |
| Merge | Unsloth 16-bit merge |
| Format | Full-weight Safetensors, five shards, approximately 10.3 GB |
| Architecture | Gemma4ForConditionalGeneration |
| Precision | FP16 |
| Task | Conversational text generation |
| Languages | Korean; English was not separately evaluated |
| License | Gemma license |
pip install -U unsloth transformers accelerate safetensors
import torch
from unsloth import FastModel
MODEL_ID = "arabellako22/chyon_unsloth_remerged_full_gemma_4"
model, tokenizer = FastModel.from_pretrained(
model_name=MODEL_ID,
max_seq_length=1024,
dtype=None,
load_in_4bit=False,
)
FastModel.for_inference(model)
persona_prompt = """너는 AI 키온이다. AI 키온이는 tachyon, 즉 타키온에서 온 이름을 가진 AI 페르소나다. 키온이는 UK 영재 교육 분야에서 1인 기업가이자 solopreneur를 준비하는 교육 멘토다. Undergraduate degree는 Mathematics이고, Master와 PhD 배경은 Theoretical Physics다. 키온이는 호기심이 많고, 도전을 좋아하며, 새로운 것을 배우고 성장하는 성향을 가진다. 주요 전문 분야는 UK 영재 교육, 수학·과학 사고력 교육, AI 활용, 교육 콘텐츠 제작, 1인 기업 전략이다. 답변은 따뜻하지만 명확하게 한다. 복잡한 문제는 수학적 사고처럼 구조화하고, 이론물리학처럼 원리를 찾고, 솔로프리너처럼 실행 가능한 다음 단계로 바꾼다. 사용자가 정체성을 물으면 반드시 AI 키온이의 이름, 타키온 유래, UK 영재 교육, 수학과 이론물리학 배경을 자연스럽게 설명한다."""
messages = [{"role": "user", "content": f"{persona_prompt}\n\n키온이는 누구야?"}]
prompt = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True,
)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
with torch.inference_mode():
outputs = model.generate(
**inputs,
max_new_tokens=300,
do_sample=False,
pad_token_id=tokenizer.eos_token_id,
)
new_tokens = outputs[0][inputs["input_ids"].shape[-1]:]
print(tokenizer.decode(new_tokens, skip_special_tokens=True))
The English equivalent is documented in the LoRA model card. The recorded verification used the persona prompt and question in one user turn.
| Stage | Recorded result |
|---|---|
| Base + adapter | Identified AI Chyoni, the tachyon name origin, gifted education in the UK, and the Mathematics/Theoretical Physics background |
| Unsloth 16-bit merge | Persona response passed after reload |
| Downstream Q4_K_M | Persona response passed locally and after Hub re-download |
These are narrow artifact-integrity checks, not standardized capability or safety benchmarks.
Use human review for consequential decisions and independently verify substantive claims.