# textcat.batch-train docs fix

**URL:** <https://support.prodi.gy/t/textcat-batch-train-docs-fix/901>\
**Category:** Uncategorized\
**Tags:** docs, textcat, done\
**Created:** [October 20, 2018, 3:14am UTC](https://support.prodi.gy/t/textcat-batch-train-docs-fix/901 "2018-10-20T03:14:12Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![NirantK](https://sea2.discourse-cdn.com/flex020/user_avatar/support.prodi.gy/nirantk/32/484_2.png) [@NirantK](https://support.prodi.gy/u/NirantK)\
**Post date:** [October 20, 2018, 3:14am UTC](https://support.prodi.gy/t/textcat-batch-train-docs-fix/901/1 "2018-10-20T03:14:12Z")

</div>

The documentation at [Workflow: Train a Model](https://prodi.gy/docs/workflow-first-steps)  
states that the training API is:

`prodigy textcat.batch-train metooindia /tmp/model --eval-split 0.2 --label LABEL`

which throws different errors, such as `OSError: [E053] Could not read meta.json from model\meta.json`  
and `prodigy textcat.batch-train: error: unrecognized arguments: --label LABEL`

The correct API which works for me in Prodigy v1.6.1 is  
`prodigy textcat.batch-train metooindia --output-model model --eval-split 0.2`

---

<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:** [October 20, 2018, 11:42am UTC](https://support.prodi.gy/t/textcat-batch-train-docs-fix/901/2 "2018-10-20T11:42:46Z")

</div>

Ah, damn, looks like we forgot to adjust this. Thanks for the heads-up, fixing! 👍