概述
本文主要演示如果在阿里云PAI EAS上面部署DeepSeek模型，以及部署模型后的SDK和API调用。
操作步骤
1、选择7B模型部署
目前7B模型一张A10卡即可满足，部署时间10分钟左右，阿里云A10卡相对充足。
2、模型信息查看及在线调试
- 请求body
{ "messages": [ { "role": "user", "content": "你是谁？" } ], "max_tokens": 2000 }
3、SDK&PostMan调用
- OpenAI SDK
from openai import OpenAI client = OpenAI( base_url="http://1991************.cn-beijing.pai-eas.aliyuncs.com/api/predict/quickstart_20250206_pbxi/v1", api_key="ZDEyMjc0ZTFh***************" ) # Chat completion response = client.chat.completions.create( model='default', messages=[ {"role": "system", "content": "You are a helpful AI assistant"}, {"role": "user", "content": """请介绍一下deepseekr1"""}, ], max_tokens=2000, stream=True ) for chunk in response: print(chunk.model_dump_json())
- Python Requests
import requests import json url = "http://19912********.cn-beijing.pai-eas.aliyuncs.com/api/predict/quickstart_20250206_pbxi/v1/chat/completions" payload = json.dumps({ "messages": [ { "role": "user", "content": "你是谁？" } ], "max_tokens": 2000 }) headers = { 'Content-Type': 'application/json', 'Authorization': 'Bearer ZDEyMjc0ZTFhNGI***********==' } response = requests.request("POST", url, headers=headers, data=payload) print(response.text)
- 测试结果
{"id":"26a3f338-b328-4bb0-b03a-d266b7bd0ea9","choices":[{"finish_reason":"stop","index":0,"logprobs":null,"message":{"role":"assistant","content":"<think>\n\n</think>\n\n您好！我是由中国的深度求索（DeepSeek）公司开发的智能助手DeepSeek-R1。如您有任何任何问题，我会尽我所能为您提供帮助。"}}],"object":"chat.completion","usage":{"prompt_tokens":7,"completion_tokens":40,"total_tokens":47},"error_info":null}
- PostMan
