Skip to content

Mapping a variable-length sentence to a fixed-length vector using BERT model

License

Notifications You must be signed in to change notification settings

JayYip/bert-multitask-service

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

bert-multitask-as-service

A serving service for bert-multitask-learning

Install

pip install bert-multitask-server
pip install bert-multitask-client

Getting Started

  1. Train and export model.

    A typical trained checkpoint dir looks like below.

    bert_serving_ckpt/
    ├── *_label_encoder.pkl
    ├── bert_config.json
    ├── export_model
    ├── params.json
    └── vocab.txt
    
  2. Start server using CLI

    bert-multitask-serving-start -model_dir ~/CWS_NER_POS_ckpt/ -num_worker=4 -problem "CWS|NER|POS"
  3. Use Client to Get Prediction

    from bert_serving.client import BertClient
    bc = BertClient()
    bc.encode(['我爱北京天安门'])

Bert多任务学习服务

一个部署Bert多任务学习的服务

安装

pip install bert-multitask-server
pip install bert-multitask-client

开始使用

  1. 训练模型并导出模型.

    导出后的模型目录应该有以下文件

    bert_serving_ckpt/
    ├── *_label_encoder.pkl
    ├── bert_config.json
    ├── export_model
    ├── params.json
    └── vocab.txt
    
  2. 用CLI启动服务

    bert-multitask-serving-start -model_dir models/ -num_worker=4 -problem "fake_problem"
  3. 用客户端获取预测结果

    from bert_serving.client import BertClient
    bc = BertClient()
    bc.encode(['我爱北京天安门'])

About

Mapping a variable-length sentence to a fixed-length vector using BERT model

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 99.3%
  • Other 0.7%