本接口（
/ai/service/embedding）用于根据指定的 Embedding 模型将输入的文本信息转化为特征向量。Method 与 URL
POST https://{实例访问 IP 地址}:{实例网络端口}/ai/service/embedding使用示例
curl -i -X POST \\-H 'Content-Type: application/json' \\-H 'Authorization: Bearer account=root&api_key=A5VOgsMpGWJhUI0WmUbY********************' \\http://10.0.X.X:80/ai/service/embedding \\-d '{"model": "BAAI/bge-m3","modelParams": {"retrieveDenseVector": true,"retrieveSparseVector": true},"dataType": "text","data": ["一种专门存储和检索向量数据的服务", "一种基于向量空间模型的信息检索方法"]}}'
请求参数
参数名 | 参数含义 | 是否必须 | 参数配置及限制 |
model | 指定需使用的 Embedding 模型的名称 | 是 | bge-large-zh-v1.5：适用中文，1024维，推荐使用。 bge-base-zh-v1.5：适用中文，768维。 bge-large-zh：适用中文，1024维。 bge-base-zh：适用中文，768维。 m3e-base：适用中文，768维。 e5-large-v2：适用英文，1024维。 text2vec-large-chinese：适用中文，1024维。 multilingual-e5-base：适用于多种语言类型，768维。 BAAI/bge-m3：适用于多种语言类型，1024维，支持生成稀疏向量。 |
modelParams | 模型参数，仅部分模型需要配置 | 否 | retrieveDenseVector：指定是否返回稠密向量，默认为 true。 retrieveSparseVector：指定是否返回稀疏向量，默认为 false。仅 BAAI/bge-m3 模型，支持生成稀疏向量。 |
dataType | 传入数据的类型 | 是 | 目前仅支持传入 text。 |
data | 需要向量化的数据 | 是 | 目前仅支持传入文本，支持批量传入，最大支持写入100条数据。 |
输出参数
{"code":0,"msg":"Operation success","requestId":"c60ef8a9affb52184d80923251d79f6f","tokenUsed":22,"denseVector":[[-0.02559197,0.057043996,-0.019487744,0.02586666,-0.017580172,0.011239408,-0.0021975215,-0.006760431,0.00048023096,-0.074288435,0.001995319,0.014772228,-0.039494347,0.01291044,0.039189134,0.011712485,0.009865956,-0.07044277,-0.029727584,0.020952757,0.028735647,-0.017366525,0.047002543,0.015886249,-0.02936133,0.017076574,-0.033481684,0.013513232,0.07196883,-0.0074738623,0.022570377,-0.004665918,-0.02073911,0.05390032,0.016496673,-0.0220973,-0.026278695,0.039219655,0.00887402,0.03549608,-0.0060737054,-0.014932464,0.018205855,-0.009018995,-0.0010253193,-0.040196333,-0.0036606283,-0.0033935686,-0.025469886,-0.020662807,-0.010621354,0.033115428,-0.03604546,0.0029166758,-0.08631376,-0.0024207074,-0.013101196,-0.017183397,-0.08350582,-0.002556145,-0.022753505,-0.038120896,-0.022234645,-0.023379188,0.015382651,-0.039738514,0.053473026,-0.0025256237,0.025561448,-0.030872125,-0.028247308,0.016557714,-0.003622477,0.00025132246,-0.05844797,0.002329144,-0.016038856,-0.020555982,-0.009949889,-0.022600899,0.0085916985,-0.030368527,0.044652417,-0.010339034,-0.02885773,-0.01777856,0.040959362,-0.035618164,0.016527193,-0.007573056,0.018968884,-0.006191975,0.008378051,0.03287126,0.004173765,0.0040211594,-0.017732779,-0.022524595,0.042454895,0.002413077,0.022753505,0.06293458,-0.041173007,0.027820013,0.018007468,-0.012788354,0.011002868,0.0033649548,-0.03531295,-0.017000271,-0.0061843446,-0.0024207074,-0.03906705,-0.030750042,-0.033298556,0.0065887496,-0.017854862,0.014199957,-0.022951892,-0.007710401,0.032718655,-0.002142202,......]],"sparseVector":[{"一种":0.21533203,"专门":0.21313477,"向":0.28149414,"和":0.070739746,"存储":0.24731445,"数据":0.2529297,"检":0.171875,"的服务":0.28466797,"索":0.17932129,"量":0.27197266},{"一种":0.14697266,"向":0.24475098,"基于":0.09118652,"方法":0.16345215,"检":0.19555664,"模型":0.22473145,"的信息":0.19226074,"空间":0.22839355,"索":0.22351074,"量":0.1973877}]}
参数名 | 参数含义 |
tokenUsed | Embedding 消耗的 Token 数量 |
denseVector | 稠密向量。 |
sparseVector | 稀疏向量，仅当模型支持生成稀疏向量，且 returnSparseVector 设置为 true 时返回该参数。 |
