Meaning of NER Training values using Spacy-CodePudding

Please explain the meaning of the columns when training Spacy NER model:

E    #       LOSS TOK2VEC  LOSS NER  ENTS_F  ENTS_P  ENTS_R  SCORE 
---  ------  ------------  --------  ------  ------  ------  ------
  0       0          0.00     78.11   26.82   22.88   32.41    0.27
 26     200         82.40   3935.97   94.44   94.44   94.44    0.94
 59     400         50.37   2338.60   94.91   94.91   94.91    0.95
 98     600         66.31   2646.82   92.13   92.13   92.13    0.92
146     800         85.11   3097.20   94.91   94.91   94.91    0.95
205    1000         92.20   3472.80   94.91   94.91   94.91    0.95
271    1200        124.10   3604.98   94.91   94.91   94.91    0.95

I know that ENTS_F ENTS_P and ENTS_R represent the F-score, precision, and recall respectively and the SCORE is the overall model score. What is the formula for SCORE? Where can I see the documentation about these columns? What are the # and E columns stand for? Please guide or send me to the relevant docs, I didn't find a proper documentation about the columns except here.

CodePudding user response：

# refers to iterations (or batches), and E refers to epochs.

The score is calculated as a weighted average of other metrics, as designated in your config file. This is documented here.