Open Models Almanac
Methodsmodel architecturechannel mixer

single mention · model architecture

Parallel MLP-Mixer-based action decoder

A parallel MLP-Mixer-based action module used alongside a pretrained vision-language model.

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

Also called -Style Decoder.

Filed under model architecture › channel mixer