ollama官网
首先前往ollama官网下载ollama
然后现在拉取模型

ollama run deepseek-R1

经过漫长的等待，模型下载完成
我们用命令行测试其默认api

curl -X POST http://localhost:11434/v1/chat/completions ^
More? -H "Content-Type: application/json" ^
More? -d "{\"messages\": [{\"role\": \"user\", \"content\": \"你是谁\"}], \"model\": \"deepseek-R1\", \"stream\": false, \"temperature\": 0.01, \"max_tokens\": 1024}"

响应：

{"id":"chatcmpl-884","object":"chat.completion","created":1739873175,"model":"deepseek-R1","system_fingerprint":"fp_ollama","choices":[{"index":0,"message":{"role":"assistant","content":"\u003cthink\u003e\n\n\u003c/think\u003e\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。"},"finish_reason":"stop"}],"usage":{"prompt_tokens":5,"completion_tokens":40,"total_tokens":45}}

此请求如果使用flask，spring boot等框架发送，即可使用于自己搭建的应用里。