[Question]: 如何在一个SimpleServer里面部署2个不同的taskflow任务呢？

[bfgf52]

请提出你的问题

这是服务端代码

from paddlenlp import Taskflow, SimpleServer

car_schema = ["大众","丰田","别克","奥迪","本田","奔驰","无"]
car_utc = Taskflow("zero_shot_text_classification",
               model="utc-nano",
               schema=car_schema,
               precision="fp32")


book_schema = ["水浒传","三国演义","西游记","红楼梦","无"]
book_utc = Taskflow("zero_shot_text_classification",
               model="utc-nano",
               schema=book_schema,
               precision="fp32")


app = SimpleServer()
app.register_taskflow("taskflow/car_utc", car_utc)
app.register_taskflow("taskflow/book_utc", book_utc)

这是客户端代码

import json
from pprint import pprint

import requests

if __name__ == "__main__":
    url = "http://127.0.0.1:8990/taskflow/car_utc"

    headers = {"Content-Type": "application/json"}


    texts = ["三国志"]

    data = {"data": {"text": texts}}
    r = requests.post(url=url, headers=headers, data=json.dumps(data))
    datas = json.loads(r.text)
    pprint(datas)
    results = []
    for item in datas['result']:
        if len(item['predictions']) > 0:
            label = item['predictions'][0]['label']
            score = item['predictions'][0]['score']
        else:
            label = ''
            score = 0
        results.append({"text":item['text_a'],"label":label,"score":score})
    print(results)
    # [{'text': '三国志', 'label': '三国演义', 'score': 0.9488116450766004}]

问题是：为什么我调用的是car_utc服务，但是预测结果出来的labels是book_schema里面的结果，请问要如何使用，才能使得car_utc对应的是car_schema呢

paddlenlp版本:2.5.2.post

[w5688414]

您好，SimpleServer已经停止更新了，不过SimpleServer是基于fastAPI来写的，您可以直接使用fastAPI，另外，欢迎开发者进行贡献。