Llava（图像识别）

正式环境

https://api.302.ai

POST

/302/submit/llava

302.AI的API，来自于我们在云GPU上自己部署的模型。部分模型为开源模型，部分模型为我们自己微调或开发的。

图像识别API，通过上传图片识别图片内容，模型为llava-13b。来自于：https://github.com/haotian-liu/LLaVA
平均运行时长5-10s
价格：0.01 PTC/次

请求参数

Header 参数

Body 参数multipart/form-data

返回响应

🟢200OK

application/json

Body

请求示例

Shell

JavaScript

Java

Swift

curl --location --request POST 'https://api.302.ai/302/submit/llava' \
--header 'Authorization: Bearer ' \
--form 'image=@""' \
--form 'prompt="describe the image"' \
--form 'stream="false"'

响应示例

{
    "completed_at": "",
    "created_at": "2026-01-14T05:59:51.811Z",
    "error": "",
    "id": "17683703914253",
    "model": "llava",
    "output": "[\"The \",\"image \",\"features \",\"a \",\"man \",\"wearing \",\"a \",\"black \",\"jacket \",\"and \",\"a \",\"black \",\"shirt, \",\"standing \",\"in \",\"front \",\"of \",\"a \",\"white \",\"background. \",\"He \",\"appears \",\"to \",\"be \",\"looking \",\"directly \",\"at \",\"the \",\"camera, \",\"possibly \",\"posing \",\"for \",\"a \",\"portrait. \",\"The \",\"man's \",\"facial \",\"expression \",\"is \",\"serious, \",\"and \",\"he \",\"seems \",\"to \",\"be \",\"staring \",\"intently \",\"into \",\"the \",\"distance. \",\"The \",\"overall \",\"atmosphere \",\"of \"]",
    "started_at": "2026-01-14T06:00:21.915208549Z",
    "status": "processing"
}

修改于 2026-01-14 06:01:58

Face-to-many（人像照片风格化）

Relight（二次打光）