Online entity extractor demo results differs from downloaded 2.0.0 model

[einarbmag]

How to reproduce the behaviour

I've noticed that the online model produces different results than the downloaded spaCy model. In particular, the sentence "who is Florence Ng at Value Partners" gets parsed correctly in the online demo:
https://explosion.ai/demos/displacy-ent?text=who%20is%20Florence%20Ng%20at%20Value%20Partners&model=en_core_web_md&ents=person%2Corg%2Cgpe%2Cloc%2Cproduct%2Cnorp%2Cdate%2Cper%2Cmisc

but not in the locally downloaded "en-core-web-md" model, where all of "Florence Ng at Value Partners" gets classified as an ORG.

import spacy
nlp = spacy.load("en_core_web_md")
doc = nlp("who is Florence Ng at Value Partners")
doc.ents

Output: (Florence Ng at Value Partners,)

What could be the reason for the discrepancy? Is the online demo actually using a different model?

Your Environment

• Operating System: Windows 7
• Python Version Used: 3.6.1
• spaCy Version Used: 2.0.12
• Environment Information:

[honnibal]

Thanks, I'd linked the wrong model to en_core_web_md on the demo. It should be resolved now. I've also made en_core_web_lg available there too.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.