The goal of Bert-As-Service is to give a sentence, and the service can represent the sentence as a semantic vector and return it to the user. [Bert model](https://arxiv.org/abs/1810.04805) is a popular model in the current NLP field. It has achieved good results on a variety of public NLP tasks. The semantic vector calculated by the Bert model is used as input to other NLP models, which will also greatly improve the performance of the model. Bert-As-Service allows users to easily obtain the semantic vector representation of text and apply it to their own tasks. In order to achieve this goal, we have shown in four steps that using Paddle Serving can build such a service in ten minutes. All the code and files in the example can be found in [Example](https://github.com/PaddlePaddle/Serving/tree/develop/python/examples/bert) of Paddle Serving.
Paddle Serving supports various models trained based on Paddle, and saves the serviceable model by specifying the input and output variables of the model. For convenience, we can load a trained bert Chinese model from paddlehub and save a deployable service with two lines of code. The server and client configurations are placed in the `bert_seq20_model` and` bert_seq20_client` folders, respectively.
| model | server configuration and model file path |
| thread | server-side threads |
| port | server port number |
| gpu_ids | GPU index number |
| 参数 | 含义 |
#### Step3: data preprocessing logic on Client Side
| ------- | -------------------------- |
| model | server端配置与模型文件路径 |
| thread | server端线程数 |
| port | server端端口号 |
| gpu_ids | GPU索引号 |
#### Step3：客户端数据预处理逻辑
Paddle Serving has many built-in corresponding data preprocessing logics. For the calculation of Chinese Bert semantic representation, we use the ChineseBertReader class under paddle_serving_app for data preprocessing. Model input fields of multiple models corresponding to a raw Chinese sentence can be easily fetched by developers
We tested the performance of Bert-As-Service based on Padde Serving based on V100 and compared it with the Bert-As-Service based on Tensorflow. From the perspective of user configuration, we used the same batch size and concurrent number for stress testing. The overall throughput performance data obtained under 4 V100s is as follows.
