webshart/conceptual-captions-12m-webdataset-metadata
Updated
•
82
AI & ML interests
None defined yet.
Recent Activity
View all activity
Organization Card
The webshart format is a community-driven, loosely-organised attempt at pushing a better standard for dataset metadata.
The datasets here are either a converted dataset from a third-party source (such as CC12M) or were created by SimpleTuner or CaptionFlow community members.
models
0
None public yet