Infrastructure requirement

Hi, when we are loading spacy module we are getting out of memory. Please suggest what are the server requirements for this?

For the English models anything with 4gb or more of memory should be okay. If the documents are very long, that might also be an issue.

How much memory have you been trying with?

I have been trying with 16Gb ram.

we have been loading en_core_web_sm, en_core_web_lg models.

That shouldn’t be running out of memory just to load spaCy. If you try the following:

import spacy

nlp = spacy.load('en_core_web_sm')

You should find that loads fine, right?

I suspect you might have a very long document in your dataset, which currently causes a problem for spaCy. Would you be able to check the lengths of the texts you’re loading to confirm?

We followed the DRUG example to train separate tags.Can u give any suggestions?