Open Models Almanac
Methodsdata curationtokenization

single mention · data curation

Single-image tokenization for multi-camera video inputs

A method that uses single-image tokenization by default for multi-camera video inputs.

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

Also called single-image tokenizer.

Filed under data curation › tokenization