tzem-instruct-v1.1 / README.md
blueapple8259's picture
Update README.md
2ae5ff3 verified
metadata
license: mit
datasets:
  - MarkrAI/KOpen-HQ-Hermes-2.5-60K
language:
  - ko
base_model:
  - blueapple8259/tzem-instruct
pipeline_tag: question-answering
library_name: transformers

tzem-instruct λͺ¨λΈμ— μΆ”κ°€λ‘œ νŒŒμΈνŠœλ‹ ν•œ λͺ¨λΈμž…λ‹ˆλ‹€.

ν”„λ‘¬ν”„νŠΈ ν…œν”Œλ¦Ώ

{system}
**μ‚¬μš©μž:** {prompt}
**인곡지λŠ₯:**

μ‚¬μš©

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "blueapple8259/tzem-instruct-v1.1"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

system = "여기에 μ‹œμŠ€ν…œ λ©”μ‹œμ§€ μž…λ ₯"
prompt = "여기에 ν”„λ‘¬ν”„νŠΈ μž…λ ₯"

text = f"{system}\n**μ‚¬μš©μž:** {prompt}\n**인곡지λŠ₯:".strip()

inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(
    **inputs,
    max_new_tokens=100,
    pad_token_id=tokenizer.pad_token_id,
    eos_token_id=tokenizer.eos_token_id,
    do_sample=True,
    top_p=0.2,
)

output = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(output)

μ˜ˆμ‹œ

**μ‚¬μš©μž:** 건강을 μœ μ§€ν•˜κΈ° μœ„ν•œ μ„Έ κ°€μ§€ νŒμ„ μ•Œλ €μ£Όμ„Έμš”.
**인곡지λŠ₯:** 1. κ· ν˜• 작힌 식단을 μ„­μ·¨ν•˜μ„Έμš”: 과일, μ±„μ†Œ, 톡곑물, μ €μ§€λ°© λ‹¨λ°±μ§ˆ, μ €μ§€λ°© μœ μ œν’ˆμ„ λ‹€μ–‘ν•˜κ²Œ μ„­μ·¨ν•˜μ„Έμš”. κ°€κ³΅μ‹ν’ˆ, 섀탕 음료, κ³Όλ„ν•œ μ–‘μ˜ 포화 μ§€λ°©κ³Ό 트랜슀 지방을 ν”Όν•˜μ„Έμš”.

2. κ·œμΉ™μ μœΌλ‘œ μš΄λ™ν•˜μ„Έμš”: μ£Ό 5일, 30λΆ„ μ΄μƒμ˜ 쀑강도 μš΄λ™μ„ ν•˜μ„Έμš”. μ£Ό 3일 이상, μ£Ό 4일 이상, μ£Ό 1일 이상 μš΄λ™ν•΄μ•Ό ν•©λ‹ˆλ‹€.

3. μΆ©λΆ„ν•œ μˆ˜λ©΄μ„ μ·¨ν•˜μ„Έμš”: λŒ€λΆ€λΆ„μ˜ 성인은 ν•˜λ£»λ°€μ— 7~8μ‹œκ°„μ˜ 수면이 ν•„μš”ν•©λ‹ˆλ‹€. 맀일 7~8μ‹œκ°„μ˜ μˆ˜λ©΄μ„ λͺ©ν‘œλ‘œ ν•˜μ„Έμš”.

데이터셋