Thank you for the confirmation @kostajh.

Thanks @kostajh. @MGerlach and I have started running the training pipelines. We shall keep you updated on the progress.

Thank you for proposing this approach @elukey. I have implemented it on draftquality images and the sizes have reduced.

@MGerlach is transitioning maintenance of the link recommendation algorithm from Research to ML-team.

Thank you for the swift changes @elukey, I have +2'd.

Thanks @mmodell marking this as resolved.

Thanks for the update @elukey, it's great that the changes worked :)

I tried generating an authservice_session cookie from KFServing v1.3 dex auth using curl and kept running into a "CSRF check failed" issue. I didn't face this issue with KFServing v1.1.

@ACraze, thank you for suggesting that we add inference-client.sh scripts to the repo.

@ACraze thank you for uploading the model and sharing its storage uri.

@ACraze, thank you for confirming that the enwiki-drafttopic service returns a prediction.

@ACraze thank you for suggesting that we use a KFServing transformer for the data processing step before we make a prediction.

@ACraze, thank you for testing the articlequality inferenceservice.

Thank you for the breakdown @ACraze. Yes, we can close this task.

Yes, this makes sense. Thank you for thinking through the repo structure @ACraze. The previous structure was getting a bit confusing with the duplication but this new structure is much better and coherent. Going to push patches for it today.

@ACraze, thank you for sharing the public URI for the enwiki.damaging model. I have been able to create two inference services for two models (enwiki.goodfaith and enwiki.damaging) using one generic container.

On the KFv1.1 sandbox, I have been able to create a generic revscoring inference service using the files below;

Thanks for the feedback @ACraze. I added authservice-session-generator.sh shell script to enable us easily generate authservice_session cookies on the fly.

Thank you so much for this explanation @ACraze, it helped fill the gaps I had about how the KFv1.1 sandbox setup runs inference services and returns predictions.

The ORES deployment with rebuilt viwiki models has been completed. Thanks to @elukey!

@elukey, that's fine, I set up a meeting and shared it with you on your calendar. Please feel free to adjust the time to a favourable one.

Thank you for the review @Halfak, I have updated the PR with the suggestions you made.

I documented the steps that I took to host Label Studio on Toolforge because they were surprisingly more than I had anticipated. This documentation will also help anyone who might want to host a similar tool (especially Django applications) on Toolforge.

I've created a simple example on how to classify text (Wikipedia edits for example) with three labels: Positive, Negative, and Neutral.

The ORES deployment has been completed. Thanks to @elukey and @klausman.

Thank you so much @elukey, I moved my work to /srv/home/kevinbazira and cleared /home/kevinbazira/. This did the trick.

Thanks to @elukey, I was able to get access using both ssh ores-misc-01.ores-staging.eqiad1.wikimedia.cloud and ssh ores-misc-01.eqiad.wmflabs.

I seem not to have access to ores-misc01.eqiad.wmflabs

Thank you @ACraze and @Theofpa for your suggestions.

The unanswered question I have is whether it is possible to migrate these models out of ORES without retraining them from scratch. I hope so, but I need to dig into how KFServing standalone works before I could say for certain.

I have pushed a PR for this task here where I was able to replace preq with the api-testing REST client within 11 out of 13 tests.

Thanks for the great work! I have had a look at the document and made a couple of suggestions too.

With @daniel's guidance, I pushed a patch for this task here. Below are the steps I followed to test this patch that was added to MW 1.35;

@calbon, I created this PR that added 2020 trwiki data configurations to editquality.

@ACraze, I added the "badwords" and "informals" lists to this page:

I consulted @daniel who confirmed that the WikiCommons API uses the MediaWiki API as this documentation suggests.

@Halfak please see the PR here: https://github.com/wikimedia/editquality/pull/227

@Halfak I followed these instructions and took the following steps;

Thank you for the explanation @Halfak. I have managed to reproduce this issue and fixed the bug.

I tried to reproduce this issue by creating a new edit and tried to undo it here: https://en.wikipedia.beta.wmflabs.org/w/index.php?title=Foobar&action=edit&undoafter=418935&undo=438261