precision, recall and F-Score

ryanwesslen · August 26, 2022, 1:22pm

Thanks for your question and welcome to the Prodigy community

The scoring is on the evaluation.

The default scoring all happens within spaCy and doesn't depend on Prodigy. You can view the code (for spaCy 3.0) here:

explosion/spaCy/blob/master/spacy/scorer.py

from collections import defaultdict
from typing import (
    TYPE_CHECKING,
    Any,
    Callable,
    Dict,
    Iterable,
    List,
    Optional,
    Set,
    Tuple,
)

import numpy as np

from .errors import Errors
from .morphology import Morphology
from .tokens import Doc, Span, Token
from .training import Example
from .util import SimpleFrozenList, get_lang_class

This file has been truncated. show original

We also have several posts that explain the scoring (e.g., what does it mean) more in detail like:

Topic		Replies	Views
Better score with learning-curve than with train? usage , ner , nightly	4	570	July 18, 2021
Understanding outputs for new texcat model usage , textcat , spacy , training	3	709	October 19, 2021
Output from 'train' recipe - calculating performance metrics ner , spacy	1	521	January 16, 2020
Easiest way to get stats about a model trained in prodigy? ner , spacy , solved	2	363	March 4, 2022
After train: Which model is saved? usage , solved	1	472	September 10, 2020

precision, recall and F-Score

Related Topics