添加设置 timeout 参数逻辑

[caiiiac]

在使用 PaddleHub 搭建 OCR 服务的时候发现一个问题,在开启"use_multiprocess"的情况下,由于采用了"Gunicorn",其默认超时时间比较短,在进行多张图片识别的时候服务器经常报超时错误([CRITICAL] WORKER TIMEOUT)
所以添加了设置"timeout"参数的功能,

1. 可以通过命令行 --timeout 或者 -t 赋值,默认为300秒,当值为0就表示禁用超时设置.如:
   hub serving start -m ocr_system --use_multiprocess --workers --timeout 0

2. 也可以在"config.json "中添加
   {
   "modules_info": {
   "ocr_system": {
   "init_args": {
   "version": "1.0.0",
   "use_gpu": false
   },
   "predict_args": {
   }
   }
   },
   "port": 8868,
   "use_multiprocess": true,
   "workers": 2,
   "timeout": 0
   }

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[MyOnlyCat]

感谢兄弟,才遇到这种超时问题