Model performance meta.json file

, ,
1

Hi I was looking at the meta.json file, after training a model, and got a bit confused on what the f, p, r performance score mean. Can someone explain to me about it?

"ents_f":0.8446115288,
"ents_p":0.851010101,
"ents_r":0.8383084577,

Or for a significant label I get :

"EVENTS":{
    "p":0.8103448276,
    "r":0.8103448276,
    "f":0.8103448276
2

Hi! The metrics here are the precision (p), recall (r) and f-score (f). You can also read more about the output and score types in the "Understanding the training output and score types" section here: https://spacy.io/usage/training#score-types

1 Like