Some problems about loading two models

[blackforever]

Here is the situation: i added “nlp_spacy” to pipeline, then added an additional model to it, it looks like this:

pipeline:

-name: “nlp_spacy”

model: “…/somePath”

and it worked well.

But recently i find that this “spacy” model is quite the same as the model trained by rasa_nlu(but do have some differences), when i load the bot, these two models will be loaded to the memory at the same time and it takes a lot of time. I was wondering if there is any solution to solve this problem(like to merge these two models or some other ways)?

[Ghostvv (Vladimir Vlasov)]

@blackforever

this “spacy” model is quite the same as the model trained by rasa_nlu

what do you mean? spacy model is language model to get word vectors from. rasa_nlu trains classification model. Looking at the code, we load provided spacy model only (please note depending on given model, it can take long time)