your service is much slower than orignal version

[duongkstn]

Hi,
I tested your docker service with request call by python. I requested by posting 1000 raw sentences.

And I saw that your service (10 seconds) is much slower than orignal version of VnCoreNLP (more than 1 second) (same number of raw sentences, py_vncorenlp library, no multithreading, no multipricessing).

Am I wrong ? Please let me know your solution, thanks