Open Models Almanac
Methodsdata curationdata mixture & curriculum

single mention · data curation

Heterogeneous training data pyramid

Structuring VLA training corpora as a data pyramid spanning heterogeneous training data.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See data mixture & curriculum for the methods filed beside it.

Also called a data pyramid.

Filed under data curation › data mixture & curriculum