SPACY and language issue (using pre trained model ) word2vec to Spacy Format

[jish (Jishnu Nair)]

I’m trying to use a pre-trained model

Using word2vec I have converted the data to ZIP format But using the latest Spacy 3.0 version I am not able to convert it into Spacy usable format

I tried python -m spacy init-model en ./spacy.word2vec.model --vectors-loc word2vec_shopping.txt.gz

word2vec_shopping.txt.gz is the file name

This is the error I’m getting

Error: No such command ‘init-model’.

I tried doing it with older versions it worked fine but in that case, I am not able to add it to RASA

Can someone help us with this

HOW DO I SOLVE THE FIRST ISSUE AND THE SECOND ONE

I AM TRYING TO USE A PRE TRAINED MODEL

PLEASE HELP ME WITH THIS OR IF POSSIBLE PLEASE SUGGEST AN ALTERNATIVE WAY ACHIEVE IT

[souvikg10 (Souvik Ghosh)]

when you say a pre-trained model do you mean a custom one trained by you in spacy’s format?

I am speaking of spaCy 2.x if you have a custom spacy model trained into a folder

then in the config.yml language you need to provide the path to that spacy model instead of en_core_web_md , you don’t need to link that model in particular. providing the path to the model also works fine

[jish (Jishnu Nair)]

Hi Souvik

Thank you For your Response

Yes, custom one trained in spacy’s format

Let me try doing this way Ill update you if it works

[jish (Jishnu Nair)]

Thank you @souvikg10 t worked for me

probably the mistake I did was

spacy 3.0

Also, I had a Doubt which is if I want to add two or more such models how do I do it? In the current approach its not possible

Eg: If I have 3 intent in my NLU file I want to train models for all three (is it possible ?)

if not possible can you give me some tips, how I can make my chatbot more efficient in understanding user inputs?