Where can I find explanation of what every parameter in config.cfg means?

I'm experimenting with training and this would help me a lot, but I can't seem to find it.

Also, what is the config that prodigy train uses? It differs a lot from spacy train with default config.

Hi! You can find an overview of the available config settings per section, as well as their expected types, default values and what they mean here:

The individual pipeline components and similar functions also document their config settings. For example, here's the config for the entity recognizer: https://spacy.io/api/entityrecognizer#config

The config Prodigy generates in prodigy train and prodigy data-to-spacy is identical to the default config produced by spacy init config with the given components that you want to train. If you have a config with different settings that you want to use instead, you can provide it via the --config option. In this case, you just need to make sure that it includes all the relevant components that you have created data for in Prodigy :slightly_smiling_face:

1 Like