如何使用硅基流动提供的免费deepseek-ocr api

🛠工具与编程 💻编程
, ,
1 
import requests
import base64

url = "https://api.siliconflow.cn/v1/chat/completions"

payload = {
    "model": "deepseek-ai/DeepSeek-OCR",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                "image_url": { 
                "url": "data:application/pdf;base64," + base64.b64encode(
                            open("test_image.jpg", "rb").read()).decode("utf-8")
                            },
                    "type": "image_url"
                },
                {
                    "type": "text",
                    # text 是提示词，根据要识别的东西写合适的提示词
                    "text": "TABLE_EXTRACTION = '\u003cimage\u003e\\n\u003c|grounding|\u003eExtract tables from this document and convert to markdown format.'"
                }
            ]
        }
    ],
    # "max_tokens": 8002,
    "stream": False,
    "response_format": { "type": "text" }
}
headers = {
    "Authorization": "Bearer 你的 API 密钥",
    # 比如     "Authorization": "Bearer testtesttest", (bearer和apikey之间加上空格)
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

# 提取content内容
result = response.json()
content = result['choices'][0]['message']['content']
print(content)
multi-ocr-sdk：支持多种ocr引擎的pip包