-
Notifications
You must be signed in to change notification settings - Fork 4.7k
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
Force retraining of outdated models #7073
Force retraining of outdated models #7073
Conversation
Thanks for submitting a pull request 🚀 @b-quachtran will take a look at it as soon as possible ✨ |
@alwx 🙏 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thank you for your contribution and sorry that it took me so much time to get it reviewed.
The code looks good. I was thinking that it might make sense to add a test for outdated models and checking if the model is actually detected as the outdated (it'd be easier to just move model_outdated
to a separate function and add a small test for this function)
However, it's a relatively small thing, so I am going to approve this PR anyway.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How did you train the second model? Via rasa train core
? And you could please add a test to verify the bug was fixed (that also helps me to understand the bug better).
Thanks for taking the initiative for the fix!
rasa_version: Text = "1.0", | ||
) -> Fingerprint: | ||
rasa_version: Text = MINIMUM_COMPATIBLE_VERSION, | ||
): |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can you please re-add the removed return type?
@pheel Please feel free to re-open if you have time to continue |
Since this, Rasa is able to boot when detected model is outdated. Since this, Core-only trainings are able to fall back on using an interpreter from the zipped model. If the zipped NLU model is outdated, Rasa will not train.
Status (please check what you already did):
black
(please check Readme for instructions)