20 lines
562 B
Python
20 lines
562 B
Python
|
import requests
|
||
|
|
|
||
|
|
OLLAMA_URL = "http://127.0.0.1:11434/api/generate"
|
||
|
|
MODEL_NAME = "qwen3:8b"
|
||
|
|
|
||
|
|
def call_qwen(prompt: str, temperature: float = 0.8, max_tokens: int = 512) -> str:
|
||
|
|
payload = {
|
||
|
|
"model": MODEL_NAME,
|
||
|
|
"prompt": prompt,
|
||
|
|
"stream": False,
|
||
|
|
"options": {
|
||
|
|
"temperature": temperature,
|
||
|
|
"num_predict": max_tokens
|
||
|
|
}
|
||
|
|
}
|
||
|
|
resp = requests.post(OLLAMA_URL, json=payload, timeout=60)
|
||
|
|
resp.raise_for_status()
|
||
|
|
data = resp.json()
|
||
|
|
return data.get("response", "") or data.get("text", "")
|