# ner train output for Prodigy nightly v1.11.0a11

**URL:** <https://support.prodi.gy/t/ner-train-output-for-prodigy-nightly-v1-11-0a11/4484>\
**Category:** Uncategorized\
**Tags:** usage, spacy, solved\
**Created:** [July 23, 2021, 3:06pm UTC](https://support.prodi.gy/t/ner-train-output-for-prodigy-nightly-v1-11-0a11/4484 "2021-07-23T15:06:23Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![robinsonkwame](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/robinsonkwame/32/1584_2.png) [@robinsonkwame](https://support.prodi.gy/u/robinsonkwame)\
**Post date:** [July 23, 2021, 3:06pm UTC](https://support.prodi.gy/t/ner-train-output-for-prodigy-nightly-v1-11-0a11/4484/1 "2021-07-23T15:06:23Z")

</div>

from `prodigy train --ner first_forms -LS --base-model en_core_web_lg XYZ`,

 ![image](https://us1.discourse-cdn.com/flex020/uploads/prodigy/original/2X/4/49ec55c1672072096a025ecd85a5728a8f7f5801.png)

The `train` output differs quite a bit from [what's shown in the documentation](https://prodi.gy/docs/recipes#train); is there updated documentation or a place I can go to better understand the training output?

EDIT - the output looks to be derived from [spaCy scorer](https://spacy.io/api/scorer) even though previous versions presented differently. Even the spaCy scorer documentation doesn't explicitly spell out what some of this mean. [This stack overflow comment](https://stackoverflow.com/questions/50644777/understanding-spacys-scorer-output) was helpful, but I would've expected documentation on something so product critical, although prodigy is the best thing since sliced bread!

---

<div class="post-metadata">

**Author:** ![ines](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/ines/32/3_2.png) [@ines](https://support.prodi.gy/u/ines)\
**Post date:** [July 24, 2021, 1:10am UTC](https://support.prodi.gy/t/ner-train-output-for-prodigy-nightly-v1-11-0a11/4484/2 "2021-07-24T01:10:02Z")

</div>

The output of the new `prodigy train` command is identical to the output of `spacy train` (since this is what it calls into under the hood). The reported scores are the scores of the individual components, plus a weighted final score. You can read more about this in the spaCy documentation: [https://spacy.io/usage/training#metrics](https://spacy.io/usage/training#metrics)