# how to use --init-token2vec in prodigy train

**URL:** <https://support.prodi.gy/t/how-to-use-init-token2vec-in-prodigy-train/4765>\
**Category:** Uncategorized\
**Tags:** usage, ner, spacy, training\
**Created:** [October 6, 2021, 1:08am UTC](https://support.prodi.gy/t/how-to-use-init-token2vec-in-prodigy-train/4765 "2021-10-06T01:08:37Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![yzzhong](https://avatars.discourse-cdn.com/v4/letter/y/a6a055/32.png) [@yzzhong](https://support.prodi.gy/u/yzzhong)\
**Post date:** [October 6, 2021, 1:08am UTC](https://support.prodi.gy/t/how-to-use-init-token2vec-in-prodigy-train/4765/1 "2021-10-06T01:08:38Z")

</div>

Hi, I am trying to train en\_core\_web\_trf model on some annotations I generated. I saw you use --init-token2vec in the [video tutorial](https://youtu.be/59BKHO_xBPA) however I didn't see it is supported in the [documentation](https://prodi.gy/docs/recipes#training).

Is this still supported? If it is, where can I find the word vector for the en\_core\_web\_trf model?

---

<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 6, 2021, 10:42am UTC](https://support.prodi.gy/t/how-to-use-init-token2vec-in-prodigy-train/4765/2 "2021-10-06T10:42:01Z")

</div>

Hi! In Prodigy v1.11 and spaCy v3, we were able to standardise a lot of this stuff because it can now be specified explicitly in the config, so we don't need all of these Prodigy-specific arguments on the `train` command anymore. You can now specify the path to it as the `init_tok2vec` setting in the `[initialize]` block of your config: [Training Pipelines & Models · spaCy Usage Documentation](https://spacy.io/usage/training#initialization) You can then provide your training config via the `--config` argument. This lets you specify all settings in one file.

> [@yzzhong](#):
>
> If it is, where can I find the word vector for the en\_core\_web\_trf model?

Just for clarification, the `en_core_web_trf` model doesn't include word vectors. It's initialised with transformer weights. If you want to initialize your model with word vectors, you can use the `en_core_web_lg` model.