snipsco / snips-nlu

Snips Python library to extract meaning from text
https://snips-nlu.readthedocs.io
Apache License 2.0
3.9k stars 512 forks source link

nlu_engine.parse() it takes too much time #894

Open harrak2020 opened 3 years ago

harrak2020 commented 3 years ago

it does not matter if interactively by running: snips-nlu parse path/to/trained_engine on the CMD or by python script : nlu_engine.parse(text,intents=["intent1"]) time increases exponentially if you just add a slot in text for example: text = "one two three four five six seven eight nine ten two three four five six" 15 slots ; takes 32.138450s. text = "one two three four five six seven eight nine ten two three four five" 14 slots ; takes 14.838444s. text = "one two three four five six seven eight nine ten two three four" 13 slots ; takes 7.012628s. text = "one two three four five six seven eight nine ten two three" 12 slots ; takes 3.441175s.

"entity": "snips/number"

is there a solution to this problem