i am sorrory about my english that is poor. Now. i'd like to talk about combine machine learning and openresty.
In my opnion, i have 3 ideas:
local host:
1. use lua-resty-shell in local host (model deployed in host, then use shell cmd to process it)
2. use ffi to call machine learning framework in local host
remote server:
3. use http or rpc request to remote host (machine learning compute in remote host, just upload it to remote server with http request or rpc request)
So, did you have best choice about that?
中文版:-_-+
有一个想法,不知道怎么样能够结合Openresty和机器学习模型做计算,对请求的流量做近些计算。或者是demo耍耍。
我能想到的有三种方式
1. 本机调用shell命令
2. 本机用ffi模式调用
3. 把流量镜像传到服务器端计算
看起来第三种似乎更适合一些,但是有没有更好的方法呢?
请求各位前辈指点。