AidGenSE API

💡注意

该版本是 AidGenSE 的一个临时版本，下一个发布的版本是正式版，接口调用方式会有细微改动，使用案例也会更丰富

所有api都支持两个请求地址，一个带了api版本号，一个是默认api请求地址。

例如下面两个接口访问了同样的功能：

GET /v1/models HTTP/1.1

GET /models HTTP/1.1

1 异常访问

1.1 所有非实现访问

Response 403

Not support method or endpoint, Please check your request.

特别注意：在V0.10A版本中，当--device是cpu/gpu时，未对请求进行拦截，因此不会有如上信息。

1.2 设备未激活

在此状态下，所有接口返回如下Response

{
	"status": "error",
	"msg": "This device has not been activated, please contact aplux."
}

设备未激活相关信息请参考UserGuide文档。

特别注意：在V0.10A版本中，当--device是cpu/gpu时，不校验设备是否激活，因此不会有如上信息。

2 模型相关

2.1 查询可用模型

💡注意

此 API 非 OpenAI 适配接口

api参考： https://platform.openai.com/docs/api-reference/models/list

Request

GET /v1/models HTTP/1.1

Ok Response

{
	"object": "list",
	"loaded_id":"aplux_qwen2_7b",
	"data": [
		{
			"id": "aplux_qwen2_0-5b",
			"created": "1733304878498",
			"owned_by": "aplux",
			"object": "model"
		},
		{
			"id": "aplux_qwen2_1-5b",
			"created": "1733304878498",
			"owned_by": "aplux",
			"object": "model"
		},
		{
			"id": "aplux_qwen2_7b",
			"created": "1733304878498",
			"owned_by": "aplux",
			"object": "model"
		},
		{
            "model_id":"aplux_qwen2-5_3b",
            "model_create":"1736215538226",
            "model_owner":"aplux",
            "cfg_path":"./res/qwen2.5-3b-htp.json"
        }
	]
}

2.2 切换模型

💡注意

此 API 非 OpenAI 适配接口

Request

GET /v1/models/switch/model_id HTTP/1.1

Ok Response

{
	"id": "aplux_qwen2_7b",
	"object": "model"
}

Error Response

{
	"error": {
		"message": "Model does not exist or the switch is wrong",
		"type": "invalid_request_error",
		"param": null,
		"code": "aplux_qwen2_71b"
	}
}

注意，不要频繁切换模型，否则会导致api发生模型加载错误而退出。

特别注意：在V0.10A版本中，当--device是cpu/gpu时，当前后端只输入了一个模型的信息，因此不支持/v1/models/switch api。

3 聊天相关

3.1 创建聊天

api参考： https://platform.openai.com/docs/api-reference/chat/create

No-Stream Request (返回chat completion object)

POST /v1/chat/completions HTTP/1.1
Content-Type: application/json

{
  "model": "aplux_qwen2_7b",
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful assistant."
    },
    {
      "role": "user",
      "content": "给我写一个python打印hello world的例子"
    }
  ]
}

Stream Request (返回chat completion chunk object)

POST /v1/chat/completions HTTP/1.1
Content-Type: application/json

{
  "model": "aplux_qwen2_7b",
  "messages": [
    {
      "role": "user",
      "content": "给我写一个python打印hello world的例子"
    }
  ],
  "stream": true
}

Ok Response (For No-Stream Chat)

{
	"id": "chatcmpl-sV1IqqE1ONeyBDGw9aak0JSxBLcwHjRl",
	"choices": [
		{
			"finish_reason": "stop",
			"index": 0,
			"message": {
				"content": "当然，下面是一个简单的Python程序，用于打印\"Hello World!\"。\n\n```python\nprint(\"Hello World!\")\n```\n\n只需将上述代码复制到Python环境中并运行，即可看到\"Hello World!\"被打印出来。",
				"role": "assistant",
				"refusal": null
			},
			"logprobs": null
		}
	],
	"created": 1733364801,
	"model": "aplux_qwen2_7b",
	"object": "chat.completion",
	"usage": {
		"completion_tokens": 46,
		"prompt_tokens": 0,
		"total_tokens": 46
	},
	"system_fingerprint": "fp_nj0KZoURMe3zP90dZWUjjIJZvNIOt8Fn"
}

Ok Response (For Stream Chat)

data: {"choices":[{"finish_reason":null,"index":0,"delta":{"role":"assistant"}}],"created":1733364537,"id":"chatcmpl-wuwZhLmQDyfcX6YAUSC7xL5pcZDDkOb1","model":"aplux_qwen2_7b","object":"chat.completion.chunk"}

data: {"choices":[{"finish_reason":null,"index":0,"delta":{"content":"当然"}}],"created":1733364537,"id":"chatcmpl-wuwZhLmQDyfcX6YAUSC7xL5pcZDDkOb1","model":"aplux_qwen2_7b","object":"chat.completion.chunk"}

....

data: {"choices":[{"finish_reason":null,"index":0,"delta":{"content":"出来"}}],"created":1733364543,"id":"chatcmpl-wuwZhLmQDyfcX6YAUSC7xL5pcZDDkOb1","model":"aplux_qwen2_7b","object":"chat.completion.chunk","system_fingerprint":"fp_Oy1cNzDki9Qgiyuv01gkzXtqPkzbQvfU"}

data: {"choices":[{"finish_reason":null,"index":0,"delta":{"content":"。"}}],"created":1733364543,"id":"chatcmpl-wuwZhLmQDyfcX6YAUSC7xL5pcZDDkOb1","model":"aplux_qwen2_7b","object":"chat.completion.chunk","system_fingerprint":"fp_5BAbMptvnugjLMsCfaum4x6Vfx5jhpLT"}

data: {"choices":[{"finish_reason":"stop","index":0,"delta":{}}],"created":1733364543,"id":"chatcmpl-wuwZhLmQDyfcX6YAUSC7xL5pcZDDkOb1","model":"aplux_qwen2_7b","object":"chat.completion.chunk","system_fingerprint":"fp_ciYnF3vTj3fV38yX0wuhZUbJZfQIe0tE","usage":{"completion_tokens":46,"prompt_tokens":0,"total_tokens":46}}

data: [DONE]

Error Response

{
	"error": {
		"code": 400,
		"message": "request model(aplux_qwen2_71b) is not available.",
		"type": "invalid_request_error"
	}
}

{
	"error": {
		"code": 400,
		"message": "request model(aplux_qwen2_1-5b) is not loaded, please switch model",
		"type": "invalid_request_error"
	}
}

3.2 停止当前对话

💡注意

此 API 非 OpenAI 适配接口

Request

POST /v1/chat/completed HTTP/1.1
Content-Type: application/json


{
    "chat_id": "chatcmpl-sV1IqqE1ONeyBDGw9aak0JSxBLcwHjRl"
}

Ok Response

{
    "status":"ok",
    "msg":""
}

Error Response

{
	"status": "error",
	"msg": "the chat_id = chatcmpl-sV1IqqE1ONeyBDGw9aak0JSxBLcwHjRl not found"
}