-
-
Notifications
You must be signed in to change notification settings - Fork 4.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Simplify adding new component to existing model with CLI #4342
Labels
Comments
svlandeg
added
feat / cli
Feature: Command-line interface
training
Training and updating models
labels
Sep 30, 2019
This was referenced Jan 15, 2020
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Labels
Feature description
It looks like these kinds of cases aren't easy to handle with the train CLI:
I want to train a model from scratch given vectors, one corpus for the tagger, and one corpus for the parser. (See example in Multiple roots per sentence #4306.)
I have a model with a tagger and want to add a parser trained on a separate corpus.
For the internal spacy models, it looks like each component is trained separately and then they are combined using custom scripts.
Could it make sense to have a CLI component that combines models/components (with compatibility checks, of course)?
Case (2) is really just a minor variant of (1), but the train CLI might be able to handle it relatively easily by combining the components in
model-final
, for example. I think case (1) could be handled by the train CLI in theory, but the command-line options would get too complicated and it would be easier to handle it with a separate CLI command.(What is the right way to combine the
vocab
directories (aside fromvectors
) for multiple models?)The text was updated successfully, but these errors were encountered: