2022 November Virtual Minihackathon

[PGijsbers]

The dates are set to 17-18 November

Hi everyone,

We are looking to organize a virtual 2 day mini hackathon in November. The main goal of this hackathon is to have a discussion about the OpenML roadmap and ongoing engineering efforts. Please use this doodle to indicate your availability: https://doodle.com/meeting/participate/id/ejqrE0Wd

While the workshop will be open to everyone who wants to join the discussion, it is not meant to on-board people for code contributions. We will discuss (and possibly work on) the following topics (more may be added):

• Prioritization of tasks and roadmapping the short-term future
• Support for image data on OpenML
• Updated task and dataset definitions: https://github.com/orgs/openml/discussions/3
• API V2 (rollout procedure, first looks(?))
• Remaining issues in parquet rollout: Migration ARFF to Parquet on the OpenML server openml-data#50 (comment)

Hope to see you (virtually) in November!

Pieter

[PGijsbers]

Also to discuss: anonymous uploading

[joaquinvanschoren]

Marcel: Working on Java client (login, configuration)
Jan: Documentation of evaluation engine API function

[joaquinvanschoren]

One cool outcome would be a full example (in a notebook) of building deep learning models on large datasets, especially MetaAlbum

[PGijsbers]

If you would like to join a particular discussion, but are not available the entire hackathon, don't be afraid to suggest a timeslot for the topic!

[PGijsbers]

We scheduled the following discussions (all times CET) join us on gathertown:

Thursday

• 14:00 Anonymous Uploading
• 14:30 Open Parquet Issues
• 16:00 Roadmap

Friday

• 10:00 Dataset and Task Definitions
• 14:00 Dataset extensions for images and other files
• 15:00 API V2 Demo and Discussion