Qwen

# 1. Transformers로 로컬 실행 (GPU 필요)
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen2.5-7B-Instruct"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

messages = [
    {"role": "system", "content": "당신은 친절한 AI 비서입니다."},
    {"role": "user", "content": "파이썬으로 피보나치 수열을 구현해주세요."}
]

text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True
)
inputs = tokenizer(text, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=512,
    temperature=0.7,
    do_sample=True
)
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

# 2. Ollama로 간편 실행 (설치: ollama pull qwen2.5:7b)
import ollama

response = ollama.chat(
    model='qwen2.5:7b',
    messages=[
        {'role': 'user', 'content': '한국의 수도는 어디인가요?'}
    ]
)
print(response['message']['content'])

# 3. DashScope API 호출 (Alibaba Cloud)
import dashscope
from dashscope import Generation

dashscope.api_key = "your-api-key"

response = Generation.call(
    model="qwen-max",
    messages=[
        {"role": "system", "content": "당신은 전문 번역가입니다."},
        {"role": "user", "content": "Translate to Korean: Hello, how are you?"}
    ],
    result_format='message'
)
print(response.output.choices[0].message.content)

# 4. OpenAI 호환 API (vLLM 서버)
from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8000/v1",  # vLLM 서버 주소
    api_key="not-needed"
)

response = client.chat.completions.create(
    model="Qwen/Qwen2.5-7B-Instruct",
    messages=[
        {"role": "user", "content": "JSON 형식으로 한국 대도시 5개를 알려주세요."}
    ],
    response_format={"type": "json_object"}
)
print(response.choices[0].message.content)

# 5. Qwen-VL 멀티모달 (이미지 + 텍스트)
from transformers import Qwen2VLForConditionalGeneration, AutoProcessor

model = Qwen2VLForConditionalGeneration.from_pretrained(
    "Qwen/Qwen2-VL-7B-Instruct",
    torch_dtype="auto",
    device_map="auto"
)
processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-7B-Instruct")

messages = [{
    "role": "user",
    "content": [
        {"type": "image", "image": "https://example.com/image.jpg"},
        {"type": "text", "text": "이 이미지를 설명해주세요."}
    ]
}]
# ... 처리 및 생성